BUTZNER, Circuit Judge:
The Charlotte-Mecklenburg School District appealed from an order of the district court requiring the faculty and student body of every school in the system to be racially mixed. We approve the provisions of the order dealing with the faculties of all schools1 and the assignment of pupils to high schools and junior high schools, but we vacate the order and remand the case for further consideration of the assignment of pupils attending elementary schools. We recognize, of course, that a change in the elementary schools may require some modification of the junior and senior high school plans, and our remand is not intended to preclude this.
I.
The Charlotte-Mecklenburg school system serves a population of over 600,000 people in a combined city and county area of 550 square miles. With 84,500 pupils attending 106 schools, it ranks as the nation’s 43rd largest school district. In Swann v. Charlotte-Mecklenburg Bd. of Ed., 369 F.2d 29 (4th Cir. 1966), we approved a desegregation plan based on geographic zoning with a free transfer provision. However, this plan did not eliminate the dual system of schools. The district court found that during the 1969-70 school year, some 16,000 black pupils, out of a total of 24,700, were attending 25 predominantly black schools, that faculties had not been integrated, and that other administrative practices, including a free transfer plan, tended to perpetuate segregation.
[141]*141Notwithstanding- our 1965 approval of the school board’s plan, the district court properly held that the board was impermissibly operating a dual system of schools in the light of subsequent decisions of the Supreme Court, Green v. County School Bd. of New Kent County, 391 U.S. 430, 435, 88 S.Ct. 1689, 20 L.Ed.2d 716 (1968), Monroe v. Bd. of Com’rs, 391 U.S. 450, 88 S.Ct. 1700, 20 L.Ed.2d 733 (1968), and Alexander v. Holmes County Bd. of Ed., 396 U.S. 19, 90 S.Ct. 29, 24 L.Ed.2d 19 (1969).
The district judge also found that residential patterns leading to segregation in the schools resulted in part from federal, state, and local governmental action. These findings are supported by the evidence and we accept them under familiar principles of appellate review. The district judge pointed out that black residences are concentrated in the northwest quadrant of Charlotte as a result of both public and private action. North Carolina courts, in common with many courts elsewhere, enforced racially restrictive covenants on real property 2 until Shelley v. Kraemer, 334 U.S. 1, 68 S.Ct. 836, 92 L.Ed. 1161 (1948), prohibited this discriminatory practice. Presently the city zoning ordinances differentiate between black and white residential areas. Zones for black areas permit dense occupancy, while most white areas are zoned for restricted land usage. The district judge also found that urban renewal projects, supported by heavy federal financing and the active participation of local government, contributed to the city’s racially segregated housing patterns. The school board, for its part, located schools in black residential areas and fixed the size of the schools to accommodate the needs of immediate neighborhoods. Predominantly black schools were the inevitable result. The interplay of these policies on both residential and educational segregation previously has been recognized by this and other courts.3 The fact that similar forces operate in cities throughout the nation. under the mask of de facto segregation provides no justification for allowing us to ignore the part that government plays in creating segregated neighborhood schools.
The disparity in the number of black and white pupils the Charlotte-Mecklen-burg School Board busses to predominantly black and white schools illustrates how coupling residential patterns with the location of schools creates segregated schools. All pupils are eligible to ride school buses if they live farther than 1% miles from the schools to which they are assigned. Overall statistics show that about one-half of the pupils entitled to transportation ride school buses. Only 541 pupils were bussed in October 1969 to predominantly black schools, which had a total enrollment of over 17,000. In contrast, 8 schools located outside the black residential area have in the aggregate only 96 students living within 1 y2 miles. These schools have a total enrollment of about 12,184 pupils, of whom 5,349 ride school buses.
[142]*142II.
The school board on its own initiative, or at the direction of the district court, undertook or proposed a number of reforms in an effort to create a unitary school system. It closed 7 schools and reassigned the pupils primarily to increase racial mixing. It drastically gerrymandered school zones to promote desegregation. It created a single athletic league without distinction between white and black schools or athletes, and at its urging, black and white PTA councils were merged into a single organization. It eliminated a school bus system that operated on a racial basis, and established nondiscriminatory practices in other facets of the school system. It modified its free transfer plan to prevent resegregration, and it provided for integration of the faculty and administrative staff.
The district court, after a painstaking analysis of the board’s proposals and the relevant authorities, disapproved the board’s final plan, primarily because it left ten schools nearly all black. In reaching this decision, the district court held that the board must integrate the student body of every school to convert from a dual system of schools, which had been established by state action, to a unitary system.
The necessity of dealing with segregation that exists because governmental policies foster segregated neighborhood schools is not confined to the Charlotte-Mecklenburg School District. Similar segregation occurs in many other cities throughout the nation, and constitutional principles dealing with it should be applied nationally. The solution is not free from difficulty. It is now well settled that school boards operating dual systems have an affirmative duty “to convert to a unitary school system in which racial discrimination would be eliminated root and branch.” Green v. School Bd. of New Kent County, 391 U.S. 430, 437, 88 S.Ct. 1689, 1694, 20 L.Ed.2d 716 (1968). Recently the Supreme Court defined a unitary school system as one “within which no person is to be effectively excluded from any school because of race or color.” Alexander v. Holmes County Bd. of Ed., 396 U.S. 19, 20, 90 S.Ct. 29, 30, 24 L.Ed.2d 19 (1969). This definition, as the Chief Justice noted in Northcross v. Board of Ed. of Memphis, 397 U.S. 232, 90 S.Ct. 891, 893, 25 L.Ed.2d 246 (1970), leaves open practical problems, “including whether, as a constitutional matter, any particular racial balance must be achieved in the schools; to what extent school districts and zones may or must be altered as a constitutional matter; to what extent transportation may or must be provided to achieve the ends sought by prior holdings of the Court.”
Several of these issues arise in this case. To resolve them, we hold: first, that not every school in a unitary school system need be integrated; second, nevertheless, school boards must use all reasonable means to integrate the schools in their jurisdiction; and third, if black residential areas are so large that not all schools can be integrated by using reasonable means, school boards must take further steps to assure that pupils are not excluded from integrated schools on the basis of race. Special classes, functions, and programs on an integrated basis should be made available to pupils in the black schools. The board should freely allow majority to minority transfers and provide transportation by bus or common carrier so individual students can leave the black schools. And pupils who are assigned to black schools for a portion of their school careers should be assigned to integrated schools as they progress from one school to another.
We adopted the test of reasonableness — instead of one that calls for absolutes — because it has proved to be a reliable guide in other areas of the [143]*143law. Furthermore, the standard of reason provides a test for unitary school systems that can be used in both rural and metropolitan districts. All schools in towns, small cities, and rural areas generally can be integrated by pairing, zoning, clustering, or consolidating schools and transporting pupils. Some cities, in contrast, have black ghettos so large that integration of every school-is an improbable, if not an unattainable, goal. Nevertheless, if a school board makes every reasonable effort to integrate the pupils under its control, an intractable remnant of segregation, we believe, should not void an otherwise exemplary plan for the creation of a unitary school system. Ellis v. Board of Public Instruc. of Orange County, 423 F.2d 203 (5th Cir. 1970).
III.
The school board’s plan proposes that pupils will be assigned to the system’s ten high schools according to geographic zones. A typical zone is generally fan shaped and extends from the center of the city to the suburban and rural areas of the county. In this manner the board was able to integrate nine of the high schools with a percentage of black students ranging from 17% to 36%. The projected black attendance at the tenth school, Independence, which has a maximum capacity of 1400 pupils, is 2%.
The court approved the board’s high school plan with one modification. It required that an additional 300 pupils should be transported from the black residential area of the city to Independence School.
The school board proposed to rezone the 21 Junior high school areas so that black attendance would range from 0% to 90% with only one school in excess of 38%. This school, Piedmont, in the heart of the black residential area, has an enrollment of 840 pupils, 90% of whom are black. The district court disapproved the board’s plan because it maintained Piedmont as a predominantly black school. The court gave the board four options to desegregate all the junior high schools: (1) rezoning; (2) two-way transportation of pupils between Piedmont and white schools; (3) closing Piedmont and reassigning its pupils; and (4) adopting a plan proposed by Dr. John A. Finger, Jr., a consultant appointed by the court, which combined zoning with satellite districts. The board, expressing a preference for its own plan, reluctantly adopted the plan proposed by the court’s consultant.
Approximately 31,000 white and 13,-000 black pupils are enrolled in 76 elementary schools. The board’s plan for desegregating these schools is based entirely upon geographic zoning. Its proposal left more than half the black elementary pupils in nine schools that remained 86% to 100% black, and assigned about half of the white elementary pupils to schools that are 86% to 100% white. In place of the board’s plan, the court approved a plan based on zoning, pairing, and grouping, devised by Dr. Finger, that resulted in student bodies that ranged from 9% to 38% black.
The court estimated that the overall plan which It approved would require this additional transportation:
No. of pupils No. of buses Operating cost
Senior High 1.500 20 $ 30,000
Junior High 2.500 28 $ 50,000
Elementary 9,300 90 $186,000
TOTAL 13,300 138 $266,000
In addition, the court found that a new bus cost about $5,400, making a total outlay for equipment of $745,-200. The total expenditure for the first year would be about $1,011,200.
The school board computed the additional transportation requirements under the court approved plan to be:
No. of No. of Operating pupils buses cost
Senior High 2,497 69 $ 96,000
Junior High 4,359 84 $116,800
Elementary 12,429 269 $374,000
TOTAL 19,285 422 $586,800
[144]*144In addition to the annual operating cost, the school board projected the following expenditures:
Cost of buses $2,369,100
Cost of parking areas 284,800
Cost of additional personnel 166,200
Based on these figures, the school board computed the total expenditures for the first year would be $3,406,-700 under the court approved plan.4
Both the findings of the district court and the evidence submitted by the board are based on estimates that rest on many variables. Past practice has shown that a large percentage of students eligible for bus transportation prefer to provide their own transportation. However, it is difficult to accurately predict how many eligible students will accept transportation on the new routes and schedules. The number of students that a bus can carry each day depends in part on the number of trips the bus can make. Scheduling two trips for a bus generally reduces costs. But student drivers may not be able to spend the time required for two trips, so that adult drivers will have to be hired at substantially higher salaries. It is difficult to accurately forecast how traffic delays will affect the time needed for each trip, for large numbers of school buses themselves generate traffic problems that only experience can measure.
The board based its projections on each 54-passenger bus carrying about 40 high school pupils or 54 junior high and elementary pupils for one roundtrip a day. Using this formula, it arrived at a need of 422 additional buses for transporting 19,285 additional pupils. This appears to be a less efficient operation than the present system which transports 23,600 pupils with 280 buses, but the board’s witnesses suggest that prospects of heavier traffic justify the difference. The board also envisioned parking that seems to be more elaborate than that currently used at some schools.
In making its findings, the district court applied factors derived from present bus operation, such as the annual operating cost per student, the average number of trips each bus makes, the capacity of the buses — including permissible overloads, and the percentage of eligible pupils who use other forms of transportation. The district court also found no need for expensive parking facilities or for additional personnel whose costs could not be absorbed by the amount allocated for operating expenses. While we recognize that no estimate— whether submitted by the board or made by the court — can be absolutely correct, we accept as not clearly erroneous the findings of the district court.
Opposition to the assignment of pupils under both the board’s plan and the plan the court approved centered on bussing, [145]*145which numbers among its critics both black and white parents. This criticism, however, cannot justify the maintenance of a dual system of schools. Cooper v. Aaron, 358 U.S. 1, 78 S.Ct. 1401, 3 L.Ed.2d 5 (1958). Bussing is neither new nor unusual. It has been used for years to transport pupils to consolidated schools in both racially dual and unitary school systems. Figures compiled by the National Education Association show that nationally the number of pupils bussed increased from 12 million in the 1958-59 school year to 17 million a decade later. In North Carolina 54.9% of all pupils are bussed. There the average daily roundtrip is 24 miles, and the annual cost is over $14,000,000. The Charlotte-Mecklenburg School District presently busses about 23,600 pupils and another 5,000 ride common carriers.
Bussing is a permissible tool for achieving integration, but it is not a panacea. In determining who should be bussed and where they should be bussed, a school board should take into consideration the age of the pupils, the distance and time required for transportation, the effect on traffic, and the cost in relation to the board’s resources. The board should view bussing for integration in the light that it views bussing for other legitimate improvements, such as school consolidation and the location of new schools. In short, the board should draw on its experience with bussing in general —the benefits and the defects — so that it may intelligently plan the part that bussing will play in a unitary school system.
Viewing the plan the district court approved for junior and senior high schools against these principles and the background of national, state, and local transportation policies, we conclude that it provides a reasonable way of eliminating all segregation in these schools. The estimated increase in the number of junior and senior high school students who must be bussed is about 17 % of all pupils now being bussed. The additional pupils are in the upper grades and for the most part they will be going to schools already served by buses from other sections of the district. Moreover, the routes they must travel do not vary appreciably in length from the average route of the system’s buses. The transportation of 300 high school students from the * black residential area to suburban Independence School will tend to stabilize the system by eliminating an almost totally white school in a zone to which other whites might move with consequent “tipping” or resegregation of other schools.5
We find no merit in other criticism of the plan for junior and senior high schools. The use of satellite school zones 6 ***10as a means of achieving desegregation is not improper. District courts have been directed to shape remedies that are characterized by the “practical flexibility” that is a hallmark of equity. See Brown v. Board of Ed., 349 U.S. [146]*146294, 300, 75 S.Ct. 753, 99 L.Ed. 1083 (1955). Similarly, the pairing and clustering of schools has been approved. Green v. School Bd. of New Kent County, 391 U.S. 430, 442 n. 6, 88 S.Ct. 1689, 20 L.Ed.2d 716 (1968); Hall v. St. Helena Parish School Bd., 417 F.2d 801, 809 (5th Cir.), cert. denied, 396 U.S. 904, 90 S.Ct. 218, 24 L.Ed.2d 180 (1969).
The school board also asserts that §§ 401(b) and 407(a) (2) of the Civil Rights Act of 1964 [42 U.S.C. §§ 2000c(b) and 2000c-6(a) (2)] forbid the bussing ordered by the district court.7 But this argument misreads the legislative history of the statute. Those provisions are not limitations on the power of school boards or courts to remedy unconstitutional segregation. They were designed to remove any implication that the Civil Rights Act conferred new jurisdiction on courts to deal with the question of whether school boards were obligated to overcome de facto segregation. See generally, United States v. School District 151, 404 F.2d 1125, 1130 (7th Cir. 1968); United States v. Jefferson County Board of Ed., 372 F.2d 836, 880 (5th Cir. 1966), aff’d on rehearing en banc 380 F.2d 385 (5th Cir.), cert. denied, sub nom. Caddo Parish School Bd. v. United States, 389 U.S. 840, 88 S.Ct. 67, 19 L.Ed.2d 103 (1967); Keyes v. School Dist. No. One, Denver, 303 F.Supp. 289, 298 (D.Colo.), stay pending appeal granted (10th Cir.), stay vacated, 396 U.S. 1215, 90 S.Ct. 12, 24 L.Ed.2d 37 (1969). Nor does North Carolina’s anti-bussing law present an obstacle to the plan, for those provisions of the statute in conflict with the plan have been declared unconstitutional. Swann v. Charlotte-Mecklenburg Bd. of Ed., 311 F.Supp. 265 (W.D.N.C.1970).8
The district court properly disapproved the school board’s elementary school proposal because it left about one-half of both the black and white elementary pupils in schools that were nearly completely segregated. Part of the difficulty concerning the elementary schools results from the board’s refusal to accept the district court’s suggestion that it consult experts from the Department of Health, Education, and Welfare. The consultants that the board employed were undoubtedly competent, but the board limited their choice of remedies by maintaining each school’s grade structure. This, in effect, restricted the means of overcoming segregation to only geographical zoning, and as a further restriction the board insisted on contiguous zones. The board rejected such legitimate techniques as pairing, grouping, clustering, and satellite zoning. Moreover, the board sought to impose a ratio in each school of not less than 60% white students. While a 60%-[147]*14740% ratio of white to black pupils might be desirable under some circumstances, rigid adherence to this formula in every school should not be allowed to defeat integration.
On the other hand, the Finger plan, which the district court approved, will require transporting 9,300 pupils in 90 additional buses. The greatest portion of the proposed transportation involves cross-bussing to paired schools — that is, black pupils in grades one through four would be carried to predominantly white schools, and white pupils in the fifth and sixth grades would be transported to the black schools. The average daily round-trip approximates 15 miles through central city and suburban traffic.
The additional elementary pupils who must be bussed represent an increase of 39% over all pupils presently being bussed, and their transportation will require an increase of about 32% in the present fleet of buses. When the additional bussing for elementary pupils is coupled with the additional requirements for junior and senior high schools, which we have approved, the total percentages of increase are: pupils, 56%, and buses, 49%. The board, we believe, should not be required to undertake such extensive additional bussing to dis-. charge its obligation to create a imitary school system.
IV.
Both parties oppose a remand. Each side is adamant that its position is correct — the school board seeks total approval of its plan and the plaintiffs insist on implementation of the Finger plan. We are favorably impressed, however, by the suggestion of the United States, which at our invitation filed a brief as amicus curiae, that the school board should consider alternative plans, particularly for the elementary schools. We, therefore, will vacate the judgment of the district court and remand the case for reconsideration of the assignment of pupils in the elementary schools, and for adjustments, if any, that this may require in plans for the junior and senior high schools.
On remand, we suggest that the district court should direct the school board to consult experts from the Office of Education of the Department of Health, Education, and Welfare, and to explore every method of desegregation, including rezoning with or without satellites, pairing, grouping, and school consolidation. Undoubtedly some transportation will be necessary to supplement these techniques. Indeed, the school board’s plan proposed transporting 2,300 elementary pupils, and our remand should not be interpreted to prohibit all bus-sing. Furthermore, in devising a new plan, the board should not perpetuate segregation by rigid adherence to the 60% white-40 % black racial ratio it favors.
If, despite all reasonable efforts to integrate every school, some remain segregated because of residential patterns, the school board must take further steps along the lines we previously mentioned, including a majority to minority transfer plan,9 to assure that [148]*148no pupil is excluded from an integrated school on the basis of race.
Alexander v. Holmes County Bd. of Ed., 396 U.S. 19, 90 S.Ct. 29, 24 L.Ed.2d 19 (1969), and Carter v. West Feliciana School Bd., 396 U.S. 290, 90 S.Ct. 608, 24 L.Ed.2d 477 (1970), emphasize that school boards must forthwith convert from dual to unitary systems. In Nes-bit v. Statesville City Bd. of Ed., 418 F.2d 1040 (4th Cir. 1969), and Whitten-berg v. School Dist. of Greenville County, 424 F.2d 195 (4th Cir. 1970), we reiterated that immediate reform is imperative. We adhere to these principles, and district courts in this circuit should not consider the stays which were allowed because of the exceptional nature of this ease to be precedent for departing from the directions stated in Alexander, Carter, Nesbit, and Whittenberg.
Prompt action is also essential for the solution of the remaining difficulties in this case. The school board should immediately consult with experts from HEW and file its new plan by June 30, 1970. The plaintiffs should file their exceptions, if any, within 7 days, and the district court should promptly conduct all necessary hearings so that the plan may take effect with the opening of school next fall. Since time is pressing, the district court’s order approving a new plan shall remain in full force and effect unless it is modified by an order of this court. After a plan has been approved, the district court may hear additional objections or proposed amendments, but the parties shall comply with the approved plan in all respects while the district court considers the suggested modifications. Cf. Nesbit v. Statesville City Bd. of Ed., 418 F.2d 1040, 1043 (4th Cir. 1969).
Finally, we approve the district court’s inclusion of Dr. Finger’s consultant fee in the costs taxed against the board. See In the Matter of Peterson, 253 U.S. 300, 312, 40 S.Ct. 543, 64 L.Ed. 919 (1920). We caution, however, that when a court needs an expert, it should avoid appointing a person who has appeared as a witness for one of the parties. But the evidence discloses that Dr. Finger was well qualified, and his dual role did not cause him to be faithless to the trust the court imposed on him. Therefore, the error, if any, in his selection, was harmless.
We find no merit in the other objections raised by the appellants or in the appellees’ motion to dismiss the appeal. The judgment of the district court is vacated, and the case is remanded for further proceedings consistent with this opinion.
SOBELOFF, Circuit Judge, with whom WINTER, Circuit Judge, joins, concurring in part and dissenting in part:
Insofar as the court today affirms the District Court's order in respect to the senior and junior high schools, I concur. I dissent from the failure to affirm the portion of the order pertaining to the elementary schools.
I
THE BASIC LAW AND THE PARTICULAR FACTS
All uncertainty about the constitutional mandate of Brown v. Board of Education, 347 U.S. 483, 74 S.Ct. 686, 98 L.Ed. 873 (1954) and 349 U.S. 294, 75 S.Ct. 753, 99 L.Ed. 1083 (1955), was put to rest when in Green v. County School Board of New Kent County the Supreme Court spelled out a school board’s “affirmative duty to take whatever steps might be necessary to convert to a unitary system in which racial discrimination would be eliminated root and branch,” 391 U.S. 430, 437-438, 88 S.Ct. 1689, 1694, 20 L.Ed.2d 716 (1968). “Disestablish [ment of] state-imposed segregation” (at 439, 88 S.Ct. 1689, 1695) entailed “steps which promise realistically to convert promptly to a system without a ‘white’ school and a ‘Negro’ school, but just schools” (at 442, 88 S.Ct. at 1696). If there could still be doubts they were answered this past year. In Alexander v. Holmes County Board of [149]*149Education, the Court held that “[ujnder explicit holdings of this Court the obligation of every school district is to terminate dual school systems at once and to operate now and hereafter only unitary schools,” 396 U.S. 19, 20, 90 S.Ct. 29, 24 L.Ed.2d 19 (1969). The command was once more reaffirmed in Carter v. West Feliciana School Board, 396 U.S. 290, 90 S.Ct. 608, 24 L.Ed.2d 477 (1970), requiring “relief that will at once extirpate any lingering vestiges of a constitutionally prohibited dual school system.” (Harlan, J., concurring at 292, 90 S.Ct. at 609).
We face in this case a school district divided along racial lines. This is not a fortuity. It is the result, as the majority has recognized, of government fostered residential patterns, school planning, placement, and, as the District Court found, gerrymandering. These factors have interacted on each other so that by this date the black and white populations, in school and at home, are virtually entirely separate.
As of November 7, 1969, out of 106 schools in the system, 57 were racially identifiable as white, 25 were racially identifiable as black.1 Of these, nine were all white schools and eleven all black. Of 24,714 black students in the system, 16,000 were in entirely or predominantly black schools.
There are 76 elementary schools with over 44,000 pupils. In November 1969, 43 were identifiable as white, 16 as black, with 13 of the latter 98% or more black, and none less than 65%.. For the future the Board proposes little improvement. There would still be 25 identifiably white elementary schools and approximately half of the white elementary students would attend schools 86 to 100% white. Nine schools would remain 83 to 100% black, serving 6,432 students or over half the black elementary pupils.
To call either the past or the proposed distribution a “unitary system” would be to embrace an illusion.2 And the majority does not contend that the system is unitary, for it holds that “the district court properly disapproved the school board’s elementary school proposal because it left about one-half of both the black and white elementary pupils in schools that were nearly completely segregated.” The Board’s duty then is plain and unarguable: to convert to a unitary system. The duty is absolute. It is not to be tempered or watered down. It must be done, and done now.
II
THE COURT-ORDERED PLAN
A. The Necessity of the Court-Ordered Plan
The plan ordered by the District Court works. It does the job of desegregating the schools completely. This “places a heavy burden upon the board to explain its preference for an apparently less effective method.” Green, supra, 391 U.S. at 439, 88 S.Ct. at 1695.
The most significant fact about the District Court’s plan is that it — or one like it — is the only one that can work. Obviously, when the black students are all on one side of town, the whites on [150]*150the other, only transportation will bring them together. The District Judge is quite explicit:
Both Dr. Finger and the school board staff appear to have agreed, and the court finds as a fact, that for the present at least, there is no way to desegregate the all-black schools in Northwest Charlotte without providing (and continuing to provide) bus or other transportation for thousands of children. All plans and all variations of plans considered for this purpose lead in one fashion or another to that conclusion.
The point has been perceived by the counsel for the Board who have candidly informed us that if the job must be done then the Finger plan is the way to do it.
The only suggestion that there is a possible alternative middle course came from the United States, participating as amicus curiae. Its brief was prefaced by the following revealing confession:
We understand that the record in the case is voluminous, and we would note at the outset that we have been unable to analyze the record as a whole. Although we have carefully examined the district court’s various opinions and orders, the school board’s plan, and those pleadings readily available to us, we feel that we are not conversant with all of the factual considerations which may prove determinative of this appeal. Accordingly, we here attempt, not to deal extensively with factual matters, but rather to set forth some legal considerations which may be helpful to the Court.
Notwithstanding this disclaimer, the Government went on to imply in oral argument — and has apparently impressed on this court — that HEW could do better. No concrete solution is suggested but the Government does advert to the possibility of pairing and grouping of schools. Two points stand out. First, pairing and grouping are precisely what the Finger plan, adopted by the District Court, does. Second, in the circumstances of this case, these methods necessarily entail bussing.
I am not “favorably impressed” by the Government’s performance. Its vague and noncommital representations do little but obscure the real issues, introduce uncertainty and fail to meet the “heavy burden” necessary to overturn the District Court’s effective plan.3
[151]*151B. The Feasibility of the Plan
Of course it goes without saying that school boards are not obligated to do the impossible. Federal courts do not joust at windmills. Thus it is proper to ask whether a plan is feasible, whether it can be accomplished. There is no genuine dispute on this point. The plan is simple and quite efficient. A bus will make one pickup in the vicinity of the children’s residences, say in the white residential area. It then will make an express trip to the inner-city school. Because of the non-stop feature, time can be considerably shortened and a bus could make a return trip to pick up black students in the inner city and to convey them to the outlying school. There is no evidence of insurmountable traffic problems due to the increased bussing.4 Indeed, straight line bussing promises to be quicker. The present average one-way trip is over 15 miles and takes one hour and fourteen minutes; under the plan the average one-way trip for elementary students will be less than seven miles and 35 minutes. The cost of all of the additional bussing will be less than one week’s operating budget.5
C. The Standard of Review
In Brown II, the Supreme Court charged the district courts with the enforcement of the dictates of Brown I. The lower courts were to have “a practical flexibility in shaping * * * remedies.” 349 U.S. at 300, 75 S.Ct. at 756, 99 L.Ed. 1083. Thus, in subsuming these cases under traditional equity principles, the Supreme Court brought the desegregation decree within the rule that to be overturned it “must [be] demonstrate [d] that there was no reasonable basis for the District Judge’s decision.” United States v. W. T. Grant Co., 345 U.S. 629, 634, 73 S.Ct. 894, 898, 97 L.Ed. 1303 (1953). This court has paid homage to this maxim of appellate review when, in the past, a district judge has ordered less than comprehensive relief. Bradley v. School Board of City of Richmond, 345 F.2d 310, 320 (4th Cir. 1965), vac., 382 U.S. 103, 86 S.Ct. 224, 15 L.Ed. 2d 187 (1965). What is called for here is similar deference to an order that would finally inter the dual system and not preserve a nettlesome residue. As the Supreme Court made clear in Green, supra, those who would challenge an effective course of action bear a “heavy burden.” The Finger plan is a remarkably economical scheme when viewed in the light of what it accomplishes. There has been no showing that it can be improved or replaced by better or more [152]*152palatable means. It should, then, be sustained.
III
OBJECTIONS RAISED AGAINST THE COURT-ORDERED PLAN
A. The “Illegal” Objective of the Plan
My Brother Bryan expresses concern about the plan, regardless of cost, because it undertakes, in his view, an illegal objective: “achieving racial balance.” Whatever might be said for this view abstractly or in another context, it is not pertinent here. We are confronted in this case with no question of bussing for mere balance unrelated to a mandatory constitutional goal. What the District Court has ordered is compliance with the constitutional imperative to disestablish the existing segregation. Unless we are to palter with words, desegregation necessarily entails integration, that is to say integration in some substantial degree. The dictum to the contrary in Briggs v. Elliott, 132 F. Supp. 776 (E.D.S.C.1955), was rejected by necessary implication by the Supreme Court in Green, supra, and explicitly by this court in Walker v. County School Board of Brunswick Co., 413 F.2d 53, 54 n. 2 (4th Cir. 1969).
As my Brother Winter shows, there is no more suitable way of achieving this task than by setting, at least initially, a ratio roughly approximating that of the racial population in the school system. The District Judge adopted this ad hoe measurement as a starting guide, expressed a willingness to accept a degree of modification,6 and departed from it where circumstances required.
B. The “Unreasonableness” of the Plan
The majority does not quarrel with the plan’s objective, nor, accepting the findings of the District Court, does it really dispute that the plan can be achieved. Rather, we are told, the plan is an unreasonable burden.
This notion must be emphatically rejected. At bottom it is no more than an abstract, unexplicated judgment — a conclusion of the majority that, all things considered, desegregation of this school system is not worth the price. This is a conclusion neither we nor school boards are permitted to make.
In making policy decisions that are not constitutionally dictated, state authorities are free to decide in their discretion that a proposed measure is worth the cost involved or that the cost is unreasonable, and accordingly they may adopt or reject the proposal. This is not such a case. Vindication of the plaintiffs’ constitutional right does not rest in the school board’s discretion, as the Supreme Court authoritatively decided sixteen years ago and has repeated with increasing emphasis. It is not for the Board or this court to say that the cost of compliance with Brown is “unreasonable.”
That a subjective assessment is the operational part of the new “reasonableness” doctrine is highlighted by a study of the factors the majority bids school boards take into account in making bus-sing determinations. “[A] school board should take into consideration the age of the pupils, the distance and time required for transportation, the effect on traffic, and the cost in relation to the board’s resources.” But as we have seen, distance and time will be compara[153]*153tively short, the effect on traffic is un-demonstrated, and the incremental cost is marginal. As far as age is concerned, it has never prevented the bussing of pupils in Charlotte-Mecklenburg, or in North Carolina generally, where 70.9% of all bussed students are elementary pupils.
If the transportation of elementary pupils were a novelty sought to be introduced by the District Court, I could understand my brethren’s reluctance. But, as is conceded, bussing of children of elementary school age is an established tradition. Bussing has long been used to perpetuate dual systems.7 More importantly, bussing is a recognized educational tool in Charlotte-Mecklenburg and North Carolina. And as the National Education Association has admirably demonstrated in its brief, bussing has played a crucial role in the evolution from the one-room schoolhouse in this nation. Since the majority accepts the legitimacy of bussing, today’s decision totally baffles me.
In the final analysis, the elementary pupil phase of the Finger plan is disapproved because the percentage increase in bussing is somehow determined to be too onerous.8 Why this is so we are not told. The Board plan itself would bus 5000 additional pupils. The fact remains that in North Carolina 55% of all pupils are now being bussed. Under the Finger plan approximately 47% of the Charlotte-Mecklenburg student population would be bussed. This is well within the existing percentage throughout the state.
The majority’s proposal is inherently ambiguous. The court-ordered plan is said to be unreasonable. Yet the School Board’s own plan has also been disapproved. Does the decision — that the Finger plan is unreasonable — depend on the premise that an intermediate course is available? Would the amount of segregation retained in the School Board’s plan be avowedly sanctioned if it were recognized that nothing short of the steps delineated in the District Court’s plan will suffice to eliminate it? Since there is no practicable alternative, must we assume that the majority is willing to tolerate the deficiencies in the Board plan?
These questions remain unresolved and thus the ultimate meaning of the “reasonableness” doctrine is undefined. Suffice it to say that this case is not an appropriate one in which to grapple with the theoretical issue whether the law can endure a slight but irreducible remnant of segregated schools. This record presents no such problem. The remnant of racially identifiable elementary schools, to which the District Court addressed itself, encompasses over half the elementary population. This large fraction cannot be called slight; nor, as the Finger plan demonstrates, is it irreducible.
I am even more convinced of the unwisdom of reaching out to fashion a new “rule of reason,” when this record is far from requiring it, because of the serious consequences it would portend for the general course of school desegrega[154]*154tion. Handed a new litigable issue — the so-called reasonableness of a proposed plan — school boards can be expected to exploit it to the hilt. The concept is highly susceptible to delaying tactics in the courts. Everyone can advance a different opinion of what is reasonable. Thus, rarely would it be possible to make expeditious disposition of a board’s claim that its segregated system is not “reasonably” eradieable. Even more pernicious, the new-born rule furnishes a powerful incentive to communities to perpetuate and deepen the effects of race separation so that, when challenged, they can protest that belated remedial action would be unduly burdensome.
Moreover, the opinion catapults us back to the time, thought passed, when it was the fashion to contend that the inquiry was not how much progress has been made but the presence or absence of good faith on the part of the board. Whether an “intractable remnant of segregation” can be allowed to persist, apparently will now depend in large measure on a slippery test: an estimate of whether the Board has made “every reasonable effort to integrate the pupils under its control.” 9
[155]*155The Supreme Court having barred further delay by its insistent emphasis on an immediate remedy, we should not lend ourselves to the creation of a new loophole by attenuating the substance of desegregation.