We review and affirm an order approving the 1988-89 Student Assignment Plan for the San Jose Unified School District.

An en banc panel of this court ruled that the school district acted with segregative intent in maintaining racially imbalanced schools, and ordered the district court to fashion an appropriate remedy. Diaz v. San Jose Unified School Dist., 733 F.2d 660 (9th Cir.1984), cert. denied, 471 U.S. 1065 , 105 S.Ct. 2140 , 85 L.Ed.2d 497 (1985). The district court, after receiving proposals from plaintiffs, the class of all Spanish-sur-named students enrolled in the San Jose district and their parents, and the defendant school district, drafted an elaborate desegregation order. 633 F.Supp. 808 (N.D.Cal.1985). It set interim annual goals, projecting that 90% of all students would attend desegregated schools by 1990.

The court adopted an approach designed “to maximize the ability of a student and his or her parents to choose voluntarily which school the student will attend.” Id. at 815 . To facilitate voluntary student movement to desegregated schools, it ordered the district to establish (1) district-wide magnet schools; (2) schools with specialty enrichment programs; and (3) “schools with certain ‘programs of excellence’ in which a student’s attendance would improve the racial balance of both the sending and receiving school.” Id.

The district expected almost complete desegregation by 1990 under the voluntary plan, but the court included a mandatory “backup” mechanism. If voluntary choices were not leading to sufficient desegregation, the district was authorized to impose caps to limit the enrollment of new students of particular ethnicity at racially-isolated schools.

To implement its desegregation plan, the court established a student assignment procedure that relied on an annual registration program. All students were subject initially to the program, and thereafter only those new to the district and those beginning the elementary, middle or high school level.

The district was to inform students and parents of educational and other opportunities offered at each school, and they were to respond with a ranking of school preference. Students were to be assigned to their first choices unless the school’s physical or programmatic capacity would be exceeded, or enrollment of certain students was barred by an ethnic enrollment cap. The court provided assignment priorities for schools that were over-subscribed. Students at the same priority level were assigned by random drawing.

The district prepared a student assignment plan for the 1988-89 school year. It applied to students who were required to participate in the registration process (students changing school levels, incoming students, and students on waiting lists). The plan was divided into two phases. During Phase I the district collected student registration requests through March 1988. From that pool, it filled schools according to the students’ choices pursuant to court-ordered priorities and up to the facility or ethnic capacities. Ninety-three percent of registering students were assigned to their first choices.

Phase II covered students who registered after April 1, 1988. Students who enrolled during this phase could not be considered for assignment to schools that had reached their facility or ethnic capacity by the end of Phase I. Students who did not receive their first choices were placed on waiting lists and would be considered again during the next registration period. By the end of Phase II, the district expects that 92% of its students will attend desegregated schools, and 35 of its 38 schools will be desegregated.

In fulfilling its supervisory role over the remedial order, the district court held a hearing in March 1988 to review the plan and other aspects of its desegregation order. Plaintiffs challenged the assignment plan. They argued that Phase I does little to achieve desegregation and that only by imposing caps in Phase II is the district able to desegregate schools. They complained that because 72% of the participants in Phase II are minority students, they shoulder the burden of desegregation disproportionately. They sought to have the district’s plan rejected and their various proposals implemented.

After the parties submitted briefs and the one day hearing, the court approved the 1988-89 assignment plan. It explained:

Plaintiffs appeal. They ask that the order be reversed and that their proposed plan be implemented or, in the alternative, that the case be remanded for further consideration and an evidentiary hearing. The district moved to dismiss the appeal as untimely. We raised sua sponte the question of statutory jurisdiction.

ANALYSIS

I. Jurisdiction

A. Timeliness

The district argues that plaintiffs present an untimely appeal of the 1985 remedial order. Plaintiffs did not appeal that order.

To the extent that plaintiffs seek to attack the 1985 order, their appeal is untimely. Some of their arguments, however, challenge only the validity of the 1988-89 assignment plan. Those issues are timely.

B. Statutory Jurisdiction

The question of statutory jurisdiction is controlled by our decision in United States v. State of Washington, 761 F.2d 1404 (9th Cir.1985), cert. denied sub nom. 474 U.S. 1100 , 106 S.Ct. 879 , 88 L.Ed.2d 916 (1986). In Washington, the district court pursuant to its continuing jurisdiction in the Northwest Indian Fisheries litigation issued an order that adopted recommendations of the Fishery Advisory Board regarding allocation of fish for the 1983 fall season. We ruled that the order was sufficiently final for jurisdiction under 28 U.S.C. § 1291 because, as a post-judgment order, there was little danger of piecemeal review and because immediate review provided the only opportunity for meaningful review. Id. at 1407 .

Our decision in Washington recognized that “the inconvenience and costs of piecemeal review on the one hand and the danger of denying justice by delay on the other” are the most important competing considerations when deciding questions of finality. See Gillespie v. United States Steel Corp., 379 U.S. 148, 152-53 , 85 S.Ct. 308, 310-11 , 13 L.Ed.2d 199 (1964) (quoting Dickinson v. Petroleum Conversion Corp., 338 U.S. 507, 511 , 70 S.Ct. 322, 324 , 94 L.Ed. 299 (1950)). The order here, like that in Washington, was entered after the underlying dispute had been decided, and was supervisory in nature. Concern about piecemeal review is not as significant under these circumstances. Washington, 761 F.2d at 1406 . Plaintiffs allege immediate adverse effects as a result of the assignment plan. Although we might effectively review at a later date a general claim of disproportionate burden, we could never review the merits of the assignment plan because it will cease to exist at the end of this school year. In effect, delay in review would deny relief to those plaintiffs who are currently subject to the plan.

We are ever mindful of the policy against obstructing or impeding an ongoing judicial proceeding by interlocutory appeal. United States v. Nixon, 418 U.S. 683, 690 , 94 S.Ct. 3090, 3098 , 41 L.Ed.2d 1039 (1973). Orders issued by a district court pursuant to its supervisory jurisdiction over a remedial desegregation order present an opportunity for improper appellate interference. But the facts here are virtually indistinguishable from those in Washington, and we conclude that the order must be treated as an appealable final order. Defendants' motion to dismiss this appeal for lack of jurisdiction is denied.

II. Review of the Order

A. Standard of Review

This circuit has not articulated previously the standard for reviewing a school desegregation remedy. Desegregation orders are wholly equitable in nature. Brown v. Bd. of Educ., 349 U.S. 294, 300 , 75 S.Ct. 753, 756 , 99 L.Ed. 1083 (1955). The abuse of discretion standard that governs our review of equitable remedies is equally applicable here. Cf. SEC v. Hardy, 803 F.2d 1034 , 1038 (9th Cir.1986) (supervision of equitable receivership). That review is especially deferential because district courts have wide latitude in fashioning desegregation remedies. See, e.g., Columbus Bd. of Educ. v. Penick, 443 U.S. 449, 471 , 99 S.Ct. 2941, 2952 , 61 L.Ed.2d 666 (1979) (Stewart, J., concurring and dissenting); Lee v. Anniston City School System, 737 F.2d 952 (11th Cir.1984).

B. Compliance with the Prior Remedial Order

Plaintiffs argue that the assignment plan is inconsistent with the court’s remedial order in that: (1) the court did not authorize a dual assignment process (Phase I and Phase II); (2) it did not authorize inconsistent use of caps; and (3) it did not authorize a freedom of choice plan at Phase I.

The district’s use of bifurcated registration process and more extensive use of caps during Phase II does not conflict with the 1985 order. It stated specifically:

633 F.Supp. at 818

The court did not need to authorize a dual assignment process or inconsistent use of caps. It had already placed those decisions within the district’s discretion. Despite plaintiffs’ “freedom of choice” label, the assignment plan is not inconsistent with the 1985 order.

C.Green v. County School Bd. of New Kent County

Plaintiffs argue that the assignment plan violates the mandate of Green v. County School Bd. of New Kent County, 391 U.S. 430 , 88 S.Ct. 1689 , 20 L.Ed.2d 716 (1968). They argue that the district is operating a freedom of choice plan, contrary to Green , and that the district court failed to assess adequately the effectiveness of the plan.

The plaintiffs in Green attacked a court-ordered desegregation plan that allowed students to choose between the district’s two schools. The district argued that it had fully discharged its desegregation obligation “by adopting a plan by which every student, regardless of race, may 'freely’ choose the school he will attend.” Id. at 437 , 88 S.Ct. at 1693 .

The Court rejected the plan as a sufficient step to effectuate desegregation. After three years of operation, not a single white student had chosen to attend the all-black school, and the number of blacks enrolled in the predominantly white school was not sufficient to avoid the conclusion that the school system remained a dual system. Merely giving students freedom to choose their school was not an adequate remedy.

The Court tempered its ruling, somewhat:

391 U.S. at 439-40 , 88 S.Ct. at 1694-95 .

Although plaintiffs characterize Phase I as a freedom of choice plan, it differs sig nificantly from the inadequate plan of Green . The desegregation remedy here is based on voluntary school selection, but the overall plan, including Phase I, includes controls to ensure that it is effective. Unlike Green , the program here “relies heavily on the use of magnet schools and specialty enrichment programs to encourage voluntary transfers for the purpose of desegregation.” 633 F.Supp. at 812 . And unlike Green , if the voluntary selection process does not result in attainment of the interim desegregation goals, the district is required to employ ethnic enrollment caps to further desegregation. The plan here reflects Green’s admonition that “ ‘freedom of choice’ is not an end in itself,” but that it may have a proper place in a desegregation plan.

The Green Court reiterated the district court’s duties to oversee desegregation plans:

Id. at 439 , 88 S.Ct. at 1696 .

Plaintiffs complain that the desegregation plan does not comply with this mandate: “[T]wo and [a] half years after the court’s remedial order, 26% of minorities [sic] elementary students are still in segregated schools, 6 schools predominantly minority are still segregated, schools in the north primarily minority are under capacity, and schools in the south, primarily majority are over capacity, and two out of every three students eligible for transportation is a minority student.”

This argument addresses overall effectiveness of the court's remedial order. Plaintiffs did not appeal that order, and we would be hesitant to fault it now when the district has exceeded consistently the interim desegregation goals. Moreover, the district court, in formulating the 1985 order, considered proposals from the district and plaintiffs. It acknowledged its duty under Green and did not adopt the district’s proposals in full. 633 F.Supp. at 812 .

D. Disproportionate Burdens

Plaintiffs argue that the proposed student assignment plan is inequitable and that the burden to desegregate disproportionately affects the victims of de jure segregation. They complain that Phase II students, predominantly minority, are bused mandatorily in order to desegregate the southern majority-populated schools.

This circuit has not addressed previously a claim of inequitable burden. Others, however, have articulated several principles that are helpful to our analysis.

Charges that the burdens of desegregation have been distributed inequitably must be scrutinized closely. See Higgins v. Bd. of Educ. of City of Grand Rapids, 508 F.2d 779, 793 (6th Cir.1974); Parent Assoc. of Andrew Jackson High School v. Ambach, 598 F.2d 705, 717 (2d Cir.1979) (most exacting form of review required). District courts must ensure that the burdens of desegregation are distributed equally and without discrimination. See Higgins, 508 F.2d at 793 ; Arvizu v. Waco Independent School Dist., 495 F.2d 499, 504 (5th Cir.1974).

Whether a singular group is burdened impermissibly turns on “the validity of the [bjoard’s justifications for its proposals and the availability of feasible alternatives” to the objectionable measures. Id. at 504 ; see Keyes v. School Dist. No. 1, Denver, Colorado, 521 F.2d 465, 479 (10th Cir.1975), cert. denied, 423 U.S. 1066 , 96 S.Ct. 806 , 46 L.Ed.2d 657 (1976).

We need not demarcate permissible from impermissible burdens today because plaintiffs have failed to establish that they are burdened disproportionately. They contend that at least 70% of Phase II students are minorities, and that a great majority of them are burdened by Phase II.

The 1988-89 plan had not been implemented fully at the time of the hearing, but the parties agree that it does not deviate materially from prior years’ plans. Among the group of students registering during the 1987-88 Phase II, nearly, equal proportions of majority and minority students were not assigned to their first choice schools. Overall, however, 98% of majority students got their first choices, while 94% of minority students received theirs. Plaintiffs have not established that assignments to other than first choice schools affect them disproportionately. Nor have they offered any evidence that they are subject to mandatory busing at significantly higher rates than majority students. Their claim of disproportionate burden cannot stand solely on the assertion that a majority of the students registering during Phase II are minorities.

E. Evidentiary Hearing

Plaintiffs contend that the court erred by denying their request to present evidence. Denial of an evidentiary hearing is reviewed for abuse of discretion. See United States v. Dicesare, 765 F.2d 890, 895 (9th Cir.1985). An evidentiary hearing is required when the litigants offer proof “sufficiently definite, specific, detailed, and nonconjectural to enable the court to conclude” that a contested issue of fact is in question. Id. at 896 .

The court acknowledged the plaintiffs’ request for a hearing, and commented that they had made some recommendations and comments but failed to file any declarations or affidavits supporting their position. They have failed to identify any offer of proof at the hearing that would take their assertions from the realm of mere conjecture. The court did not abuse its discretion.

III. Attorneys’ Fees and Costs

Plaintiffs seek attorneys’ fees pursuant to 42 U.S.C. § 1988 . The district suggests that plaintiffs should be sanctioned under Fed.R.App.P. 38.

Courts have discretion to award reasonable attorneys’ fees to the prevailing party in a section 1983 action. 42 U.S.C. § 1988 . “A ‘prevailing party’ under section 1988 means ‘a party [who] has prevailed on the merits of at least some of his claims.” Jensen v. City of San Jose, 806 F.2d 899, 900 (9th Cir.1986) (en banc) (quoting Hanrahan v. Hampton, 446 U.S. 754, 758 , 100 S.Ct. 1987, 1989 , 64 L.Ed.2d 670 (1980)). To be more specific, parties prevail “ ‘if they succeed on any significant issue in litigation which achieves some of the benefit the parties sought in bringing the suit.’ ” Id. at 900 (quoting Hensley v. Eckerhart, 461 U.S. 424, 433 , 103 S.Ct. 1933, 1939 , 76 L.Ed.2d 40 (1983)).

Plaintiffs sought to have the 1988-89 plan disapproved. They did not prevail. Nor are sanctions warranted here.

The order is AFFIRMED.