with whom Justice Blackmun and Justice Stevens join, dissenting.
Oklahoma gained statehood in 1907. For the next 65 years, the Oklahoma City School Board (Board) maintained segregated schools — initially relying on laws requiring dual school systems; thereafter, by exploiting residential segregation that had been created by legally enforced restrictive covenants. In 1972 — 18 years after this Court first found segregated schools unconstitutional — a federal court finally interrupted this cycle, enjoining the Board to implement a specific plan for achieving actual desegregation of its schools.
The practical question now before us is whether, 13 years after that injunction was imposed, the same Board should have been allowed to return many of its elementary schools to their former one-race status. The majority today suggests that 13 years of desegregation was enough. The Court remands the case for further evaluation of whether the purposes of the injunctive decree were achieved sufficient to justify the decree’s dissolution. However, the inquiry it commends to the District Court fails to recognize explicitly the threatened reemergence of one-race schools as a relevant “vestige” of de jure segregation.
In my view, the standard for dissolution of a school desegregation decree must reflect the central aim of our school desegregation precedents. In Brown v. Board of Education, 347 U. S. 483 (1954) (Brown I), a unanimous Court declared that racially “[sjeparate educational facilities are inherently *252unequal.” Id., at 495. This holding rested on the Court’s recognition that state-sponsored segregation conveys a message of “inferiority as to th[e] status [of Afro-American school children] in the community that may affect their hearts and minds in a way unlikely ever to be undone.” Id., at 494. Remedying this evil and preventing its recurrence were the motivations animating our requirement that formerly de jure segregated school districts take all feasible steps to eliminate racially identifiable schools. See Green v. New Kent County School Bd., 391 U. S. 430, 442 (1968); Swann v. Charlotte-Mecklenburg Bd. of Education, 402 U. S. 1, 25-26 (1971).
I believe a desegregation decree cannot be lifted so long as conditions likely to inflict the stigmatic injury condemned in Brown I persist and there remain feasible methods of eliminating such conditions. Because the record here shows, and the Court of Appeals found, that feasible steps could be taken to avoid one-race schools, it is clear that the purposes of the decree have not yet been achieved and the Court of Appeals’ reinstatement of the decree should be affirmed. I therefore dissent.1
I
In order to assess the full consequence of lifting the decree at issue in this case, it is necessary to explore more fully than does the majority the history of racial segregation in the Oklahoma City schools. This history reveals nearly unflagging resistance by the Board to judicial efforts to dismantle the city’s dual education system.
When Oklahoma was admitted to the Union in 1907, its Constitution mandated separation of Afro-American children *253from all other races in the public school system. Dowell v. School Bd. of Oklahoma City Public Schools, 219 F. Supp. 427, 431 (WD Okla. 1963). In addition to laws enforcing segregation in the schools, racially restrictive covenants, supported by state and local law, established a segregated residential pattern in Oklahoma City. 677 F. Supp. 1503, 1506 (WD Okla. 1987). Petitioner Board exploited this residential segregation to enforce school segregation, locating “all-Negro” schools in the heart of the city’s northeast quadrant, in which the majority of the city’s Afro-American citizens resided. Dowell, supra, at 433-434.
Matters did not change in Oklahoma City after this Court’s decision in Brown I and Brown v. Board of Education, 349 U. S. 294 (1955) (Brown II). Although new school boundaries were established at that time, the Board also adopted a resolution allowing children to continue in the schools in which they were placed or to submit transfer requests that would be considered on a case-by-case basis. Dowell, 219 F. Supp., at 434. Because it allowed thousands of white children each year to transfer to schools in which their race was the majority, this transfer policy undermined any potential desegregation. See id., at 440-441, 446.
Parents of Afro-American children relegated to schools in the northeast quadrant filed suit against the Board in 1961. Finding that the Board’s special transfer policy was “designed to perpetuate and encourage segregation,” id., at 441, the District Court struck down the policy as a violation of the Equal Protection Clause, id., at 442. Undeterred, the Board proceeded to adopt another special transfer policy which, as the District Court found in 1965, had virtually the same effect as the prior policy — “perpetuat[ion] [of] a segregated system.” Dowell v. School Bd. of Oklahoma City Public Schools, 244 F. Supp. 971, 975 (WD Okla. 1965), aff’d in part, 375 F. 2d 158 (CA10), cert. denied, 387 U. S. 931 (1967).
*254The District Court also noted that, by failing to adopt an affirmative policy of desegregation, the Board had reversed the desegregation process in certain respects. For example, eight of the nine new schools planned or under construction in 1965 were located to serve all-white or virtually all-white school zones. 244 F. Supp., at 975. Rather than promote integration through new school locations, the District Court found that the Board destroyed some integrated neighborhoods and schools by adopting inflexible neighborhood school attendance zones that encouraged whites to migrate to all-white areas. Id., at 976-977. Because the Board’s pupil assignments coincided with residential segregation initiated by law in Oklahoma City, the Board also preserved and augmented existing residential segregation. Ibid.
Thus, by 1972, 11 years after the plaintiffs had filed suit and 18 years after our decision in Brown /, the Board continued to resist integration and in some respects the Board had worsened the situation. Four years after this Court’s admonition to formerly de jure segregated school districts to come forward with realistic plans for immediate relief, see Green v. New Kent County School Bd., supra, at 439, the Board still had offered no meaningful plan of its own. Instead, “[i]t rationalize^] its intransigence on the constitutionally unsound basis that public opinion [was] opposed to any further desegregation.” Dowell v. Board of Education of Oklahoma City Public Schools, 338 F. Supp. 1256, 1270 (WD Okla.), aff’d, 465 F. 2d 1012 (CA10), cert. denied, 409 U. S. 1041 (1972). The District Court concluded: “This litigation has been frustratingly interminable, not because of insuperable difficulties of implementation of the commands of the Supreme Court. . . and the Constitution . . . but because of the unpardonable recalcitrance of the . . . Board.” 338 F. Supp., at 1271. Consequently, the District Court ordered the Board to implement the only available plan that exhibited the promise of achieving actual desegregation — the “Finger Plan” offered by the plaintiffs. Id., at 1269.
*255In 1975, after a mere three years of operating under the Finger Plan, the Board filed a “Motion to Close Case,” arguing that it had “‘eliminated all vestiges of state imposed racial discrimination in its school system.’” Dowell v. Board of Education of Oklahoma City Public Schools, 606 F. Supp. 1548, 1551 (WD Okla. 1985) (quoting motion), rev’d, 795 F. 2d 1516 (CA10), cert. denied, 479 U. S. 938 (1986). In 1977, the District Court granted the Board’s motion and issued an “Order Terminating Case.” The court concluded that the Board had “operated the [Finger] Plan properly” and stated that it did not “foresee that the termination of . . . jurisdiction will result in the dismantlement of the [Finger] Plan or any affirmative action by the defendant to undermine the unitary system.” App. 174-175. The order ended the District Court’s active supervision of the school district but did not dissolve the injunctive decree. The plaintiffs did not appeal this order.
The Board continued to operate under the Finger Plan until 1985, when it implemented the Student Reassignment Plan (SRP). The SRP superimposed attendance zones over some residentially segregated areas. As a result, considerable racial imbalance reemerged in 33 of 64 elementary schools in the Oklahoma City system with student bodies either greater than 90% Afro-American or greater than 90% non-Afro-American. Dowell, 606 F. Supp., at 1553. More specifically j 11 of the schools ranged from 96.9% to 99.7% Afro-American, and approximately 44% of all Afro-American children in grades K-4 were assigned to these virtually all-Afro-American schools. See 890 F. 2d 1483, 1510, n. 4. (CA10 1989) (Baldock, J., dissenting).2
In response to the SRP, the plaintiffs moved to reopen the case. Ultimately, the District Court dissolved the deseg*256regation decree, finding that the school district had been “unitary” since 1977 and that the racial imbalances under the SRP were the consequence of residential segregation arising from “personal preferences.” 677 F. Supp., at 1512. The Court of Appeals reversed, finding that the Board had not met its burden to establish that “the condition the [decree] sought to alleviate, a constitutional violation, has been eradicated.” 890 F. 2d, at 1491.
r — I H — I
I agree with the majority that the proper standard for determining whether a school desegregation decree should be dissolved is whether the purposes of the desegregation litigation, as incorporated in the decree, have been fully achieved. Ante, at 247, citing United States v. Swift & Co., 286 U. S. 106 (1932). See United States v. United Shoe Machinery Corp., 391 U. S. 244, 248 (1968); Pasadena City Bd. of Education v. Spangler, 427 U. S. 424, 436-437 (1976); id., at 444 (Marshall, J., dissenting) (“We should not compel the District Court to modify its order unless conditions have changed so much that ‘dangers, once substantial, have become attenuated to a shadow,’” quoting, Swift, supra, at 119).3 I strongly disagree with the majority, however, on what must be shown to demonstrate that a decree’s purposes *257have been fully realized.4 In my view, a standard for dissolution of a desegregation decree must take into account the unique harm associated with a system of racially identifiable schools and must expressly demand the elimination of such schools.
A
Our pointed focus in Brown I upon the stigmatic injury caused by segregated schools explains our unflagging insistence that formerly de jure segregated school districts extinguish all vestiges of school segregation. The concept of stigma also gives us guidance as to what conditions must be eliminated before a decree can be deemed to have served its purpose.
In the decisions leading up to Brown I, the Court had attempted to curtail the ugly legacy of Plessy v. Ferguson, 163 U. S. 537 (1896), by insisting on a searching inquiry into whether “separate” Afro-American schools were genuinely “equal” to white schools in terms of physical facilities, curricula, quality of the faculty, and certain “intangible” considerations. See, e. g., Sweatt v. Painter, 339 U. S. 629 (1950); Sipuel v. Board of Regents of Univ. of Okla., 332 U. S. 631 (1948). In Brown I, the Court finally liberated the Equal Protection Clause from the doctrinal tethers of Plessy, declaring that “in the field of public education the doctrine of ‘separate but equal’ has no place. Separate educational facilities are inherently unequal.” 347 U. S., at 495.
The Court based this conclusion on its recognition of the particular social harm that racially segregated schools inflict on Afro-American children.
*258“To separate them from others of similar age and qualifications solely because of their race generates a feeling of inferiority as to their status in the community that may affect their hearts and minds in a way unlikely ever to be undone. The effect of this separation on their educational opportunities was well stated by a finding in the Kansas case by a court which nevertheless felt compelled to rule against the Negro plaintiffs:
“‘Segregation of white and colored children in public schools has a detrimental effect upon the colored children. The impact is greater when it has the sanction of the law; for the policy of separating the races is usually interpreted as denoting the inferiority of the negro group. A sense of inferiority affects the motivation of a child to learn. Segregation with the sanction of law, therefore, has a tendency to [retard] the educational and mental development of negro children and to deprive them of some of the benefits they would receive in a racially] integrated school system.’” Id., at 494.
Remedying and avoiding the recurrence of this stigmatizing injury have been the guiding objectives of this Court’s desegregation jurisprudence ever since. These concerns inform the standard by which the Court determines the effectiveness of a proposed desegregation remedy. See Green v. New Kent County School Bd., 391 U. S. 430 (1968). In Green, a school board sought to implement the mandate of Brown I and Brown II by adopting a “freedom of choice” plan under which individual students could specify which of two local schools they would attend. The Court held that this plan was inadequate because it failed to redress the effect of segregation upon “every facet of school operations — faculty, staff, transportation, extracurricular activities and facilities.” 391 U. S., at 435. By so construing the extent of a school board’s obligations, the Court made clear that the Equal Protection Clause demands elimination of every indicium of a “[r]acial[ly] identifiable]” school system that will inflict the stigmatizing injury that Brown I sought to cure. Ibid. *259Accord, Swann v. Charlotte-Mecklenburg Bd. of Education, 402 U. S., at 15.
Concern with stigmatic injury also explains the Court’s requirement that a formerly de jure segregated school district provide its victims with “make whole” relief. In Milliken v. Bradley, 418 U. S. 717 (1974) (Milliken I), the court concluded that a school desegregation decree must “restore the victims of discriminatory conduct to the position they would have occupied in the absence of such conduct.” Id., at 746. In order to achieve such “make whole” relief, school systems must redress any effects traceable to former de jure segregation. See Milliken v. Bradley, 433 U. S. 267, 281-288 (1977) (Milliken II) (upholding remedial education programs and other measures to redress the substandard communication skills of Afro-American students formerly placed in segregated schools). The remedial education upheld in Milli-ken II was needed to help prevent the stamp of inferiority placed upon Afro-American children from becoming a self-perpetuating phenomenon. See id., at 287.
Similarly, avoiding reemergence of the harm condemned in Brown I accounts for the Court’s insistence on remedies that ensure lasting integration of formerly segregated systems. Such school districts are required to “make every effort to achieve the greatest possible degree of actual desegregation and [to] be concerned with the elimination of one-race schools.” Swann, supra, at 26 (emphasis added). See Dayton Bd. of Education v. Brinkman, 443 U. S. 526, 538 (1979); Columbus Bd. of Education v. Penick, 443 U. S. 449, 460 (1979); Raney v. Board of Education of Gould School Dist., 391 U. S. 443, 449 (1968) (endorsing the “ ‘goal of a desegregated, non-racially operated school system [that] is rapidly and finally achieved,”’ quoting Kelley v. Altheimer, 378 F. 2d 483, 489 (CA8 1967) (emphasis added)). This focus on “achieving and preserving an integrated school system,” Keyes v. School Dist. No. 1, Denver, Colo., 413 U. S. 189, 251, n. 31 (1973) (Powell, J., concurring in part and dissent*260ing in part) (emphasis added), stems from the recognition that the reemergence of racial separation in such schools may revive the message of racial inferiority implicit in the former policy of state-enforced segregation.5
Just as it is central to the standard for evaluating the formation of a desegregation decree, so should the stigmatic injury associated with segregated schools be central to the standard for dissolving a decree. The Court has indicated that “the ultimate end to be brought about” by a desegregation remedy is “a unitary, nonracial system of public education.” Green, supra, at 436. We have suggested that this aim is realized once school officials have “eliminate[d] from the public schools all vestiges of state-imposed segregation,” Swann, supra, at 15 (emphasis added), whether they inhere in the school’s “faculty, staff, transportation, extracurricular activities and facilities,” Green, supra, at 435, or even in “the community and administration's] attitudes toward [a] school,” Keyes, supra, at 196. Although the Court has never explicitly defined what constitutes a “vestige” of state-enforced segregation, the function that this concept has per*261formed in our jurisprudence suggests that it extends to any condition that is likely to convey the message of inferiority implicit in a policy of segregation. So long as such conditions persist, the purposes of the decree cannot be deemed to have been achieved.
B
The majority suggests a more vague and, I fear, milder standard. Ignoring the harm identified in Brown I, the majority asserts that the District Court should find that the purposes of the decree have been achieved so long as “the Oklahoma City School District [is now] being operated in compliance with the commands of the Equal Protection Clause” and “it [is] unlikely that the Board would return to its former ways.” Ante, at 247. Insofar as the majority instructs the District Court, on remand, to “conside[r] whether the vestiges of de jure segregation ha[ve] been eliminated as far as practicable,” ante, at 250, the majority presumably views elimination of vestiges as part of “operating] in compliance with the commands of the Equal Protection Clause.” But as to the scope or meaning of “vestiges,” the majority says very little.
By focusing heavily on present and future compliance with the Equal Protection Clause, the majority’s standard ignores how the stigmatic harm identified in Brown I can persist even after the State ceases actively to enforce segregation.6 It was not enough in Green, for example, for the school district to withdraw its own enforcement of segregation, leaving it up to individual children and their families to “choose” *262which school to attend. For it was clear under the circumstances that these choices would be shaped by and perpetuate the state-created message of racial inferiority associated with the school district’s historical involvement in segregation. In sum, our school-desegregation jurisprudence establishes that the effects of past discrimination remain chargeable to the school district regardless of its lack of continued enforcement of segregation, and the remedial decree is required until those effects have been finally eliminated.
HH HH
Applying the standard I have outlined, I would affirm the Court of Appeals’ decision ordering the District Court to restore the desegregation decree. For it is clear on this record that removal of the decree will result in a significant number of racially identifiable schools that could be eliminated.
As I have previously noted:
“Racially identifiable schools are one of the primary vestiges of state-imposed segregation which an effective desegregation decree must attempt to eliminate. In Swann, supra, for example, we held that ‘[t]he district judge or school authorities . . . will thus necessarily be concerned with the elimination of one-race schools.’ 402 U. S., at 26. There is ‘a presumption,’ we stated, ‘against schools that are substantially disproportionate in their racial composition. ’ Ibid. And in evaluating the effectiveness of desegregation plans in prior cases, we ourselves have considered the extent to which they discontinued racially identifiable schools. See, e. g., Green v. County School Board of New Kent County, supra; Wright v. Council of the City of Emporia, [407 U. S. 451 (1972)]. For a principal end of any desegregation remedy is to ensure that it is no longer ‘possible to identify a “white school” or a “Negro school,”’ Swann, supra, at 18. The evil to be remedied in the dismantling of a dual system is the ‘[r]acial identification of the *263system’s schools.’ Green, 391 U. S., at 435. The goal is a system without white schools or Negro schools-^a system with ‘just schools.’ Id., at 442. A school authority’s remedial plan or a district court’s remedial decree is to be judged by its effectiveness in achieving this end. See Swann, supra, at 25; Davis [v. Board of School Comm’rs of Mobile County, 402 U. S. 33, 37 (1971)]; Green, supra, at 439.” Milliken I, 418 U. S., at 802-803 (Marshall, J., dissenting).
Against the background of former state-sponsorship of one-race schools, the persistence of racially identifiable schools perpetuates the message of racial inferiority associated with segregation. Therefore, such schools must be eliminated whenever feasible.
It is undisputed that replacing the Finger Plan with a system of neighborhood school assignments for grades K-4 resulted in a system of racially identifiable schools. Under the SRP, over one-half of Oklahoma City’s elementary schools now have student bodies that are either 90% Afro-American or 90% non-Afro-American. See supra, at 255. Because this principal vestige of de jure segregation persists, lifting the decree would clearly be premature at this point. See Davis v. East Baton Rouge Parish School Bd., 721 F. 2d 1425, 1434 (CA5 1983) (“[T]he continued existence of one-race schools is constitutionally unacceptable when reasonable alternatives exist”).
The majority equivocates on the effect to be given to the reemergence of racially identifiable schools. It instructs the District Court to consider whether those “‘most important indicia of a segregated system’” have been eliminated, reciting the facets of segregated school operations identified in Green — faculty, staff, transportation, extracurricular activities and facilities.’” Ante, at 250. And, by rendering “res nova” the issue whether residential segregation in Oklahoma City is a vestige of former school segregation, ante at 250, n. 2, the majority accepts at least as a theoretical possibility *264that vestiges may exist beyond those identified in Green. Nonetheless, the majority hints that the District Court could ignore the effect of residential segregation in perpetuating racially identifiable schools if the court finds residential segregation to be “the result of private decisionmaking and economics.” Ibid. Finally, the majority warns against the application of a standard that would subject formerly segregated school districts to the “Draconian” fate of “judicial tutelage for the indefinite future.” Ante, at 249.7
This equivocation is completely unsatisfying. First, it is well established that school segregation “may have a profound reciprocal effect on the racial composition of residential neighborhoods.” Keyes, 413 U. S., at 202; see also Columbus Bd. of Education, 443 U. S., at 465, n. 13 (acknowledging the evidence that “school segregation is a contributing cause of housing segregation”). The record in this case amply demonstrates this form of complicity in residential segregation on the part of the Board.8 The District Court *265found as early as 1965 that the Board’s use of neighborhood schools “serve[d] to . . . exten[d] areas of all Negro housing, destroying in the process already integrated neighborhoods and thereby increasing the number of segregated schools.” 244 F. Supp., at 977. It was because of the Board’s responsibility for residential segregation that the District Court refused to permit the Board to superimpose a neighborhood plan over the racially isolated northeast quadrant. See id., at 976-977.
Second, there is no basis for the majority’s apparent suggestion that the result should be different if residential segregation is now perpetuated by “private decisionmaking.” The District Court’s conclusion that the racial identity of the northeast quadrant now subsists because of “personal preference[s],” 677 F. Supp., at 1512, pays insufficient attention to the roles of the State, local officials, and the Board in creating what are now self-perpetuating patterns of residential segregation. Even more important, it fails to account for the unique role of the School Board in creating “all-Negro” schools clouded by the stigma of segregation — schools to which white parents would not opt to send their children. That such negative “personal preferences” exist should not absolve a school district that played a role in creating such “preferences” from its obligation to desegregate the schools to the maximum extent possible.9
*266I also reject the majority’s suggestion that the length of federal judicial supervision is a valid factor in assessing a dissolution. The majority is correct' that the Court has never contemplated perpetual judicial oversight of former de jure segregated school districts. Our jurisprudence requires, however, that the job of school desegregation be fully completed and maintained so that the stigmatic harm identified in Brown I will not recur upon lifting the decree. Any doubt on the issue whether the School Board has fulfilled its remedial obligations should be resolved in favor of the Afro-Amercan children affected by this litigation.10
*267In its concern to spare local school boards the “Draconian” fate of “indefinite” “judicial tutelage,” ante, at 249, the majority risks subordination of the constitutional rights of Afro-American children to the interest of school board autonomy.11 The courts must consider the value of local control, but that factor primarily relates to the feasibility of a remedial measure, see Milliken II, 433 U. S., at 280-281, not whether the constitutional violation has been remedied. Swann establishes that if further desegregation is “reasonable, feasible, and workable,” 402 U. S., at 31, then it must be undertaken. In assessing whether the task is complete, the dispositive question is whether vestiges capable of inflicting stigmatic harm exist in the system and whether all that can practicably be done to eliminate those vestiges has been done. The Court of Appeals concluded that “on the basis of the record, it is clear that other measures that are feasible remain available to the Board [to avoid racially identifiable schools].” 890 F. *2682d, at 1505. The School Board does not argue that further desegregation of the one-race schools in its system is unworkable and in light of the proven feasibility of the Finger Plan, I see no basis for doubting the Court of Appeals’ finding.
We should keep in mind that the court’s active supervision of the desegregation process ceased in 1977. Retaining the decree does not require a return to active supervision. It may be that a modification of the decree which will improve its effectiveness and give the school district more flexibility in minimizing busing is appropriate in this case. But retaining the decree seems a slight burden on the school district compared with the risk of not delivering a full remedy to the Afro-American children in the school system.12
I — I <1
Consistent with the mandate of Brown I, our cases have imposed on school districts an unconditional duty to eliminate any condition that perpetuates the message of racial inferiority inherent in the policy of state-sponsored segregation. The racial identifiability of a district’s schools is such a condition. Whether this “vestige” of state-sponsored segregation will persist cannot simply be ignored at the point where a district court is contemplating the dissolution of a desegregation decree. In a district with a history of state-sponsored school segregation, racial separation, in my view, remains inherently unequal.
I dissent.
The issue of decree modification is not before us. However, I would not rule out the possibility of petitioner demonstrating that the purpose of the decree at issue could be realized by less burdensome means. Under such circumstances a modification affording petitioner more flexibility in redressing the lingering effects of past segregation would be warranted. See infra, at 268.
As a result of school closings, currently there are 10 all-Afro-American elementary schools in the system, 890 F. 2d, at 1512, n. 7 (Baldock, J., dissenting). According to respondents, all but one of these schools are located in the northeast quadrant. Brief for Respondents 17.
I also strongly agree with the majority’s conclusion that, prior to the dissolution of a school desegregation decree, plaintiffs are entitled to a precise statement from a district court. Ante, at 246. Because of the sheer importance of a desegregation decree’s objectives, and because the dissolution of such a decree will mean that plaintiffs will have to mount a new constitutional challenge if they wish to contest the segregative effects of the school board’s subsequent actions, the district court must give a detailed explanation of how the standards for dissolution have been met. Because the District Court’s 1977 order terminating its “active jurisdiction” did not contain such a statement, that order does not bar review of its 1987 order expressly dissolving the decree.
Perhaps because of its preoccupation with overturning the Court of Appeals’ invocation of the “grievous wrong” language from United States v. Swift, 286 U. S. 106 (1932), see ante, at 243-244, the majority’s conception of the purposes of a desegregation decree is not entirely clear. See infra, at 263-264.
Because of the relative indifference of school boards toward all-Afro-American schools, many of these schools continue to suffer from high student-faculty ratios, lower quality teachers, inferior facilities and physical conditions, and lower quality course offerings and extracurricular programs. See Note, 87 Colum. L. Rev. 794, 801 (1987); see also Camp, Thompson, & Crain, Within-District Equity: Desegregation and Microeco-nomic Analysis, in The Impacts of Litigation and Legislation on Public School Finance 273,282-286 (J. Underwood & D. Verstegen eds. 1990) (citing recent studies indicating that because of systematic biases, predominately minority public schools typically receive fewer resources than other schools in the same district).
Indeed, the poor quality of a system’s schools may be so severe that nothing short of a radical transformation of the schools within the system will suffice to achieve desegregation and eliminate all of its vestiges. See Jenkins v. Missouri, 855 F. 2d 1295, 1301-1307 (CA8 1988), aff’d in part and rev’d in part on other grounds, 495 U. S. 33 (1990) (desegregation plan required every high school, every middle school, and half of the elementary schools in the school system to become magnet schools).
Faithful compliance with the decree admittedly is relevant to the standard for dissolution. The standard for dissolution should require that the school district have exhibited faithful compliance with the decree for a period sufficient to assure the District Court that the school district is committed to the ideal of an integrated system. Cf. Morgan v. Nucci, 831 F. 2d 313, 321 (CA1 1987) (addressing whether the school district has exhibited sufficient good faith “to indicate that further oversight of [student] assignments is not needed to forestall an imminent return to the unconstitutional conditions that led to the court’s intervention”).
The majority also instructs the District Court to consider whether dissolution was appropriate “as of 1985,” ante, at 249, prior to the Board’s adoption of the SRP. However, the effect of the Board’s readoption of neighborhood attendance zones cannot be ignored arbitrarily. A district court, in evaluating whether dissolution of a desegregation decree is warranted, must consider whether conditions exist that are capable of inflicting the stigmatic harms associated with the original violation. The SRP demonstrates that lifting the decree would result in one-race schools which the decree was designed to eliminate. Even in cases lacking such tangible evidence of unremoved vestiges, a district court must anticipate what effect lifting a decree will have in order to assess dissolution.
Again, our commitment to “make whole” relief requires that any injurious condition flowing from the constitutional violation must be remedied to the maximum extent practicable. See Milliken II, 433 U. S. 267, 280-281, 287-288 (1977). Therefore, beyond eliminating vestiges concerning “faculty, staff, transportation, extracurricular activities and facilities,” Green v. New Kent County School Bd., 391 U. S. 430, 435 (1968), other measures may be necessary to treat a “root condition shown by [the] record.” Milliken II, supra, at 288. The remedial obligations of a school board, therefore, are defined by the effects of the board’s past discriminatory conduct. On the issue whether residential segregation is a vestige, *265the relevant inquiry is whether the record shows that the board’s past actions were a “contributing cause” to residential segregation. Columbus Bd. of Education v. Penick, 443 U. S. 449, 465, n. 13 (1979).
Resistance to busing and the desire to attract white students to the public school system have been among the key motivations for incorporating magnet schools into desegregation plans. See Selig, The Reagan Justice Department and Civil Rights: What Went Wrong, 1985 U. Ill. L. Rev. 785, 802, n. 57 (noting the Reagan Administration’s touting of ‘“special magnet schools’ ” as a means of improving education for all children without “ ‘forced transportation’ ”). The absence of magnet schools in the Oklahoma City desegregation plan suggests much untapped potential for changing attitudes towards schools in the system.
The majority does not discuss the burden of proof under its test for dissolution of a school desegregation decree. However, every presumption we have established in our school desegregation cases has been against the school district found to have engaged in de jure segregation. See Dayton Bd. of Education v. Brinkman, 443 U. S. 526, 537 (1979) (conduct resulting in increased segregation was presumed to be caused by past intentional discrimination where dual system was never affirmatively remedied); Keyes v. School Dist. No. 1, Denver, Colo., 413 U. S. 189, 208 (1973) (proof of state-imposed segregation in a substantial portion of a school district will support a prima facie finding of a systemwide violation, thereby shifting the burden to school authorities to show that current segregation is not caused by past intentional discrimination); Swann v. Charlotte-Mecklenburg Bd. of Education, 402 U. S. 1, 26 (1971) (establishing a presumption against racially identifiable schools once past state discrimination has been shown, thereby shifting the burden to the school district to show that current segregation was not caused by past intentional discrimination). Moreover, in addition to the “affirmative duty” placed upon school districts to eliminate vestiges of their past discrimination, Green, 391 U. S., at 437-438, school districts initially have the burden of coming forward with desegregation plans and establishing that such plans promise to be effective. Id., at 439. And, while operating under a decree, a school board has a “heavy burden” to justify use of less effective or resegregative methods. Ibid. Accord, Dayton, supra, at 538; Wright v. Council of City of Emporia, 407 U. S. 451, 467 (1972).
Given the original obligation placed on formerly de jure segregated school districts to provide an effective remedy that will eliminate all vestiges of its segregated past, a school district seeking dissolution of an in-junctive decree should also bear the burden of proving that this obligation *267has been fulfilled. Cf. Keyes, supra, at 211, n. 17 (noting that the plaintiffs should not bear the burden of proving “non-attenuation”).
That “judicial tutelage” over the Oklahoma City School Board subsists at this late date is largely due to the Board’s failure to take advantage of opportunities it had at its disposal at the outset. It could have abolished and located new schools with a view toward promoting integration and shaping (rather than following) public attitudes toward its schools. See supra, at 254. It could have come forward with its own meaningful desegregation plan — a plan that would have been tailored to its particular concerns, including minimizing busing. Ibid. A school district’s failures in this regard, however, should not lead federal courts, charged with assuring that constitutional violations are fully remedied, to renounce supervision of unfinished tasks because of the lateness of the hour.
The concepts of temporariness and permanence have no direct relevance to courts’ powers in this context because the continued need for a decree will turn on whether the underlying purpose of the decree has been achieved. “The injunction ... is ‘permanent’ only for the temporary period for which it may last. It is justified only by the violence that induced it and only so long as it counteracts a continuing intimidation. Familiar equity procedure assures opportunity for modifying or vacating an injunction when its continuance is no longer warranted.” Milk Wagon Drivers v. Meadowmoor Dairies, Inc., 312 U. S. 287, 298 (1941).
Research indicates that public schools with high concentrations of poor and minority students have less access to experienced, successful teachers and that the slow pace of instruction at such schools may be “hindering] students’ academic progress, net of their own aptitude levels.” See Gamoran, Resource Allocation and the Effects of Schooling: A Sociological Perspective, in Microlevel School Finance: Issues and Implications for Policy 207, 214 (D. Monk & J. Underwood eds. 1988).