BUTZNER, Circuit Judge:

In these consolidated cases, several counties, organizations, and individuals challenge Virginia Acts of Assembly, 1981 Special Session, Chapter 12, August 11, 1981 (hereafter referred to as the Act of August 11), which reapportioned the electoral districts for the House of Delegates.

The defendants are a number of Virginia officials, including those who have responsibility for conducting the State's elections. Various intervenors join the defendants in supporting the entire Act or specific provisions pertaining to their counties and cities.

The plaintiffs and their supporting intervenors attack the Act on one or more of the following grounds:

It violates the Equal Protection Clause of the Fourteenth Amendment because it does not provide for substantial population equality in electoral districts;

It violates the Fourteenth and Fifteenth Amendments because it invidiously discriminates against Virginia's black citizens;

It violates Article IV, § 4 of the United States Constitution by denying Virginians their right to a republican form of government;

It violates the Equal Protection Clause of the Fourteenth Amendment because neither the Census Bureau nor the General Assembly counted as residents of Norfolk, the home port of their fleet, approximately 9,000 naval personnel deployed at sea.

It violates Article II, § 6 of the Virginia Constitution which provides: "Every 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."

One or more of the plaintiffs and intervenors seek the following relief:

A declaration that the Act of August 11, 1981, is unconstitutional;

An injunction prohibiting the State Board of Elections from conducting an election in 1981 for members of the House of Delegates on the basis of the August 11 Act;

An order requiring the 1981 elections to be conducted on the basis of the 1971 apportionment Act;

An order reapportioning the State into single-member districts;

An injunction requiring the State to include in the population of Norfolk approximately 9,000 naval personnel deployed at sea.

Several parties have recommended alternative redistricting plans. Several intervenors request redistricting of specific counties, leaving the remainder of the August 11 Act intact.

Because we conclude that the Act violates the Equal Protection Clause of the Fourteenth Amendment and we have afforded appropriate relief, we find it unnecessary to discuss in detail each specific complaint.

I

Article II, § 6 of the Virginia Constitution requires the boundaries of its Senate and House of Delegates electoral districts to be redrawn every ten years. The House Privileges and Elections Committee had primary responsibility for the development of a redistricting plan. After receipt of the 1980 census figures in February, 1981, the Committee scheduled a series of public hearings on redistricting throughout the state. Starting on March 20, the Committee considered at least a dozen redistricting plans, including some that proposed a more even distribution of population in each district than the plan eventually enacted.

On April 7, 1981, the Committee reported a plan to the House of Delegates. During the floor debate on April 8, objections were raised charging departures from the one-person, one-vote ideal and dilution of minority voting strength. Nonetheless, the bill was passed by the House. Following Senate approval, the Governor signed the bill on April 10, 1981.

The Act then was submitted to the Attorney General of the United States on April 30 for the preclearance required by § 5 of the Voting Rights Act of 1965. On August 1, the Attorney General notified the State that the Act could not be given clearance because some districts in Southside Virginia, "appear[ed] to dilute and fragment black voting strength unnecessarily."

The leadership of the House of Delegates and representatives of various interested parties promptly held a series of meetings with Justice Department officials. Eventually, the State and the Department reached an informal understanding of the changes necessary for approval, and the Department agreed to review expeditiously any new enactment of the General Assembly.

On August 11, the Committee held another public hearing at which plans likely to increase the black membership of the House of Delegates through the use of single-member districts were presented. The Committee rejected these plans and approved one incorporating the changes discussed with the Department of Justice. Later the same day, the House also rejected a single-member plan and adopted the Committee plan. The Senate promptly approved the House bill and the Governor signed it.

The Act of August 11, 1981, was reviewed by the Department of Justice on August 12. Late that afternoon, the Department notified the State that "the Attorney General does not interpose any objections to the changes [made on August 11]." The Attorney General, however, reserved the right to object within 60 days if new evidence cast matters in a different light.

The August 11 Act provides for 49 electoral districts embracing Virginia's 95 counties and 41 incorporated cities with an aggregate population of 5,346,279. With a 100-delegate House, the ideal population for each single-member district is therefore 53,463 citizens. The amount of deviation from the ideal in any one district is typically measured in percentage terms. The maximum statewide deviation is the sum of the deviations of the two districts with the greatest deviations above and below the ideal.

When the ratio of citizens to delegates in two or more adjacent districts varies greatly from the ideal, the deviation can be corrected either by redrawing the district lines or by establishing an additional "floater" district encompassing the two or more underlying districts. Floater districts provide for one or more at-large delegates to represent all of the underlying districts.

By reducing the number of floater districts, the August 11 Act virtually moots earlier disputes concerning the proper method for calculating deviations from the ideal in floater districts. By the traditional House method, the August 11 Act has a maximum deviation of 26.63%, the total of district 42 (-14.16%) and district 3 (+ 12.47%). By the shared floater method, that total rises to 27.72%, the sum of the deviations in districts 42 and 46 (14.16% + 13.56%).

Not counting the geographically isolated district 42, the Eastern Shore counties, the plan has a total deviation of 22.13% by the House method and 25.01% by the shared floater method. The average deviation is ±4.90%, with 20 districts exceeding ±5%, by the House method. By the shared floater method, the average is ±5.26%, and 22 districts exceed ±5%.

II

A preliminary issue involves this court's authority to decide the constitutional controversy presented in these cases. In McDaniel v. Sanchez, ___ U.S ___, ___, 101 S.Ct. 2224, 2238, 68 L.Ed.2d 724 (1981), and Connor v. Waller, 421 U.S. 656, 95 S.Ct. 2003, 44 L.Ed.2d 486 (1975), the Court held that a district court should defer consideration of an apportionment plan drawn by a legislative body until after the plan has been reviewed by the Attorney General pursuant to § 5 of the Voting Rights Act.

Section 5 of the Voting Rights Act, 42 U.S.C. § 1973c, was amended in 1975 to enable expedited approval within the 60 days allowed for the Attorney General's review. The Act permits the Attorney General to indicate affirmatively that no objection will be made, reserving the right, nevertheless, to reexamine the submission if additional information comes to his attention before 60 days have elapsed. The Attorney General has implemented this statutory provision by adopting a regulation to be codified as 28 C.F.R. § 51.42. See 46 Fed.Reg. 878 (1981).

The Attorney General's statement of August 12, 1981, that he "does not interpose any objections to the changes" in the August 11 Act specifically noted that he was acting pursuant to 28 C.F.R. § 51.42. Accordingly, he reserved the right to reexamine the submission if additional information came to his attention within the remainder of the 60 day period.

Section 5 of the Voting Rights Act states that the purpose of the procedure the Attorney General followed is "to facilitate an expedited approval within sixty days after [the state's] submission. . . ." Deferral of our consideration of the constitutional questions presented by these cases until the expiration of the remainder of the 60 day period allowed the Attorney General would frustrate this salutary provision of the Act. We therefore conclude that the Attorney General's review, though subject to reexamination, lifts the restriction that McDaniel and Connor, supra, would otherwise place on our authority to decide these cases.

III

In Reynolds v. Sims, 377 U.S. 533, 568, 84 S.Ct. 1362, 1385, 12 L.Ed.2d 506 (1964), the Court held that the "Equal Protection Clause requires that the seats in both houses of a bicameral state legislature must be apportioned on a population basis." To implement this constitutional requisite, the Court explained, a state must "make an honest and good faith effort to construct districts . . . as nearly of equal population as is practicable." 377 U.S. at 577, 84 S.Ct. at 1390. This, too, is the command of Article II, § 6 of the Virginia Constitution. Recognizing that strict mathematical equality is not imposed on state legislatures by the Constitution, the Court held that some deviations from the ideal are permissible if they "are based on legitimate considerations incident to the effectuation of a rational state policy. . . ." 377 U.S. at 579, 84 S.Ct. at 1391.

Virginia urges that the population deviations in the August 11 Act are justified by the General Assembly's consideration of legitimate state interests. The State's principal witness described the policies or guidelines considered by the House Committee in drafting a reapportionment plan as follows:

I think we should start with the Constitutional requirement or mandate, that the district be as equal as practicable in population.

The second guideline, and after population, certainly I think the most important to the committee, was the preservation of local subdivisions as entities. That is, not violate the integrity of the boundaries of the districts.

A third guideline, or a third aim of the committee within the first two was to create as many single-member districts as they could, where that opportunity presented itself, and where the choices were such, they would generally opt for the single-member district.

The committee felt that they could seek to retain existing legislative districts, as far as it was consistent with the population standards.

The committee felt that incumbency concerns could be taken into account. The whole area of communities of interest, natural boundaries, and those sort of factors that helped to define the district could be considered.

The committee, as it did its work, I believe, applied a general rule that multi-member districts, especially in the rural areas, should not be too big. Too big is something that is hard to define, but the committee was concerned about creating excessively large geographical districts, taking in a large number of localities in rural areas.

At one time or another, the Supreme Court has expressly or tacitly recognized that most of the policies considered by Virginia permit deviation from strict numerical equality. The Court has never held, however, that these state interests, though legitimate, may be aggregated to justify population variances as large as those disclosed by the Virginia Act.

Indeed, dicta that we deem persuasive are to the contrary. The Court dealt with Virginia's 1971 reapportionment statute in Mahan v. Howell, 410 U.S. 315, 93 S.Ct. 979, 31 L.Ed.2d 793 (1973). There, recognizing that Virginia had a rational interest in preserving the integrity of political subdivisions, the Court held that a deviation of 16.4% did not exceed constitutional limits. Nevertheless, the Court cautioned that "this percentage may well approach tolerable limits. . . ." 410 U.S. at 329, 93 S.Ct. at 987.

Later the Court reviewed the principles that have evolved from its reapportionment decisions. Referring to the upper limits of permissible deviations, the Court said in Gaffney v. Cummings, 412 U.S. 735, 744, 93 S.Ct. 2321, 2327, 37 L.Ed.2d 298 (1973):

As these pronouncements have been worked out in our cases, it has become apparent that the larger variations from substantial equality are too great to be justified by any state interest so far suggested. There were thus the enormous variations struck down in the early cases beginning with Reynolds v. Sims, as well as the much smaller, but nevertheless unacceptable deviations, appearing in later cases such as Swann v. Adams, 385 U.S. 440, 87 S.Ct. 569, 17 L.Ed.2d 501 (1967); Kilgarlin v. Hill, 386 U.S. 120, 87 S.Ct. 820, 17 L.Ed.2d 771 (1967); and Whitcomb v. Chavis, 403 U.S. 124, 161-163, 91 S.Ct. 1858, 1878-1879, 29 L.Ed.2d 363 (1971).

The population variations that the Court deemed "too great to be justified by any state interest so far suggested" were Swann v. Adams (25.65%), Kilgarlin v. Hill (26.48%), and Whitcomb v. Chavis (24.78%). Virginia's deviation is comparable to these.

We cannot accept Virginia's argument that the Supreme Court invalidated reapportionment plans with deviations in the neighborhood of 25% simply because the states involved in those cases did not articulate peculiar interests justifying departure from practicable equality. It is true that in Swann v. Adams, 385 U.S. 440, 87 S.Ct. 569, 17 L.Ed.2d 501 (1967), Florida presented no acceptable reason for a deviation of 25.65%. In Kilgarlin v. Hill, 386 U.S. 120, 87 S.Ct. 820, 17 L.Ed.2d 771 (1967), where the deviation was 26.48%, Texas similarly failed to articulate policies sufficient to justify the variations among the populations of the legislative districts. Plainly the Court found it unnecessary in these cases to address the question whether any policy could justify such large deviations.

In a subsequent case, however, the Court dealt with the preservation of subdivision boundaries, the most important policy Virginia presses. In Whitcomb v. Chavis, 403 U.S. 124, 91 S.Ct. 1858, 29 L.Ed.2d 363 (1971), the Court referred to the following conclusion of the district court: "It may not be possible for the Indiana general assembly to comply with the state constitutional requirement prohibiting crossing or dividing counties for senatorial apportionment and still meet the requirements of the Equal Protection Clause. . . ." 403 U.S. at 136, 91 S.Ct. at 1865. Without rejecting the district court's conclusion, the Supreme Court, nevertheless, affirmed an order requiring reapportionment. There the deviations were 28.20% for the Senate districts and 24.78% for the House. The ratio of the largest to the smallest Senate districts was 1.327 to 1 and for the House, 1.279 to 1. This evidence, the Court said, established a "convincing showing of malapportionment. . . ." 403 U.S. at 162, 91 S.Ct. at 1879.

The deviations and ratios that the Court found unacceptable in Whitcomb are comparable to those disclosed in the Virginia reapportionment Act.

Guided by the precepts that the Supreme Court has explained in the cases we have cited, we conclude that the Act of August 11 is facially unconstitutional because the deviation among the populations of the districts that it creates exceeds the limits tolerated by the Equal Protection Clause.

Our decision does not hinge on which method is used to calculate the deviation in the Act's single floater district or on whether the Eastern Shore counties should be excluded in computing the deviation among the populations of the districts. The deviations in the August 11 Act range from 22.13% to 27.72%, and the ratios of largest to smallest districts range from 1.24:1 to 1.31:1, depending on the factors of the calculus. Even accepting the figures most favorable to the State, we believe that these variances are too large. The Supreme Court has not held a reapportionment statute with a deviation of this magnitude to be constitutional.

IV

Apart from the facial unconstitutionality of the Virginia Act, we find that the State's announced policies do not necessitate or justify the deviations among the populations of the legislative districts. This deficiency is a separate and independent reason for holding the Act unconstitutional. Kilgarlin v. Hill, 386 U.S. 120, 123, 87 S.Ct. 820, 822, 17 L.Ed.2d 771 (1967).

The primary policy asserted by the state to justify the population deviations is the preservation of the integrity of political subdivision boundaries. Quite naturally the State relies on Mahan v. Howell, 410 U.S. 315, 93 S.Ct. 979, 35 L.Ed.2d 320 (1973), which held this policy to be a legitimate justification for a deviation of 16.4%.

In Mahan, the Court relied on uncontradicted evidence that it was impossible to reduce population disparities and still maintain the integrity of county and city boundaries. 410 U.S. at 319-20, 326, 93 S.Ct. at 982-83, 986. The significance of this fact was emphasized in Connor v. Finch, 431 U.S. 407, 421, 97 S.Ct. 1828, 1837, 52 L.Ed.2d 465 (1977). There the Court rejected a district court's reapportionment plan because, among other reasons, an alternative plan served Mississippi's policy against fragmenting county boundaries and yet came closer "to achieving districts that are `as nearly of equal population as is practicable.'" 431 U.S. at 420, 97 S.Ct. at 1836. Distinguishing Mahan, the Connor Court said:

Under the less stringent standards governing legislatively adopted apportionments, the goal of maintaining political subdivisions as districts sufficed to justify a 16.4% population deviation in the plan for the Virginia House of Delegates. Mahan v. Howell, 410 U.S. 315, 93 S.Ct. 979, 35 L.Ed.2d 320. But in Mahan, there was uncontradicted evidence that the legislature's plan "`produces the minimum deviation above and below the norm, keeping intact political boundaries.'" 431 U.S. at 420, 97 S.Ct. at 1836.

Here, in contrast to the situation the Court found decisive in Mahan, alternative plans presented to the Committee and alternatives tendered by the litigants demonstrate that it is now possible to maintain the integrity of Virginia's political subdivisions while substantially reducing the variations in population among the legislative districts. We deem it beyond our province to press the General Assembly to adopt any of these plans. They illustrate, however, our conclusion that Mahan does not provide controlling precedent for the decision of these consolidated cases.

Moreover, the 1971 Act approved in Mahan differs significantly from the August 11 Act. Calculated by the State's method, these differences may be summarized as follows:

1971 Act August 11, 1981 Act

1. Maximum Population Variance 16.4% 26.63%

2. Maximum Population Variance (Excluding Eastern Shore) 16.4% 22.13%

3. Average Deviation from Ideal Size ±3.89% ±4.90%

4. Average Deviation from Ideal Size (Excluding Eastern Shore) ±3.49% ±4.69%

5. Number of Districts Exceeding 5% Deviation 16 20

6. Number of Districts Exceeding 5% Deviation (Excluding Eastern Shore) 15 19

7. Ratio of Largest to Smallest Districts 1.18 to 1 1.31 to 1

8. Ratio of Largest to Smallest Districts (Excluding Eastern Shore) 1.18 to 1 1.24 to 1

9. Percent of Population Able to Elect a Majority of Delegates 49.2% 48.97%

Thus, tested by every measure, the August 11 Act departs to a much greater degree than the 1971 Act from the goal of fashioning legislative districts of substantially equal population.

The State also announced a policy of creating single-member districts whenever the opportunity to do so presented itself. The Supreme Court has indicated on numerous occasions that this is an appropriate consideration. Indeed, the Court has directed district courts that are obliged to draw their own reapportionment plans to follow this policy in the absence of exceptional circumstances. See Chapman v. Meier, 420 U.S. 1, 26-27, 95 S.Ct. 751, 765-766, 42 L.Ed.2d 766 (1975).

Virginia appears to have given this policy a low priority, because the state departed from it frequently. Of the 31 multi-member districts created by the Act, 10 are composed of a single political subdivision. These 10 elect 30 delegates. Thus, without transgressing the boundary of any political subdivision, the General Assembly could have created 30 additional single-member districts. If the three-member districts, 51 and 52, allocated to parts of Fairfax County were also subdivided, the number of single-member districts would increase to 36.

In view of this departure from the state's announced policy, we cannot accept the claim that the aim of creating single-member districts "where that opportunity presented itself" justifies the excessive population deviations disclosed by the Act.

The State argues that all the proposed plans that reduce population variance while maintaining the integrity of subdivision boundaries infringe the State's other interests. This argument may be true, but the other interests pressed by the State cannot singly or in the aggregate justify sacrificing population equality to the extent evidenced by the Virginia Act.

Of the other policies considered by the legislature, the record discloses that the State most assiduously pursued the goals of retaining existing legislative districts, preserving incumbents in office, and recognizing communities of interest. Indeed, the evidence persuades us that the State gave undue priority to these concerns.

Retention of existing legislative districts is unobjectionable when achieved without creating significant population variances. It is, however, essentially a policy of maintaining the status quo. It therefore inherently conflicts with Virginia's Constitution and federal standards requiring decennial reapportionment to alter the status quo by responding to shifts in the State's population.

Preservation of incumbency interests is closely related to retention of legislative districts and is subject to the same limitations. It, too, maintains the status quo at the expense of meeting constitutional mandates for a reapportionment that reflects changes in population. The Supreme Court has noted: "The fact that district boundaries may have been drawn in a way that minimizes the number of contests between present incumbents does not in and of itself establish invidiousness." Burns v. Richardson, 384 U.S. 73, 89 n. 16, 86 S.Ct. 1286, 1295 n. 16, 16 L.Ed.2d 376 (1966). Thus, recognition of incumbency concerns is not in itself unconstitutional. White v. Weiser, 412 U.S. 783, 791, 93 S.Ct. 2348, 2352, 37 L.Ed.2d 335 (1973). But the Court has never held that this policy justifies large disparities in district populations.

Virginia also relies on its policy of recognition of communities of interest as a justification for the Act's population deviations. To support the allocation of delegates, the state's principal witness prepared a map dividing Virginia into seven geographic regions. He then showed that the regional allocations were substantially equal despite population inequalities between the several districts composing each region.

We cannot accept this post hoc analysis as a justification for the districts' population variances. The record shows that the regional lines reflect only the judgment of the witness and not that of the legislature. It is difficult to assume that each county on the border of a region fits precisely in the region to which it is assigned and not to a neighboring region. Undoubtedly, the people in each large region share a somewhat tenuous community of interest, but their regional concerns have not been shown to eclipse either local or statewide concerns. Moreover, delegates are elected by district, not by region. The population of the districts, not that of the regions, must therefore be apportioned to achieve practicable equality.

Virginia's emphasis on community of interest also fails to heed one of the most elementary principles set forth in Reynolds v. Sims, 377 U.S. 533, 84 S.Ct. 1362, 12 L.Ed.2d 506 (1964). There the Court, after recognizing that rational state policies — including the maintenance of natural and historical boundaries — could warrant some population deviations, cautioned: "But neither history alone, nor economic or other sorts of group interests, are permissible factors in attempting to justify disparities from population-based representation. Citizens, not history or economic interests, cast votes." 377 U.S. at 579-80, 84 S.Ct. at 1391.

It may be that policies of retaining existing districts, expressing concern for incumbents, and recognizing communities of interest would not invalidate a plan with a deviation slightly in excess of 10%. But we are not persuaded that these policies, singly or in combination, can justify the deviations disclosed by the Act of August 11.

In sum, in addition to its facial invalidity, the Act of August 11, 1981, violates the Equal Protection Clause and the Virginia Constitution because the State's announced policies either do not necessitate, or are not adequate to justify, the Act's population variances.

V

Aligning themselves with the parties who challenge the Act because it violates the Equal Protection Clause regardless of race, several parties protest that the August 11 Act violates the rights of black citizens secured by the Fourteenth and Fifteenth Amendments. They assail generally the Act's creation of numerous multi-member districts as establishing both the purpose and effect of diluting black voting strength. They specifically attack the composition of districts 12 and 13.

In City of Mobile v. Bolden, 446 U.S. 55, 100 S.Ct. 1490, 64 L.Ed.2d 47 (1980), the plurality opinion of the Court reviewed the limitations imposed on the states by the Fourteenth and Fifteenth Amendments with respect to the composition of legislative districts. At issue was whether an at-large election of a city commission violated the rights of black voters. With respect to the Fourteenth Amendment, the Court adverted to the basic principle that a violation can be established only by showing purposeful discrimination. Applying this principle, the Court said:

Despite repeated constitutional attacks upon multimember legislative districts, the Court has consistently held that they are not unconstitutional per se . . . We have recognized, however, that such legislative apportionments could violate the Fourteenth Amendment if their purpose were invidiously to minimize or cancel out the voting potential of racial or ethnic minorities. . . . To prove such a purpose it is not enough to show that the group allegedly discriminated against has not elected representatives in proportion to its numbers. . . . A plaintiff must prove that the disputed plan was "conceived or operated as [a] purposeful devic[e] to further racial . . . discrimination." 446 U.S. at 66, 100 S.Ct. at 1499.

The Supreme Court has observed that criticism of multi-member districts is based in part on "their tendency to submerge minorities. . . ." Whitcomb v. Chavis, 403 U.S. 124, 159, 91 S.Ct. 1858, 1877, 29 L.Ed.2d 363 (1971). Nevertheless, the Court has consistently held that multi-member districts are not per se unconstitutional. Whitcomb, 403 U.S. at 159, 160, 91 S.Ct. at 1877, 1878; Mobile, 446 U.S. at 66. It may well be, as the plaintiffs claim, that multi-member districts will limit the number of black citizens elected to the House of Delegates. Neither the Fourteenth nor Fifteenth Amendment, however, guarantees a racial group the right to elect its candidates in proportion to its voting potential. Mobile, 446 U.S. at 65, 100 S.Ct. at 1499; White v. Regester, 412 U.S. 755, 766, 93 S.Ct. 2332, 2339, 37 L.Ed.2d 314 (1973).

The Court has repeatedly declared that "official action will not be held unconstitutional solely because it results in a racially disproportionment impact." Arlington Heights v. Metropolitan Housing Development Corp., 429 U.S. 252, 264-65, 97 S.Ct. 555, 562-63, 50 L.Ed.2d 450 (1977). Adverse racial effect often provides the initial step to proving a violation of the Fourteenth Amendment, but it must be accompanied by proof of purposeful discrimination. Such proof is lacking when the use of multi-member districts can be explained on grounds other than race. In the absence of other evidence showing that black citizens are denied access to the political process, a multi-member reapportionment plan is not unconstitutional. See Mobile, 446 U.S. at 70, 100 S.Ct. at 1501.

Here the evidence discloses that the General Assembly indiscriminately created many multi-member districts both in areas containing a concentration of black citizens and in those predominately composed of white residents. Multi-member districts were created, the evidence reveals, to further the announced policies of the State that we previously discussed. Although, as we have mentioned, the State could have created more single-member districts without crossing the boundaries of political subdivisions, the allocation of delegates to multi-member districts without regard to their racial composition negates the claim of purposeful discrimination. Moreover, unlike the situation depicted in White v. Regester, 412 U.S. 755, 765-67, 93 S.Ct. 2332, 2339-40, 37 L.Ed.2d 314 (1973), the evidence does not disclose that black voters have less opportunity than other residents of multi-member districts "to participate in the political processes and to elect legislators of their choice." 412 U.S. at 766, 93 S.Ct. at 2339.

The complaints about districts 12 and 13 are best summarized by quoting from the brief submitted by one of the plaintiffs:

District 12. Old District 13, composed of Patrick, Henry, and Pittsylvania Counties and the City of Martinsville, was 25.41% black (1980 Census). This district was restructured in the 1981 plan, and Patrick and Henry Counties and the City of Martinsville were joined with Floyd County, which is only 3.32% black. This realignment dilutes black voting strength in this area, and results in new District 12, which is only 19.95% black.

. . . .

District 13. District 14 in the 1971 plan, composed of Danville, was 29.74% black. In the 1981 plan Danville and Pittsylvania County, which also is 30% black, are combined with Campbell County, which is only 15.10% black. This combination reduces the black percentage of the Danville district from 29.74% to 25.71%.

The evidence discloses that debates in the House expressly adverted to the dilution of minority voting strength in both of these districts.

Realignment of the counties, however, only minimally dilutes black voting strength in the districts. In neither district was a black majority submerged by reapportionment. Moreover, the evidence discloses that the populations of each of the two districts deviates from the ideal by less than 2%. In view of these facts, we find no purposeful discrimination with respect to these districts that would violate the Fourteenth Amendment.

The Fifteenth Amendment provides: "The right of citizens of the United States to vote shall not be denied or abridged by the United States or by any State on account of race, color, or previous condition of servitude." Reiterating principles drawn from its apportionment cases, the Court held that "action by a State that is racially neutral on its face violates the Fifteenth Amendment only if motivated by a discriminatory purpose." Mobile, 446 U.S. at 62, 100 S.Ct. at 1497.

The evidence shows that black citizens have not been denied the right to vote in Virginia since 1965. Contemporary practices and procedures do not abridge their right to register and vote.

We therefore conclude that the August 11 Act does not violate either the Fourteenth or Fifteenth Amendment on the basis of racial discrimination. See Mobile, 446 U.S. at 65, 66-74, 100 S.Ct. at 1499, 1500-1503.

VI

Copeland offers no proof, however, of the voting residence of these 9,000 citizens. Even the figure itself, based on Navy Department records and not on an actual Census Bureau enumeration, is suspect. Copeland's exhibit No. 1, a letter from the Census Bureau, notes that, based on the Bureau's experience with nondeployed ships, "[t]he number of persons actually aboard ship can deviate considerably from Navy Department figures. . . ." We find, in any event, that 9,000 additional residents would have but slight effect on the Norfolk and Virginia Beach populations and on the calculations of deviation from the ideal district populations.

Accordingly, we decline to find that the General Assembly's reliance on the 1980 census figures denied Copeland equal protection of the laws.

VII

Having found the August 11 plan unconstitutional, we must consider the question of appropriate relief. Any remedy must, of course, be considered in light of the imminence of the 1981 elections. The primary election is scheduled for September 8; the general election for November 3. The deadline for candidates to file for election has already passed. Certification of candidates, ballot preparation, and a host of other election mechanics must be undertaken promptly if the existing schedule is to be followed.

A number of remedies have been suggested. First, the court could implement its own plan meeting the stringent requirements of Chapman v. Meier, 420 U.S. 1, 26-27, 95 S.Ct. 751, 765-766, 42 L.Ed.2d 766 (1975), and Connor v. Finch, 431 U.S. 407, 417-20, 97 S.Ct. 1828, 1835-1836, 52 L.Ed.2d 465 (1977). Such a plan might be based on one of the parties' single-member plans. Second, we could permit the Virginia General Assembly to devise a new plan of its own. Third, we could direct the upcoming elections to proceed under the August 11 Act as an interim remedy. In this event, we could either allow the newly-elected delegates to serve their usual two-year terms, as urged by the state defendants, or require them to serve only one year with new elections set for 1982, as suggested by one of the intervenors. And fourth, we could order the elections to be reorganized to follow the 1971 district lines.

We find that devising a court-ordered plan would be time-consuming and would substantially delay the November elections. Adopting a party's plan would save some time but would require a drastic change in voting mechanics, also delaying the elections. A legislatively redrawn plan would suffer from the same drawbacks.

The 1971 Act is substantially out of date because Virginia's population has grown some 15%, with that growth unevenly spread throughout the Commonwealth. Allowing elections to proceed under the 1971 Act would greatly disadvantage the citizens in Virginia's rapidly growing areas and would effect great harm to the principle of one-person, one-vote.

November 3 is already fixed as the date for a statewide election for state officials including the Governor. Experience suggests that if the statewide election proceeds on November 3 but the House of Delegates election is postponed, voter turnout for the latter will be significantly lower than otherwise. We believe that a strong and representative turnout for the House election depends on holding it on November 3.

Consequently, we conclude that the August 11 Act should be continued in effect for the November election. Interim relief using an unconstitutional apportionment plan is permissible, when, as here, necessary election machinery is already in progress for an election rapidly approaching. Reynolds v. Sims, 377 U.S. 533, 585, 84 S.Ct. 1362, 1393, 12 L.Ed.2d 506 (1964); see also Kilgarlin v. Hill, 386 U.S. 120, 121, 87 S.Ct. 820, 821, 17 L.Ed.2d 771 (1967); Toombs v. Fortson, 241 F. Supp. 65, 71 (N.D.Ga. 1965), aff'd per curiam, 384 U.S. 210, 86 S.Ct. 1464, 16 L.Ed.2d 482 (1966).

Whenever possible, of course, a state legislature should have an opportunity to redraw a plan found by the courts to be unconstitutional. Reynolds v. Sims, 377 U.S. at 586-87, 84 S.Ct. at 1394; Wise v. Lipscomb, 437 U.S. 535, 539-40, 98 S.Ct. 2493, 2496-97, 57 L.Ed.2d 411 (1978). Although we have found that it would be impractical for the General Assembly to devise a plan that would accommodate an election on November 3, we believe that that body can constitutionally reapportion the State in the period of five months ending February 1, 1982. If this is not done, we will be compelled to consider drafting a court plan. We retain jurisdiction for this purpose. The State must forthwith submit any new Act to the Attorney General pursuant to the Voting Rights Act of 1965. See McDaniel v. Sanchez, ___ U.S. ___, ___ n. 35, 101 S.Ct. 2224, 2238 n. 35, 68 L.Ed.2d 724 (1981). We have selected the deadline of February 1, 1982, because the history of this case shows that a later date will interfere with the State's general election schedule.

Because Virginia's citizens are entitled to vote as soon as possible for their representatives under a constitutional apportionment plan, we will limit the terms of members of the House of Delegates elected in 1981 to one year. We also will direct the state election officials to conduct a new election in 1982 for the House of Delegates under the General Assembly's new Act or our own plan. That election should be held the same day as the November general election. See Va. Code § 24.1-1(5)(a). Delegates elected then shall serve for the remainder of the 1982-84 term unless the General Assembly chooses to extend the term to a full two years.

VIII

At least six of the parties involved in this case have requested that the court grant them reasonable attorney's fees incurred in this litigation pursuant to 42 U.S.C. § 1973 l(e) and 1988.

Ordinarily, all claims for relief, including a party's request for attorney's fees, must be decided by the district court before a judgment becomes final and ripe for appeal. Liberty Mutual Insurance Co. v. Wetzel, 424 U.S. 737, 742, 96 S.Ct. 1202, 1205, 47 L.Ed.2d 435 (1976). Interlocutory injunctions, however, present an exception to this rule because they are immediately appealable. 28 U.S.C. § 1253.

In this case the propriety of attorney's fees has not yet been briefed nor have any affidavits or other supporting evidence been submitted to this court to justify any award. Therefore, given the proximity of the upcoming elections and the importance of allowing any aggrieved party an opportunity for immediate appeal of the other issues in this case, we will not defer the entry of our interlocutory order to address the question of attorney's fees but will reserve consideration of this question for a later time.

A P P E N D I X 1981 HOUSE OF DELEGATES DISTRICTS ACTS OF ASSEMBLY, 1981 SPECIAL SESSION, CHAPTER 12 (August 11, 1981)

DELEGATES DISTRICT POPULATION DEVIATION

1. Dickenson 19,806 Lee 25,956 Norton 4,757 Scott 25,068 Wise 43,863 ------- 2 119,450 +11.71% ===============================================================

2. Bristol 19,042 Smyth 33,366 Washington 46,487 ------ 2 98,895 -7.51% ===============================================================

3. Buchanan 37,989 Russell 31,761 Tazewell 50,511 ------- 2 120,261 +12.47% ===============================================================

4. Bland 6,349 Grayson 16,579 Wythe 25,522 Galax 6,524 ------ 1 54,974 +2.83% ===============================================================

5. Carroll 27,270 Giles 17,810 Montgomery 63,516 Pulaski 35,229 Radford 13,225 ------- 3 157,050 -2.08% ===============================================================

6. Roanoke City 100,427

2 100,427 -6.08% ===============================================================

7. Craig 3,948 Roanoke County 72,945 Salem 23,958 ------- 2 100,851 -5.68% ===============================================================

8. Alleghany 14,333 Botetourt 23,270 Clifton Forge 5,046 Covington 9,063 ------ 1 51,712 -3.28% ===============================================================

9. Bedford County 34,927 Franklin County 35,740 Rockbridge 17,911 Buena Vista 6,717 Lexington 7,292 Bedford City 5,991 ------- 2 108,578 +1.54% ===============================================================

10. Augusta 53,732 Bath 5,860 Highland 2,937 Staunton 21,857 Waynesboro 15,329 ------ 2 99,715 -6.74% ===============================================================

11. Amherst 29,122 Nelson 12,204 Lynchburg 66,743 ------- 2 108,069 +1.07% ===============================================================

12. Floyd 11,563 Henry 57,654 Martinsville 18,149 Patrick 17,585 ------- 2 104,951 -1.85% ==============================================================

13. Campbell 45,424 Pittsylvania 66,147 Danville 45,642 ------- 3 157,213 -1.98% ===============================================================

14. Charlotte 12,266 Halifax 30,418 South Boston 7,093 ------ 1 49,777 -6.89% ===============================================================

15. Greene 7,625 Rockingham 57,038 Shenandoah 27,559 Harrisonburg 19,671 ------- 2 111,893 +4.65% ===============================================================

16. Frederick 34,150 Winchester 20,217 ------ 1 54,367 +1.69% ===============================================================

17. Loudoun 57,427

1 57,427 +7.41% ===============================================================

18. Clarke 9,965 Page 19,401 Rappahannock 6,093 Warren 21,200 ------ 1 56,659 +5.98% ===============================================================

19. Culpeper 22,620 Fauquier 35,889 ------ 1 58,509 +9.44% ===============================================================

20. Stafford 40,470 Fredericksburg 15,322 ------ 1 55,792 +4.36% ===============================================================

21. Alexandria 103,217

2 103,217 -3.47% ===============================================================

22. Arlington 152,599

3 152,599 -4.86% ===============================================================

23. Manassas 15,438 Manassas Park 6,524 Prince William 144,703 ------- 3 166,665 +3.91% ===============================================================

24. Albemarle 50,689 Fluvanna 10,244 Charlottesville 45,010 ------- 2 105,943 -0.92% ===============================================================

25. Amelia 8,405 Appomattox 11,971 Buckingham 11,751 Cumberland 7,881 Prince Edward 16,456 ------ 1 56,464 +5.61% ===============================================================

26. Nottoway 14,666 Lunenburg 12,124 Mecklenburg 29,444 ------ 1 56,234 +5.18% ===============================================================

27. Brunswick 15,632 Dinwiddie 22,602 Greensville 10,903 Emporia 4,840 Sussex 10,874 Petersburg 41,055 ------- 2 105,906 -0.95% ===============================================================

29. Goochland 11,761 Louisa 17,825 Madison 10,232 Orange 17,827 ------ 1 57,645 +7.82% ===============================================================

30. Caroline 17,904 Spotsylvania 34,435 ------ 1 52,339 -2.10% ===============================================================

32. Henrico 180,735 Hanover 50,398 ------- 4 231,133 +8.08% ===============================================================

33. Richmond City 219,214

4 219,214 +2.51% ===============================================================

34. Chesterfield 141,372 Powhatan 13,062 Colonial Heights 16,509 ------- 3 170,943 +6.58% ===============================================================

35. Prince George 25,733 Charles City 6,692 Hopewell 23,397 ------ 1 55,822 +4.41% ===============================================================

36. Chesapeake 114,226

2 114,226 +6.83% ===============================================================

37. Norfolk 266,979

5 266,979 -0.13% ===============================================================

38. Virginia Beach 262,199

5 262,199 -1.91% ===============================================================

39. Portsmouth 104,577

2 104,577 -2.20% ===============================================================

41. Suffolk 47,621 Franklin City 7,308 Isle of Wight 21,603 Southampton 18,731 Surry 6,046 ------- 2 101,309 -5.25% ===============================================================

42. Accomack 31,268 Northampton 14,625 ------ 1 45,893 -14.16% ===============================================================

43. King George 10,543 Lancaster 10,129 Northcumberland 9,828 Richmond County 6,952 Westmoreland 14,041 ------ 1 51,493 -3.68% ===============================================================

44. Essex 8,864 Gloucester 20,107 King and Queen 5,968 King William 9,327 Mathews 7,995 Middlesex 7,719 ------ 1 59,980 +12.19% ===============================================================

45. Hampton 122,617

2

46. James City 22,763 New Kent 8,781 -2.63% York 35,463 Poquoson 8,726 Williamsburg 9,870 ------ 1 85,603

1 47. Floater 45 and 46 208,220 ===============================================================

48. Newport News 144,903

3 144,903 -9.66% ===============================================================

49. Fairfax 147,310 Falls Church 9,515 ------- 3 156,825 -2.22% ===============================================================

50. Fairfax 137,783 Fairfax City 19,390 ------- 3 157,173 -2.01% ===============================================================

51. Fairfax 155,204

3 155,204 -3.23% ===============================================================

52. Fairfax 156,604

3 156,604 -2.36% ===============================================================