MAHAN v. HOWELL

MR. JUSTICE REHNQUIST delivered the opinion of the Court.

Acting pursuant to the mandate of its newly revised state constitution,1 the Virginia General Assembly enacted statutes apportioning the State for the election of members of its House of Delegates2 and Senate.3 Two suits were brought challenging the constitutionality of the House redistricting statute on the grounds that there were impermissible population variances in the districts, that the multimember districts diluted representation,4 and that the use of multimember districts constituted racial gerrymandering.5 The Senate redistricting statute was attacked in a separate suit, which alleged that the city of Norfolk was unconstitutionally split into three districts, allocating Navy personnel "home-ported" in Norfolk to one district and isolating Negro voters in one district. Three three-judge district courts were convened to hear the suits pursuant to 28 U. S. C. §§ 2281 and 2284. The suits were consolidated and heard by the four judges who variously made up the three three-judge panels.

The consolidated District Court entered an interlocutory order that, inter alia, declared the legislative reapportionment statutes unconstitutional and enjoined the holding of elections in electoral districts other than those established by the court's opinion. Howell v. Mahan, 330 F.Supp. 1138, 1150 (ED Va. 1971). Appellants, the Secretary of the State Board of Elections and its members and the city of Virginia Beach, have appealed directly to this Court from those portions of the court's order, invoking our jurisdiction under 28 U. S. C. § 1253.

I

The statute apportioning the House provided for a combination of 52 single-member, multimember, and floater delegate districts from which 100 delegates would be elected. As found by the lower court, the ideal district in Virginia consisted of 46,485 persons per delegate, and the maximum percentage variation from that ideal under the Act was 16.4%—the 12th district being over-represented by 6.8% and the 16th district being under-represented by 9.6%.6 The population ratio between these two districts was 1.18 to 1. The average percentage variance under the plan was ±3.89%, and the minimum population percentage necessary to elect a majority of the House was 49.29%. Of the 52 districts, 35 were within 4% of perfection and nine exceeded a 6% variance from the ideal. With one exception, the delegate districts followed political jurisdictional lines of the counties and cities. That exception, Fairfax County, was allotted 10 delegates but was divided into two five-member districts.

Relying on Kirkpatrick v. Preisler, 394 U.S. 526 (1969), Wells v. Rockefeller, 394 U.S. 542 (1969), and Reynolds v. Sims, 377 U.S. 533 (1964), the District Court concluded that the 16.4% variation was sufficient to condemn the House statute under the "one person, one vote" doctrine. While it noted that the variances were traceable to the desire of the General Assembly to maintain the integrity of traditional county and city boundaries, and that it was impossible to draft district lines to overcome unconstitutional disparities and still maintain such integrity, it held that the State proved no governmental necessity for strictly adhering to political subdivision lines. Accordingly, it undertook its own re-districting and devised a plan having a percentage variation of slightly over 10% from the ideal district, a percentage it believed came "within passable constitutional limits as `a good-faith effort to achieve absolute equality.' Kirkpatrick v. Preisler . . . ." Howell v. Mahan, 330 F. Supp., at 1147-1148.

Appellants contend that the District Court's reliance on Kirkpatrick v. Preisler, supra, and Wells v. Rockefeller, supra, in striking down the General Assembly's reapportionment plan was erroneous, and that proper application of the standards enunciated in Reynolds v. Sims, supra, would have resulted in a finding that the statute was constitutional.

In Kirkpatrick v. Preisler and Wells v. Rockefeller, this Court invalidated state reapportionment statutes for federal congressional districts having maximum percentage deviations of 5.97% and 13.1% respectively. The express purpose of these cases was to elucidate the standard first announced in the holding of Wesberry v. Sanders, 376 U.S. 1 (1964), that "the command of Art. I, § 2, that Representatives be chosen `by the People of the several States' means that as nearly as is practicable one man's vote in a congressional election is to be worth as much as another's." Id., at 7-8 (footnotes omitted). And it was concluded that that command "permits only the limited population variances which are unavoidable despite a good-faith effort to achieve absolute equality, or for which justification is shown." Kirkpatrick v. Preisler, supra, at 531. The principal question thus presented for review is whether or not the Equal Protection Clause of the Fourteenth Amendment likewise permits only "the limited population variances which are unavoidable despite a good-faith effort to achieve absolute equality" in the context of state legislative reapportionment.7

This Court first recognized that the Equal Protection Clause requires both houses of a bicameral state legislature to be apportioned substantially on a population basis in Reynolds v. Sims, supra. In so doing, it suggested that in the implementation of the basic constitutional principle—equality of population among the districts —more flexibility was constitutionally permissible with respect to state legislative reapportionment than in congressional redistricting. Id., at 578. Consideration was given to the fact that, almost invariably, there is a significantly larger number of seats in state legislative bodies to be distributed within a State than congressional seats, and that therefore it may be feasible for a State to use political subdivision lines to a greater extent in establishing state legislative districts than congressional districts while still affording adequate statewide representation. Ibid. Another possible justification for deviation from population-based representation in state legislatures was stated to be:

The Court reiterated that the overriding objective in reapportionment must be "substantial equality of population among the various districts, so that the vote of any citizen is approximately equal in weight to that of any other citizen in the State." Id., at 579.

By contrast, the Court in Wesberry v. Sanders, supra, recognized no excuse for the failure to meet the objective of equal representation for equal numbers of people in congressional districting other than the practical impossibility of drawing equal districts with mathematical precision. Thus, whereas population alone has been the sole criterion of constitutionality in congressional redistricting under Art. I, § 2, broader latitude has been afforded the States under the Equal Protection Clause in state legislative redistricting because of the considerations enumerated in Reynolds v. Sims, supra. The dichotomy between the two lines of cases has consistently been maintained. In Kirkpatrick v. Preisler, for example, one asserted justification for population variances was that they were necessarily a result of the State's attempt to avoid fragmenting political subdivisions by drawing congressional district lines along existing political subdivision boundaries. This argument was rejected in the congressional context. But in Abate v. Mundt, 403 U.S. 182 (1971), an apportionment for a county legislature having a maximum deviation from equality of 11.9% was upheld in the face of an equal protection challenge, in part because New York had a long history of maintaining the integrity of existing local government units within the county.

Application of the "absolute equality" test of Kirkpatrick and Wells to state legislative redistricting may impair the normal functioning of state and local governments. Such an effect is readily apparent from an analysis of the District Court's plan in this case. Under Art. VII, §§ 2 and 3 of Virginia's Constitution, the General Assembly is given extensive power to enact special legislation regarding the organization of, and the exercise of governmental powers by, counties, cities, towns, and other political subdivisions. The statute redistricting the House of Delegates consistently sought to avoid the fragmentation of such subdivisions, assertedly to afford them a voice in Richmond to seek such local legislation.

The court's reapportionment, based on its application of Kirkpatrick and Wells, resulted in a maximum deviation of slightly over 10%,8 as compared with the roughly 16% maximum variation found in the plan adopted by the legislature. But to achieve even this limit of variation, the court's plan extended single and multimember districts across subdivision lines in 12 instances, substituting population equality for subdivision representation. Scott County, for example, under the Assembly's plan was placed in the first district and its population of 24,376 voted with the 76,346 persons in Dickinson, Lee, and Wise Counties for two delegates. The district thus established deviated by 8.3% from the ideal. The court transferred five of Scott County's enumeration districts, containing 6,063 persons, to the contiguous second district composed of the city of Bristol, and Smith and Washington Counties, population 87,041. Scott County's representation was thereby substantially reduced in the first district, and all but nonexistent in the second district. The opportunity of its voters to champion local legislation relating to Scott County is virtually nil. The countervailing benefit resulting from the court's readjustment is the fact that the first district's deviation from the ideal is now reduced to 1.8%.

The city of Virginia Beach saw its position deteriorate in a similar manner under the court-imposed plan. Under the legislative plan, Virginia Beach constituted the 40th district and was allocated three delegates for its population of 172,106. The resulting underrepresentation was cured by providing a floterial district, the 42d, which also included portions of the cities of Chesapeake and Portsmouth. Under the court's plan, the 42d district was dissolved. Of its 32,651 persons that constituted the deviation from the ideal for the 40th district, 3,515 were placed in the 40th, and 29,136 were transferred to Norfolk's 39th district. The 39th district is a multimember district that includes the 307,951 persons who make up the population of the city of Norfolk. Thus, those Virginia Beach residents who cast their vote in the 39th district amount to only 8.6% of that district's population. In terms of practical politics, Virginia Beach complains that such representation is no representation at all so far as local legislation is concerned, and that those 29,136 people transferred to the 39th district have in that respect been effectively disenfranchised.

We conclude, therefore, that the constitutionality of Virginia's legislative redistricting plan was not to be judged by the more stringent standards that Kirkpatrick and Wells make applicable to congressional reapportionment, but instead by the equal protection test enunciated in Reynolds v. Sims, supra. We reaffirm its holding that "the Equal Protection Clause requires that a State make an honest and good faith effort to construct districts, in both houses of its legislature, as nearly of equal population as is practicable." 377 U. S., at 577. We likewise reaffirm its conclusion that "[s]o long as the divergences from a strict population standard are based on legitimate considerations incident to the effectuation of a rational state policy, some deviations from the equal-population principle are constitutionally permissible with respect to the apportionment of seats in either or both of the two houses of a bicameral state legislature." Id., at 579.

The asserted justification for the divergences in this case—the State's policy of maintaining the integrity of political subdivision lines—is not a new one to this Court. In Davis v. Mann, 377 U.S. 678, 686 (1964), it was noted:

The then-existing substantial deviation in the apportionment of both Houses defeated the constitutionality of Virginia's districting statutes in that case, but the possibility of maintaining the integrity of political subdivision lines in districting was not precluded so long as there existed "such minor deviations only as may occur in recognizing certain factors that are free from any taint of arbitrariness or discrimination." Roman v. Sincock, 377 U.S. 695, 710 (1964).

We are not prepared to say that the decision of the people of Virginia to grant the General Assembly the power to enact local legislation dealing with the political subdivisions is irrational. And if that be so, the decision of the General Assembly to provide representation to subdivisions qua subdivisions in order to implement that constitutional power is likewise valid when measured against the Equal Protection Clause of the Fourteenth Amendment. The inquiry then becomes whether it can reasonably be said that the state policy urged by Virginia to justify the divergences in the legislative reapportionment plan of the House is, indeed, furthered by the plan adopted by the legislature, and whether, if so justified, the divergences are also within tolerable limits. For a State's policy urged in justification of disparity in district population, however rational, cannot constitutionally be permitted to emasculate the goal of substantial equality.

There was uncontradicted evidence offered in the District Court to the effect that the legislature's plan, subject to minor qualifications, "produces the minimum deviation above and below the norm, keeping intact political boundaries. . . ." (Defendants' Exhibit 8.) That court itself recognized that equality was impossible if political boundaries were to be kept intact in the process of districting. But it went on to hold that since the State "proved no governmental necessity for strictly adhering to political subdivision lines," the legislative plan was constitutionally invalid. Howell v. Mahan, supra, at 1140. As we noted above, however, the proper equal protection test is not framed in terms of "governmental necessity," but instead in terms of a claim that a State may "rationally consider." Reynolds v. Sims, supra, at 580-581.

The District Court intimated that one reason for rejecting the justification for divergences offered by the State was its conclusion that the legislature had not in fact implemented its asserted policy, "as witness the division of Fairfax County." Howell v. Mahan, supra, at 1140. But while Fairfax County was divided, it was not fragmented. And had it not been divided, there would have been one ten-member district in Fairfax County, a result that this Court might well have been thought to disfavor as a result of its opinion in Connor v. Johnson, 402 U.S. 690, 692 (1971). The State can scarcely be condemned for simultaneously attempting to move toward smaller districts and to maintain the integrity of its political subdivision lines.

Appellees argue that the traditional adherence to such lines is no longer a justification since the Virginia constitutional provision regarding reapportionment, Art, II, § 6, supra, n. 1, neither specifically provides for apportionment along political subdivision lines nor draws a distinction between the standards for congressional and legislative districting. The standard in each case is described in the "as nearly as is practicable" language used in Wesberry v. Sanders, supra, and Reynolds v. Sims, supra. But, as we have previously indicated, the latitude afforded to States in legislative redistricting is somewhat broader than that afforded to them in congressional redistricting. Virginia was free as a matter of federal constitutional law to construe the mandate of its Constitution more liberally in the case of legislative redistricting than in the case of congressional redistricting, and the plan adopted by the legislature indicates that it has done so.

We also reject the argument that, because the State is not adhering to its tradition of respecting the boundaries of political subdivisions in congressional and State Senate redistricting, it may not do so in the case of redistricting for the House of Delegates. Nothing in the fact that Virginia has followed the constitutional mandate of this Court in the case of congressional redistricting, or that it has chosen in some instances to ignore political subdivision lines in the case of the State Senate, detracts from the validity of its consistently applied policy to have at least one house of its bicameral legislature responsive to voters of political subdivisions as such.9

We hold that the legislature's plan for apportionment of the House of Delegates may reasonably be said to advance the rational state policy of respecting the boundaries of political subdivisions. The remaining inquiry is whether the population disparities among the districts that have resulted from the pursuit of this plan exceed constitutional limits. We conclude that they do not.

The most stringent mathematical standard that has heretofore been imposed upon an apportionment plan for a state legislature by this Court was enunciated in Swann v. Adams, 385 U.S. 440 (1967), where a scheme having a maximum deviation of 26% was disapproved. In that case, the State of Florida offered no evidence at the trial level to support the challenged variations with respect to either the House or Senate. Id., at 446. The Court emphasized there that "the fact that a 10% or 15% variation from the norm is approved in one State has little bearing on the validity of a similar variation in another State." Id., at 445. We, therefore, find the citations to numerous cases decided by state and lower federal courts to be of limited use in determining the constitutionality of Virginia's statute. The relatively minor variations present in the Virginia plan contrast sharply with the larger variations in state legislative reapportionment plans that have been struck down by previous decisions of this Court. See, e. g., Reynolds v. Sims, supra; Swann v. Adams, supra; and Kilgarlin v. Hill, 386 U.S. 120 (1967).

Neither courts nor legislatures are furnished any specialized calipers that enable them to extract from the general language of the Equal Protection Clause of the Fourteenth Amendment the mathematical formula that establishes what range of percentage deviations is permissible, and what is not. The 16-odd percent maximum deviation that the District Court found to exist in the legislative plan for the reapportionment of the House is substantially less than the percentage deviations that have been found invalid in the previous decisions of this Court. While this percentage may well approach tolerable limits, we do not believe it exceeds them. Virginia has not sacrificed substantial equality to justifiable deviations.

The policy of maintaining the integrity of political subdivision lines in the process of reapportioning a state legislature, the policy consistently advanced by Virginia as a justification for disparities in population among districts that elect members to the House of Delegates, is a rational one. It can reasonably be said, upon examination of the legislative plan, that it does in fact advance that policy. The population disparities that are permitted thereunder result in a maximum percentage deviation that we hold to be within tolerable constitutional limits. We, therefore, hold the General Assembly's plan for the reapportionment of the House of Delegates constitutional and reverse the District Court's conclusion to the contrary. We also affirm Weinberg v. Prichard et al., No. 71-444, held pending this disposition.10

II

The General Assembly divided the State into 40 single-member senatorial districts. Under the plan, a portion of the city of Virginia Beach was added to the city of Norfolk and the entire area was divided into three single-member districts, which the court below found conformed almost ideally, numerically, to the "one person, one vote" principle. But all naval personnel "home-ported" at the U. S. Naval Station, Norfolk, about 36,700 persons, were assigned to the Fifth Senatorial District because that is where they were counted on official census tracts.11 It was undisputed that only about 8,100 of such personnel lived aboard vessels assigned to the census tract within the Fifth District. The court had before it evidence that about 18,000 lived outside the Fifth District but within the Norfolk and Virginia Beach areas that, if true, indicated a malapportionment with respect to such personnel.12 Lacking survey data sufficiently precise to permit the creation of three single-member districts more closely representing the actual population, the court corrected the disparities by establishing one multimember district composed of the Fifth, Sixth, and Seventh Districts, encompassing the city of Norfolk and a portion of Virginia Beach. Howell v. Mahan, supra.

Appellants charge that the District Court was not justified in overturning the districts established by the General Assembly since the Assembly validly used census tracts in apportioning the area and that the imposition by the court of a multimember district contravened the valid legislative policy in favor of single-member districts. We conclude that under the unusual, if not unique, circumstances in this case the District Court did not err in declining to accord conclusive weight to the legislative reliance on census figures. That court justifiably found that with respect to the three single-member districts in question, the legislative plan resulted in both significant population disparities and the assignment of military personnel to vote in districts in which they admittedly did not reside. Since discriminatory treatment of military personnel in legislative reapportionment is constitutionally impermissible, Davis v. Mann, supra, at 691, we hold that the interim relief granted by the District Court as to the State Senate was within the bounds of the discretion confided to it.

Application of interim remedial techniques in voting rights cases has largely been left to the district courts. Reynolds v. Sims, supra, at 585. The courts are bound to apply equitable considerations and in Reynolds it was stated that "[i]n awarding or withholding immediate relief, a court is entitled to and should consider the proximity of a forthcoming election and the mechanics and complexities of state election laws . . . ." Ibid.

The court below was faced with severe time pressures.

The reapportionment plans were first forwarded to the Attorney General on March 1, 1971. By April 7, these three cases had been filed and consolidated. The first hearing was scheduled for May 24, but on May 7, the Attorney General interposed his objections pursuant to the Voting Rights Act. As a result, the May 24 hearing was largely devoted to arguing about the effect of such objections and after that hearing, the court directed the cases to be continued until June 15. It also postponed the primary elections, which had been set for June 8, until September 14. The cases were finally heard on June 16, and the court's interlocutory order was entered on July 2, just two weeks prior to the revised July 16 filing deadline for primary candidates.

Prior to the time the court acted, this Court had handed down Whitcomb v. Chavis, 403 U.S. 124 (1971), recognizing that multimember districts were not per se violative of the Equal Protection Clause. The court conscientiously considered both the legislative policy and this Court's admonition in Connor v. Johnson, supra, that in fashioning apportionment remedies, the use of single-member districts is preferred. But it was confronted with plausible evidence of substantial malapportionment with respect to military personnel, the mandate of this Court that voting discrimination against military personnel is constitutionally impermissible, Davis v. Mann, supra, at 691-692, and the fear that too much delay would have seriously disrupted the fall 1971 elections. Facing as it did this singular combination of unique factors, we cannot say that the District Court abused its discretion in fashioning the interim remedy of combining the three districts into one multimember district.13 We, therefore, affirm the order of that Court insofar as it dealt with the State Senate.

Affirmed in part, reversed in part.

MR. JUSTICE POWELL took no part in the consideration or decision of these cases.

MR. JUSTICE BRENNAN, with whom MR. JUSTICE DOUGLAS and MR. JUSTICE MARSHALL join, concurring in part and dissenting in part.

I agree with the Court in No. 71-373, City of Virginia Beach v. Howell, that the joinder by the District Court of three senatorial districts in the Norfolk-Virginia Beach area to create one multimember senatorial district for the 1971 election was permissible under the special circumstances of this case. Cf. Whitcomb v. Chavis, 403 U.S. 124, 176-179 (1971) (DOUGLAS, J., concurring and dissenting); see Fortson v. Dorsey, 379 U.S. 433, 439 (1965); Burns v. Richardson, 384 U.S. 73, 88 (1966). I dissent, however, in No. 71-364, Mahan v. Howell, from the Court's action in setting aside the District Court's finding that the apportionment of the State House of Delegates violated the Equal Protection Clause of the Fourteenth Amendment.

The Court approves a legislative apportionment plan that is conceded to produce a total deviation of at least 16.4% from the constitutional ideal.1 Of course, "the fact that a 10% or 15% variation from the norm is approved in one State has little bearing on the validity of a similar variation in another State." Swann v. Adams, 385 U.S. 440, 445 (1967). "What is marginally permissible in one State may be unsatisfactory in another, depending on the particular circumstances of the case." Reynolds v. Sims, 377 U.S. 533, 578 (1964). Since every reapportionment case presents as its factual predicate a unique combination of circumstances, decisions upholding or invalidating a legislative plan cannot normally have great precedential significance. Abate v. Mundt, 403 U.S. 182, 189 (1971) (BRENNAN, J., dissenting). But language in the Court's opinion today suggests that more may be at stake than the application of well-established principles to a novel set of facts. In my view, the problem in the case before us is in no sense one of first impression, but is squarely controlled by our prior decisions. See Kirkpatrick v. Preisler, 394 U.S. 526 (1969); Swann v. Adams, supra; Reynolds v. Sims, supra; Davis v. Mann, 377 U.S. 678 (1964); Roman v. Sincock, 377 U.S. 695 (1964). It is appropriate, therefore, to call to mind again the controlling principles and to show that, properly applied to the facts of the case before us, they preclude a reversal of the District Court's decision.

I

Virginia's recently amended Constitution provides that "members of the Senate and of the House of Delegates of the General Assembly shall be elected from electoral districts established by the General Assembly," and "[e]very electoral district shall be composed of contiguous and compact territory and shall be so constituted as to give, as nearly as is practicable, representation in proportion to the population of the district." Art. II, § 6. Pursuant to that requirement, the General Assembly in 1971 divided the Commonwealth into 52 legislative districts from which the 100 members of the House of Delegates were to be elected.

On the basis of 1970 census figures, which set the population of the Commonwealth at 4,648,494, each delegate should ideally represent 46,485 persons. While the legislature's plan does not disregard constitutional requirements to the flagrant extent of many earlier cases,2 it does, nevertheless, demonstrate a systematic pattern of substantial deviation from the constitutional ideal. Under the 1971 plan, more than 25% of the delegates would be elected from districts in which the population deviates from the ideal by more than 5%. Almost 60% of the delegates would represent districts that deviate by more than 3%. Four legislators would be elected from districts that are overrepresented or underrepresented by more than 8%. And the maximum deviation—the spread between the most overrepresented and the most underrepresented districts—would be at least 16.4%, and might be as high as 23.6%, depending on the method of calculation.

Assuming a maximum deviation of 16.4%, the legislature's plan is still significantly less representative than many plans previously struck down by state and lower federal courts.3 Appellees maintain, however, that the total deviation, properly computed, is in fact 23.6%— a figure closely approximating the 25.65% deviation that led us to invalidate the Senate plan in Swann v. Adams, supra, the 26.48% deviation that led us to invalidate the House plan in Kilgarlin v. Hill, 386 U.S. 120 (1967), and the 24.78% deviation that led us to invalidate the House plan in Whitcomb v. Chavis, 403 U.S. 124, 161-163 (1971). Appellees arrive at the figure of 23.6% by taking into account the deviations in floterial districts, see App. 81-83, and appellants seem to concede that 23.6% is an accurate indicator of the total deviation. See Brief for Appellant Commonwealth of Virginia 7.4

The District Court pointed out that the "range of deviation may exceed 16.4%," 330 F.Supp. 1138, 1139 n. 1 (ED Va. 1971), but it had no occasion to consider whether 23.6% was the more accurate figure because of its finding that "[u]nder either mode of calculation . . . the statewide range of deviation will not pass constitutional muster." Ibid. Although conceding that the District Court did not reject or disparage appellees' assertion of a 23.6% deviation, the Court nevertheless reaches the perplexing conclusion that we "confine our consideration to the figures actually found by the court and used to support its holding of unconstitutionality"— 16.4%. Ante, at 319, n. 6. But if the legislature's plan does, in fact, "pass constitutional muster" on the assumption of a 16.4% deviation, then it is surely fair to ask whether the plan would still be valid assuming a total deviation of 23.6%. The Court refuses either to confront the question directly or to render it moot by determining that the figure of 23.6% is irrelevant because improperly derived. Instead, it attempts to obscure the issue by contending that the Commonwealth and the city of Virginia Beach disputed appellees' assertion of a 23.6% total deviation. That contention is wholly incorrect. Neither in the answers filed in the District Court, nor in the briefs, nor at oral argument did the Commonwealth or the city of Virginia Beach quarrel with appellee's method of calculating the deviation in floterial districts. See n. 4, supra. The Court's refusal to consider the question can only mean that appellees have the option of reopening this litigation in the District Court in an attempt to persuade that court that the true measure of the deviation is 23.6% and that a deviation of this order is fatal to the Commonwealth's plan.

In my view, there is no need to prolong this litigation by resolution in the court below of an issue that this Court should, but inexplicably does not, decide. The District Court correctly held that deviations of the magnitude of even 16.4% are sufficient to invalidate the legislature's plan. And that court added—again correctly —that "[i]n reapportionment cases the burden is on the State to justify deviations from parity by `legitimate considerations incident to the effectuation of a rational state policy.' Reynolds v. Sims, 377 U.S. 533, 579 (1964); see Swann v. Adams, 385 U.S. 440, 444 (1967). The State has proved no governmental necessity for strictly adhering to political subdivision lines." 330 F. Supp., at 1140. Accordingly, the District Court promulgated its own apportionment plan, which significantly reduced the extent of deviation.

Under the District Court's plan, the maximum deviation would be 7.2%,5 excluding one district which is geographically isolated from the mainland of the Commonwealth.6 And, even including that isolated district, the maximum total deviation would not exceed 10.2%. But the substantial reduction in the maximum deviation does not in itself make clear the full measure of the improvement achieved by the District Court's plan. The number of delegates whose districts deviate from the norm by 3% or more would be almost cut in half, from 58 to 32. And of the 32 districts still exceeding the 3% mark, only one—the geographically isolated district—would exceed the mean by more than 3.7%. In short, while the District Court did not achieve its stated goal of "perfect mathematical division" because of the "multiplicity of delegates, the geography of the State and the diversity of population concentrations," 330 F. Supp., at 1147, its plan would still produce measurably greater equality of representation.

Appellants necessarily concede that the District Court's plan would reduce the inequality in population per district, but they defend the legislature's plan on the ground that "tolerance of political jurisdictional lines is justification for some deviation," Brief for Appellant Commonwealth of Virginia 24. They maintain that the legislature's plan achieved the highest degree of equality possible without fragmenting political subdivisions. The principal question presented for our decision is whether on the facts of this case an asserted state interest in preserving the integrity of county lines can justify the resulting substantial deviations from population equality.

II

The holdings of our prior decisions can be restated in two unequivocal propositions. First, the paramount goal of reapportionment must be the drawing of district lines so as to achieve precise equality in the population of each district.7 "[T]he Equal Protection Clause requires that a State make an honest and good faith effort to construct districts, in both houses of its legislature, as nearly of equal population as is practicable." Reynolds v. Sims, 377 U. S., at 577; see also Kirkpatrick v. Preisler, 394 U. S., at 531. The Constitution does not permit a State to relegate considerations of equality to secondary status and reserve as the primary goal of apportionment the service of some other state interest.

Second, it is open to the State, in the event that it should fail to achieve the goal of population equality, to attempt to justify its failure by demonstrating that precise equality could not be achieved without jeopardizing some critical governmental interest. The Equal Protection Clause does not exalt the principle of equal representation to the point of nullifying every competing interest of the State. But we have held firmly to the view that variations in weight accorded each vote can be approved only where the State meets its burden of presenting cogent reasons in explanation of the variations, and even then only where the variations are small. See, e. g., Abate v. Mundt, 403 U.S. 182 (1971); Kirkpatrick v. Preisler, supra; Swann v. Adams, supra.

The validity of these propositions and their applicability to the case before us are not at all diminished by the fact that Kirkpatrick v. Preisler and Wells v. Rockefeller, 394 U.S. 542 (1969)—two of the many cases in which the propositions were refined and applied— concerned the division of States into federal congressional districts rather than legislative reapportionment. Prior to today's decision, we have never held that different constitutional standards are applicable to the two situations. True, there are significant differences between congressional districting and legislative apportionment, and we have repeatedly recognized those differences. In Reynolds v. Sims, for example, we termed "more than insubstantial" the argument that "a State can rationally consider according political subdivisions some independent representation in at least one body of the state legislature, as long as the basic standard of equality of population among districts is maintained." 377 U. S., at 580. See also id., at 578; Abate v. Mundt, supra. But the recognition of these differences is hardly tantamount to the establishment of two distinct controlling standards. What our decisions have made clear is that certain state interests that are pertinent to legislative reapportionment can have no possible relevance to congressional districting. Thus, the need to preserve the integrity of political subdivisions as political subdivisions may, in some instances, justify small variations in the population of districts from which state legislators are elected. But that interest can hardly be asserted in justification of malapportioned congressional districts. Kirkpatrick v. Preisler, supra. While the State may have a broader range of interests to which it can point in attempting to justify a failure to achieve precise equality in the context of legislative apportionment, it by no means follows that the State is subject to a lighter burden of proof or that the controlling constitutional standard is in any sense distinguishable.

Our concern in Kirkpatrick v. Preisler was with the constitutional requirement that "as nearly as is practicable one man's vote in a congressional election is to be worth as much as another's." Wesberry v. Sanders, 376 U.S. 1, 7-8 (1964). We rejected the State's argument that "there is a fixed numerical or percentage population variance small enough to be considered de minimis and to satisfy without question the `as nearly as practicable' standard. . . . Since `equal representation for equal numbers of people [is] the fundamental goal for the House of Representatives,' Wesberry v. Sanders, supra, at 18, the `as nearly as practicable' standard requires that the State make a good-faith effort to achieve precise mathematical equality. See Reynolds v. Sims, 377 U.S. 533, 577 (1964)." Kirkpatrick v. Preisler, supra, at 530-531. Moreover, we held, id., at 532, that "[i]t was the burden of the State `to present . . . acceptable reasons for the variations among the populations of the various . . . districts . . . .' Swann v. Adams, supra, at 443-444."

The principles that undergirded our decision in Kirkpatrick v. Preisler are the very principles that supported our decision in Swann v. Adams, a case involving the apportionment of a state legislature. The opinion in Kirkpatrick does not suggest that a different standard might be applicable to congressional districting. On the contrary, the "as nearly as practicable" standard with which we were concerned is identical to the standard that Reynolds v. Sims specifically made applicable to controversies over state legislative apportionment. See Reynolds v. Sims, supra, at 577. See also Hadley v. Junior College District, 397 U.S. 50, 56 (1970). And the holding in Kirkpatrick that the State must bear the burden of justifying deviations from population equality not only rested squarely and exclusively on our holding in Swann v. Adams, but even defined the test by quotation from Swann. See Kirkpatrick v. Preisler, supra, at 532.

In Swann v. Adams we held that variations in the population of legislative districts must be justified by the State by presentation of "acceptable reasons for the variations." 385 U. S., at 443. And a comparison of the opinion for the Court in Swann with the views expressed by two Justices in dissent, see Swann v. Adams, supra, at 447-448 (Harlan, J., dissenting), decisively refutes any suggestion that unequal representation will be upheld so long as some rational basis for the discrimination can be found. A showing of necessity, not rationality, is what our decision in Swann requires.

If Swann does not establish the point with sufficient clarity, then surely our decision in Kilgarlin v. Hill, 386 U.S. 120 (1967), where we elucidated and applied the principles of Swann, removes all doubt. There, the District Court had sustained the state apportionment plan on two grounds, one of which we termed a "burden of proof" ruling. The lower court held that appellants "had the burden not only of demonstrating the degree of variance from the equality principle but also of `negat[ing] the existence of any state of facts which would sustain the constitutionality of the legislation.' 252 F.Supp. 404, 414." Id., at 122. We squarely rejected that statement of the controlling legal standard, and held that under Swann v. Adams, "it is quite clear that unless satisfactorily justified by the court or by the evidence of record, population variances of the size and significance evident here [a total deviation of 26.48%] are sufficient to invalidate an apportionment plan." Ibid. We also rejected the District Court's second ground of decision: namely, that the deviations were amply justified by the State's attempt, wherever possible, to respect county boundaries. Significantly, the opinion stated that "[w]e are doubtful . . . that the deviations evident here are the kind of `minor' variations which Reynolds v. Sims indicated might be justified by local policies counselling the maintenance of established political subdivisions in apportionment plans. 377 U.S. 533, 578-579. But we need not reach that constitutional question, for we are not convinced that the announced policy of the State of Texas necessitated the range of deviations between legislative districts which is evident here." Id., at 123 (emphasis supplied).

III

I would affirm the District Court's decision because, on this record, the Commonwealth of Virginia failed—just as the State of Florida failed in Swann v. Adams and the State of Texas failed in Kilgarlin v. Hill—to justify substantial variations in the population of the districts from which members of the House of Delegates are elected. The panel that heard the case below consisted of four judges, all from Virginia, and I share their unanimous view that the Commonwealth failed to prove that the variations were justified by a need to insure representation of political subdivisions or a need to respect county boundaries in the drawing of district lines.

If variations in the population of legislative districts are to be upheld, the Court must determine, before turning to the justifications that are asserted in defense of the variations, that they are "free from any taint of arbitrariness or discrimination." Ante, at 325, quoting from Roman v. Sincock, 377 U. S., at 710. Appellees alleged before the District Court that the legislature's reapportionment plan did indeed discriminate against one region of the State—the Northern Virginia suburbs of Washington, D. C. Each House seat in Northern Virginia would be underrepresented by an average of 4.3% under the 1971 plan, and several would be underrepresented by as much as 6.3%. In view of what it termed the "pervasive under-representation in districts in Northern Virginia," 330 F. Supp., at 1146, the District Court ordered the transfer of one delegate out of the systematically overrepresented Tidewater region and into Northern Virginia.

In Abate v. Mundt, supra, at 185-186, we pointed out that we have

The District Court found as a fact that the 1971 plan did include a "built-in bias tending to favor [a] particular geographic area." Conveniently, the Court discerns no need even to acknowledge this critical finding of fact, and sets it aside without explanation. We have no basis for concluding that the finding is clearly erroneous, and that finding requires an affirmance of the District Court's decision without regard to the Commonwealth's asserted justifications for the inequalities in district population.

But even assuming that the Commonwealth's plan can be considered free of any "taint of arbitrariness or discrimination," appellants have failed to meet their burden of justifying the inequalities. They insist that the legislature has followed a consistent practice of drawing district lines in conformity with county boundaries. But a showing that a State has followed such a practice is still a long step from the necessary showing that the State must follow that practice. Neither in the Virginia Constitution nor in any Act of the Assembly has Virginia explicitly indicated any interest in preserving the integrity of county lines or in providing representation of political subdivisions as political subdivisions. Cf. Reynolds v. Sims, supra, at 580-581. On the contrary, the Constitution establishes a single standard for both legislative apportionment and congressional districting, and that standard requires only that lines be drawn so as to insure, "as nearly as is practicable," representation in proportion to population.8 And the origins of the constitutional provision make clear that equality in district population, not the representation of political subdivisions, is the Commonwealth's pre-eminent goal.9

Moreover, in asserting its interest in preserving the integrity of county boundaries, the Commonwealth offers nothing more than vague references to "local legislation," without describing such legislation with precision, without indicating whether such legislation amounts to a significant proportion of the legislature's business, and without demonstrating that the District Court's plan would materially affect the treatment of such legislation.10

The Court assumes that county representation is an important goal of Virginia's reapportionment plan, ante, at 326-328, and appellants suggest that the plan can be justified, at least in part, by the effort "to give an independent voice to the cities and counties [the legislature] daily governs." Brief for Appellant Commonwealth of Virginia 33. If county representation is indeed the Commonwealth's goal, then the apportionment plan adopted in 1971 itself falls far short of that objective. Appellants describe the problem in the following terms:

That argument assumes that some significant number of issues will have an impact squarely on Wythe County, while having no impact, or a differing impact, on the surrounding areas. For on issues affecting the entire region or the Commonwealth as a whole—presumably the vast majority of issues—the critical concern is not that each vote in Wythe County be cast in a single district, but that each vote cast be precisely equal in weight to votes in every other part of the Commonwealth. And the argument also assumes that the issues affecting only one county are of predominant concern to the voters. Under a representative form of government, the voters participate indirectly through the election of delegates. It should be obvious that as a voter's concern with regional or statewide issues increases relative to his interest in county issues, the significance of voting outside the county will correspondingly diminish.

But even if a substantial number of issues do have an impact primarily on a single county, and even if those issues are of deep concern to the voters, it still does not follow that the legislature's apportionment plan is a rational attempt to serve an important state interest. The plan would by no means provide, even in the legislature's own terms, effective representation for each county. Thus, the fourth legislative district, which would elect one delegate under the 1971 plan, consists of Wythe, Grayson, and Bland Counties along with the city of Galax. Yet Wythe County alone, according to appellants' figures, comprises 22,139 of the 49,279 persons resident in the district. Since Wythe County makes up almost one-half of the population of the fourth district, the district's delegate is likely to champion Wythe County's cause should an issue arise that pits its interest against the interests of Grayson or Bland County or the city of Galax.

In short, the best that can be said of appellants' efforts to secure county representation is that the plan can be effective only with respect to some unspecified but in all likelihood small number of issues that affect a single county and that are overwhelmingly important to the voters of that county; and even then it provides effective representation only where the affected county represents a large enough percentage of the voters in the district to have a significant impact on the election of the delegate.11 But even if county representation were, in fact, a strong and legitimate goal of the Commonwealth, and even if the 1971 plan did represent a rational effort to serve that goal, it is still not clear that the legislature's plan should be upheld. The plan prepared by the District Court would achieve a much higher degree of equality in district population, and it would accomplish that salutary goal with minimal disruption of the legislature's effort to avoid fragmenting counties. Of the 134 political subdivisions in the Commonwealth, only 12 would be divided by the District Court's plan. More significant, the number of persons resident in voting districts that would be cut out of one county or city and shifted to another is 64,738, out of the total state population of 4,648,494. Thus, even making each of the logical and empirical assumptions implicit in the view that violating county lines would effectively disenfranchise certain persons on certain local issues, the number of persons affected would still be less than 1 1/2% of the total state population.

IV

On this record—without any showing of the specific need for county representation or a showing of how such representation can be meaningfully provided to small counties whose votes would be submerged in a multi-county district—I see no basis whatsoever for upholding the Assembly's 1971 plan and the resulting substantial variations in district population. Accordingly, I would affirm the judgment of the District Court holding the plan invalid under the Equal Protection Clause of the Fourteenth Amendment.