This case is not the usual attack upon an at-large district. Rather it is an appeal from the district court’s rejection of the school board’s submission of two successive plans, which followed the board’s decision during this suit to abandon an at-large system. First, the board submitted a “3-4” plan in which three school board trustees would be elected at-large and four would be elected from single-member districts. The district court rejected this plan in favor of all single-member districts. The board then proposed a “5-2” plan with 5 members elected at-large from most of the district and 2 members elected from single-member districts located in minority neighborhoods. The district court also rejected the 5-2 plan, erroneously treating black and hispanic voters as a cohesive voting unit. I am persuaded that the district court’s finding that blacks and hispanics were a single politically cohesive minority under Thornburg v. Gingles, — U.S. —, 106 S.Ct. 2752, 92 L.Ed.2d 25 (1986) is clearly erroneous. I dissent.
I
A
In the Voting Rights Act Amendments of 1982, Congress rejected intent as an element of a violation in favor of an “effects” or “results” test. At the same time, Congress provided, as a political compromise, that “nothing in this section establishes a right to have members of a protected class elected in numbers equal to their proportion in the population.” 42 U.S.C. § 1973(b). The compromise did little more than sidestep the divisive issue, leaving to the courts the task of developing a principled way to distinguish illegal vote dilution from lost races; and to do so without depending so heavily upon the degree of minority success in elections that we make proportional representation — if not in name, in fact — the true rule. The efforts of the Gingles court to identify such a principle produced a 5-4 split along a line that, in substantial part, parallels the line between effects and proportional representation.
*1504Justice Brennan’s opinion in Gingles requires three proofs to establish that a multimember district dilutes a minority’s vote: A minority group must demonstrate (1) that it is sufficiently large and geographically compact to constitute a majority in a single-member district; (2) that it is politically cohesive; and (3) that the majority votes sufficiently as a bloc to enable it usually to defeat the minority’s preferred candidate. 106 S.Ct. at 2766-67.
In a section without majority support, Justice Brennan’s opinion also rejects the argument that proof of racially polarized voting requires proof of a causal relationship between race and voting, concluding that “only the correlation between race of voter and selection of certain candidates, not the causes of the correlation, matters,” id. at 2773; relatedly, that it is the “status of the candidate as the chosen representative of a particular racial group, not the race of the candidate, that is important,” id. at 2776 (emphasis in original).
Significantly, and critically, it was here that Justice White left Justice Brennan’s opinion, explaining that Justice Brennan’s conclusion that the “crucial factor in identifying polarized voting is the race of the voter and that the race of the candidate is irrelevant ... is interest-group politics rather than a rule hedging against racial discrimination.” Id. at 2783-84. The key component in cohesion is the presence of bloc voting. Consequently, the Court’s opinion, sans agreement on the definition of bloc voting presents a troubled construct, inviting inferior courts through its uncertainty to blur distinctions between ephemeral political coalitions and cohesive racial minorities, thereby confusing losing votes with “diluted” votes. The district court and the panel majority accepted the invitation.
B
Concluding on this record that blacks and hispanics are sufficiently cohesive to be counted as a single minority in the Gingles inquiry risks that ephemeral political alliances having little or no necessary connection to discrimination will be confused with cohesive political units joined by a common disability of chronic bigotry. Stated more directly, the determination that disparate minority groups are cohesive must be consistent with the purpose of the Act to provide racial and ethnic minorities the vote that they would have in the absence of forbidden discrimination.
The purpose of the Act is to redress racial or ethnic discrimination which manifests itself in voting patterns or electoral structures. The tie to race or national origin in Justice Brennan’s opinion in Gingles is the raw correspondence in votes and outcome. Its three step inquiry assumes a group unified by race or national origin and asks if it is cohesive in its voting. If a minority group lacks a common race or ethnicity, cohesion must rely principally on shared values, socio-economic factors, and coalition formation, making the group almost indistinguishable from political minorities as opposed to racial minorities. At the least, concluding that a political group lacking the cementing and predictive force of common race or national origin is nonetheless politically cohesive under Gingles is a difficult undertaking with significant risks. The risks include the reality that diluting the requirement of cohesion expands the mission of the Act beyond the treatment of present-day manifestations of chronic bigotry to a more general device for accommodating majority government and plural constituents — thereby revealing a distrust of the ability of our republican government to do so. I turn now to the record to determine the extent to which my fears have here been realized.
II
The district court first concluded that blacks and hispanics were sufficiently large and geographically compact to constitute a majority in a single-member district. It found that blacks and hispanics live predominately in the eastern and southeastern portions of Midland, an area that “roughly correlates with the areas encompassed by City Precincts 1, 2 and 3.” The district court noted that 92.43% of the black population of Midland and 73.7% of the hispan*1505ics live in the area with “only 7.6%” of the whites. But describing the makeup of this area as percentages of the entire Midland population is not necessarily a description of ethnic insularity. The undisputed facts are that this area is fairly well integrated and the minorities are dispersed throughout. The ethnic distribution in the general population of the area is 44% hispanic, 34% black, and 22% white — and even this fails to describe accurately the ethnic distribution of eligible voters.
Defendant’s expert testified without contradiction that the percentage of voters in the population varies among blacks, hispanics and whites, but must approach 65% to constitute a political majority. Under the district’s proposed 3-4 plan, hispanics would have had one safe district. But it is doubtful that blacks alone could acquire the necessary 65% to constitute a majority in any one district. Thus, blacks alone could not successfully attack an at-large system and certainly not the district’s proposed 3-4 plan, which the same minority leaders accepted for Midland’s city government. It is then no accident, and hardly determinative of political cohesion, that blacks did not bring this suit alone; it certainly does not substitute for evidence of cohesive voting.
Under the ordered 7-0 plan, most blacks are concentrated in District 1; yet they constitute only 45.81% of its total population. Accordingly, the district court combined blacks and hispanics to meet Gingles’ required political majority. Despite its novelty and significance, the district court’s sole justification for the combined minority unit was that it would be unjust to do otherwise:
Should the Plaintiffs fail to prove that the minorities cannot form coalitions, this would go to the query presented by prong two. Thus, to not allow an aggregation for the satisfaction of prong one would be inherently unjust. Therefore, from the evidence presented, appellees did satisfy prong one of the Gingles test.
The district court then turned to the question whether the “group is politically cohesive.” It concluded that while the groups may not
always have the same political goals, it is clear that the two groups have political goals that are inseparable. As such, coalition formation will often prove to be mutually beneficial to the two groups.
In short, the district court did not, by its conclusion that blacks and hispanics were a cohesive political force, find anything more than that they occasionally join their political hands. There is no record evidence to support a finding of any greater nexus; indeed the record evidence is to the contrary.
In the original trial of this case, and before Gingles focused the court and the parties upon causation by asking whether a minority had the votes for election of a single-member district, the district court explicitly found that blacks and hispanics were not a cohesive unit. In rejecting the proposed 3-4 plan, before Gingles, the district court explained:
Should the “3-4 plan” be placed in effect by this Court, it is certain that the minority population of the MISD would be able to elect but one representative. While the Defendants’ proposed District 2 contains a large percentage of minority group members, it cannot be said that all of the minorities residing in the proposed District 2 are homogeneous. On the contrary, while perhaps Blacks and Mexican Americans share common experiences in past discrimination practices of the majority, the two groups have distinct cultures and decidedly different political goals.
(emphasis supplied). After Gingles and a two-hour hearing on remand,1 the district court reached the opposite conclusion. The sole basis for its new conclusion that blacks and hispanics were a cohesive political body was their occasional coalitions, including the prosecution of this suit. The district court simply concluded that this was sufficient under Gingles to allow aggregation. The district court did not, and could not on this record, find that blacks and hispanics *1506were cohesive in any other sense. The reality is that how much of a coalition, as distinguished from cohesion, actually exists is itself seriously at issue.
The district court pointed to low income and poor education as common and tying concerns of blacks and hispanics, ignoring the fact that 22% of the population of this same part of Midland is white with similar socio-economic characteristics. This is not to say that poor whites are similarly situated in every relevant way; in the view of the Voting Rights Act they cannot be, freed as they are from the disabilities of discrimination due to race and, presumably, national origin. Our decision accepts, as Congress has accepted, that the headwind of chronic bigotry has slowed both blacks and hispanics in the economic race, but it does not follow that any resulting common political agenda — seen broadly — gives them political cohesion.2 Such evidence as there was pointed in the opposite direction.
The board offered a study conducted by the University of Texas reflecting the attitude of Midland residents toward the policies of the school board. The study demonstrated that hispanics and blacks had significantly different views regarding Midland schools. Nevertheless, the district court found that the study was not relevant “to predict voting behavior ... or how cohesive people will be,” that voter attitudes were not relevant to how people vote — a rather startling proposition. While it should be no surprise to one who has read the survey, after the court announced its 7-0 plan with the two safe coalition seats, a hispanic candidate promptly filed against black candidates in District 1 — the predominately black district — a curious exercise in coalition politics indeed. Moreover, Dr. Taebel, the only expert who testified on remand after Gingles and whose opinion was left virtually unchallenged by other expert testimony, was asked whether in his study of the Midland Independent School District he had seen any operating coalition between blacks and hispanics. He replied that “I have seen none whatsoever.”
In any event, this newly minted minority must also meet the third step in Gingles; it “must be able to demonstrate that the white majority votes sufficiently as a bloc to enable it — in the absence of special circumstances ... usually to defeat the minority’s preferred candidate.” Gingles, 106 S.Ct. at 2767 (citations omitted). Regarding possible white bloc voting to defeat the minority’s candidates, Taebel pointed out that in all elections since 1978, the candidate preferred by the minorities was elected eleven of fifteen times for a success rate of 73.78%. The evidence to which the district judge pointed, and his candid basis for refusing throughout this case to consider anything but a 7-0 plan, was a perception that the school district would be best served by a greater number of black and hispanic board members.
In finding that blacks and hispanics in the Midland Independent School District are a cohesive group under Gingles, the district court ignored historical and cultural differences between blacks and hispanics and lacked record support that in Midland the two were politically cohesive; as it relaxed the measure of cohesion to accommodate such differences, the district court increased the risk of frustrating congression*1507al will. On this record, his order draws upon a vision of “safe” districts that is unacceptably close to proportional representation.
Ill
The point is not that there were no coalitions. There doubtlessly were from time to time. Rather, the point is that sustaining the district court’s findings that blacks and hispanics are a single cohesive minority under Gingles has consequences far beyond Midland, Texas, because it reads cohesion in a way the Gingles court, and Congress, could not have intended. The error of the district court, perpetuated by the panel majority, is that it abrades the requisite moorings to race or national origin. Without such ties, we are about judicial superintendence of election outcomes in the name of protecting those less able to fend for themselves in the political arena, when inability is indistinguished from political loss.
In short, from conflicting signals from Congress, the Supreme Court in Gingles attempted to derive an administrable measure for determining whether an election practice or voting pattern is illegally correlated with race or national origin. Under the Court’s formulation, the requirement of minority cohesion and correlation of voting behavior with, race or national origin are critically in tandem, such that dilution of one requirement dilutes the other. Attenuating the test for cohesion and correspondingly, the correlation with race and national origin, drains the district court’s “enforcement” of the Voting Rights Act of statutory authority and constitutional support. In the last measure, the judgment becomes no more than a judicial sense that its result is more “just” than the challenged plans. I had supposed that the essence of our republican arrangement is that voting minorities lose; that, absent participation inhibited by racial or ethnic discrimination, Congress and the courts lack authority to further regulate.3
IV
Even though the board’s 3-4 plan complies with the Voting Rights Act, the question remains whether the district court is free to reject the plan instead of according it the deference courts owe legislative judgments.
A
Our jurisprudence makes plain that “reapportionment is primarily the duty and responsibility of the State through its legislature or other body, rather than of a federal court.” Chapman v. Meier, 420 U.S. 1, 27, 95 S.Ct. 751, 766, 42 L.Ed.2d 766 (1975) (citations omitted). Thus, a federal district court should follow the preferences of the State as expressed in reapportionment plans, unless adherence to the preferences violates the Constitution or federal statutory law. See Upham v. Seamon, 456 U.S. 37, 41, 102 S.Ct. 1518, 1521, 71 L.Ed.2d 725 (1982). Most relevant to this case, “the fact that the reapportionment plan ... was devised in response to an order of a federal court does not change its *1508character as a legislative plan.” McDaniel v. Sanchez, 452 U.S. 130, 146, 101 S.Ct. 2224, 2234, 68 L.Ed.2d 724 (1981).
B
Plaintiffs argue that the board lacked authority under state law to propose a 3-4 plan and that the plan should therefore be denied the deference normally accorded legislative plans. However, I am persuaded that the board did have authority to propose this plan in response to a court order, and that the court should defer to the board’s judgment.
Section 23.024(b) of the Texas Education Code provides that “[t]he Board of Trustees of a school district, on its own motion, may order that trustees of the district are to be elected from single-member districts or that not fewer than 70 percent of the members of the board of trustees are to be elected from single-member trustee districts” (emphasis supplied). The MISD’s proposed plans were not in an “order” adopted on the legislative body’s “own motion.” Rather, they were proposals or judgments proffered in response to the suit. Thus, the proposals retain their legislative character because they reflect the “policy choices” of the elected representatives of the district, Wise v. Lipscomb, 437 U.S. 535, 548, 98 S.Ct. 2493, 2501, 57 L.Ed.2d 411 (1978), and therefore merit deference as legislative plans, McDaniel, 452 U.S. at 152, 101 S.Ct. at 2237-38. Because they are proposals and not self-adopted orders, however, they do not violate Tex. Educ.Code § 23.024(b). Thus, the 3-4 and 5-2 plans were legislative plans, worthy of due deference.
Admittedly, one might object that this reading of “on its own motion” enables a school district to escape the will of the Texas State legislature by hiding behind a federal remedy. The reality is that the freedom of a school board exists only when it submits a proposal to a federal court in an attack upon its voting system. That the state legislature limited the power of independent school districts to carve its districts, “on its own,” does not mean that it intended to limit its ability to respond to a federal order. On the other hand, and it gives me substantial pause, the statute can be read as expressing the state’s fear of local gerrymanders by its preference for either at-large or single-member districts. At the same time, the language arguably is only an intended distinction between board initiated and voter initiated change. Absent a violation of the Voting Rights Act or the Constitution, the ordering of the voting is the state’s business. Our task then is to follow state policy and while not free of doubt, I am persuaded that Texas did not intend to so limit school boards in their defense of voter right suits.
V
Even were the majority correct in determining that the proposed plans violate either the Voting Rights Act or the Texas Education Code, it does not follow that the remedy the district court imposed should be affirmed. The court must still consider the effect of its decision on the “Agreed Order” dated October 10, 1985, in which the board stated it was “willing for a change to be made in the election method for school board members” and proposed the 3-4 plan.
The agreement contemplated that the 3-4 plan would be implemented if found to comply with the Voting Rights Act. The agreement does not indicate what should happen if the district court finds that the 3-4 plan violates the Act or if the court refuses to enter the plan as a federal remedy, even if it satisfies the Act. Conceivably, the district board may have preferred to defend the at-large election system in that case. We should remand for the district court to determine whether the board had conceded that the at-large election system violated the Act, or whether the willingness to change was conditioned on the court’s acceptance of the 3-4 proposal.4
*1509VI
In sum, I dissent from the majority’s determination that the district court properly applied the Voting Rights Act and Gingles, especially in light of the failure of the evidence to establish cohesion of his-panic and black voters, as I would read cohesion. Having concluded that the school board’s agreement to institute a 3-4 or 5-2 plan is ambiguous and possibly invalid under the majority’s reasoning, I would remand this case for resolution of the meaning and continued vitality of the Agreed Order.
. The plaintiffs’ case on remand occupies eleven pages of transcript.'
. Compare the findings of the district court here with the observation of Justice Stevens that:
In the line-drawing process, racial, religious, ethnic, and economic gerrymanders are all species of political gerrymanders.
From the standpoint of the groups of voters that are affected by the line-drawing process, it is also important to recognize that it is the group’s interest in gaining or maintaining political power that is at stake. The mere fact that a number of citizens share a common ethnic, racial, or religious background does not create the need for protection against gerrymandering. It is only when their common interests are strong enough to be manifested in political action that the need arises. For the political strength of a group is not a function of its ethnic, racial, or religious composition; rather it is a function of numbers — specifically the number of persons who will vote in the same way.” Mobile v. Bolden, 446 U.S. 55, 100 S.Ct. 1490, 64 L.Ed.2d 47 (1980) (concurring in judgment). See Cousins v. City Council of City of Chicago, 466 F.2d 830, 851-852 (7th Cir.) (Stevens, J., dissenting), cert. denied, 409 U.S. 893, 93 S.Ct. 85, 34 L.Ed.2d 151 (1972).
. It is true that Congress could have required lesser proof. In theory at least, Congress could have so relaxed the proof that a putative minority need prove only that it was not represented according to its number, presuming that such underrepresentation had a sufficiently high likelihood of being tied to the illegality that its proscription was warranted. I do not pause to treat the difficulties posed by a statutory command of proportional representation. See Katzenbach v. Morgan, 384 U.S. 641, 86 S.Ct. 1717, 16 L.Ed.2d 828 (1966). Congress did not do so; perhaps because proportional representation as a measure of due strength will not do, but rather is itself a political alien, antithetical to our basic devotion to republican government. Concern for accommodation of factions has seldom been far away when debate about government implicates basic values. See The Federalist No. 10 (J. Madison); F. McDonald, Novas Ordo Seclorum: The Intellectual Origins of the Constitution 162-64 (1985). I emphasize that Justice Brennan’s construct, as it shifts from race as a causative force, moves toward a theory of group rights and away from a theory of individual voting rights; it is a specie of gerrymandering — sans the equal protection requirement of intent. Relatedly, given the uncertainty of the Gingles court, it is not surprising that the issue of bloc voting was not resolved. It cannot be, absent agreement regarding the fundamental role of race in the enforcement of the Voting Rights Act.
. An additional problem is, in light of the majority's statement that the board lacked authority to propose a 3-4 plan, whether the agreed order that is so heavily based on that proposal is itself void for want of authority.