OPINION AND ORDER OF THREE-JUDGE COURT GRANTING LANGSDON PLAINTIFFS’ MOTION FOR SUMMARY JUDGMENT

This case challenges Tennessee’s most recent legislative reapportionment, effected by Chapter 836 of the Tennessee Public Acts of 1992. It arises under the Equal Protection Clause of the Fourteenth Amendment and Section 2 of the Voting Rights Act, 42 U.S.C. § 1973. A three-judge panel has been convened pursuant to 28 U.S.C. § 2284.

The Langsdon plaintiffs seek a declaration that Chapter 836 is unconstitutional as to the lower house of the Tennessee legislature as well as an injunction against the holding of any further elections pursuant to its provisions. They contend that it violates the “one person, one vote” standards of the Fourteenth Amendment; that it constitutes illegal partisan gerrymandering; and that it violates the Voting Rights Act. This case has been consolidated for trial with Case No. 92-2407, Rural West Tennessee African-American Affairs Council, et al., v. Ned R. McWherter, et al. The primary question presented is whether an apportionment of the lower house which deviates 14% from district to district and breaks 30 county lines is invalid when plaintiffs have demonstrated that an apportionment which deviates less than 10% and breaks fewer county lines is possible.

Chapter 836 was passed pursuant to Article II, Section 5 of the Tennessee Constitution, which governs apportionment of the state House electoral districts. It is based upon the results of the 1990 Federal Decennial Census. Tennessee’s population under this count is 4,877,185 persons; the optimum (or average) population for each of the 99 House districts is thus 49,264. The House districts constructed by Chapter 836 have a maximum population deviation between the largest and smallest districts of 13.90%. (District 5, composed of parts of Unicoi and Washington Counties, has a variance of 6.99% over the optimum of 49,264; District 31, composed of part of Hamilton County, has a variance of 6.91% under the optimum.) Forty-three of the 99 districts have variances greater than five percent above or below the optimum. The plan divides a total of 30 counties one time each in forming multicounty House districts. Plaintiffs have presented an alternative plan, for purposes of comparison, which has a total population variance of only 9.847% and splits only 27 counties in forming multi-county districts.

The legislature has extensive computer resources available to aid its redistricting efforts, including a dedicated GIS computer system with four Sun work stations, a color laser printer, a map plotter, a custom-designed software program, and trained staff to operate this system. Approximately $400,-000 was budgeted by the legislature to purchase the system. In addition, all census data from the 1990 count were available in both hard copy and in computer readable format, as well as the Topologically Integrated Geographic Encoding and Referencing (TIGER) maps prepared by the Census Bureau. No such computerized system was available for reapportionment following the 1980 census.

*449 II.

Federal Rule of Civil Procedure 56(c) directs a court to grant summary judgment to the moving party if that party can show that there is no genuine issue as to any material fact and that the moving party is entitled to a judgment as a matter of law. The facts as set forth above are not in dispute. The Supreme Court has spoken concerning applicable federal Constitutional standards. For puiposes of our case the Tennessee legislature should stay within a 10% deviation from district to district while breaking as few county lines as is reasonably possible. In Reynolds v. Sims, 377 U.S. 533, 577, 84 S.Ct. 1362, 1389, 12 L.Ed.2d 506 (1964), the Court explained the one person, one vote requirement as it applies to reapportionment of state legislative districts: “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.” However, “[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, 84 S.Ct. at 1391. The Court applied these competing principles in White v. Regester, 412 U.S. 755, 93 S.Ct. 2332, 37 L.Ed.2d 314 (1973), a challenge to the Texas state legislative redistricting plan. In a unanimous decision, the Court concluded that the “relatively minor” total population variance of 9.9% between the largest and smallest districts, standing alone, did not violate the one person, one vote standard. Id. at 764, 93 S.Ct. at 2338.

A more difficult situation was presented in Mahan v. Howell, 410 U.S. 315, 93 S.Ct. 979, 35 L.Ed.2d 320 (1973). The Virginia General Assembly passed a redistricting statute in compliance with a new state constitutional provision requiring that the new districts keep political subdivisions intact. The resulting plan produced a population variance of 16.4% between the largest and smallest districts. The Court applied a two-part test: “whether it can reasonably be said that the state policy ... is, indeed, furthered by the plan adopted by the legislature, and whether, if so justified, the divergences are also within tolerable limits.” Id. at 326, 93 S.Ct. at 986. The Court concluded that the first prong was met because political subdivisions remained intact under the plan, with the exception of Fairfax County, which was split into two five-member districts. As to the second prong, the Court held that although the variance of 16.4% “may well approach tolerable limits, we do not believe it exceeds them.” Id. at 329, 93 S.Ct. at 987.

These three cases, read together, indicate a general three-part classification scheme for redistricting plans. Those plans producing a total population variance of 10% or less are de minimus violations of the one person, one vote standard, and a plaintiff cannot prevail on this evidence alone. Those plans producing a variance of more than 16.4% may be insupportable on any grounds. Those plans with a variance falling between those two boundaries establish a prima facie case of vote dilution, but may be justified by a plan that reasonably furthers a rational state policy. The challenged plan before us, with a variance of 13.9%, falls into this category. The problem with the reapportionment in this case is that its variance above 10% does not reasonably further a rational state policy.

In State ex rel. Lockert v. Crowell, 656 S.W.2d 836 (Tenn.1983) [Lockert II], the Tennessee Supreme Court applied the three-part test in a challenge to the state reapportionment following the 1980 census. The Tennessee Constitution provides that in multi-county districts, “each county shall adjoin at least one other county of such district; and no county shall be divided in forming such a district.” Tenn. Const, art. II § 5. The House redistricting plan under attack crossed 57 county lines and had a total deviation of 11.07%. The state trial court concluded that it would be possible, under the evidence before the court, to adopt a plan with a 10% or less maximum variance that- divided only 25 counties. In its decision to modify these restrictions, the Tennessee Supreme Court expressed some misgivings with this conclusion: “Our interpretation of the proof in this record is that it may be very difficult *450 to keep the total deviation in either body below 10% and remain close to the limits of State violations set by the Chancellor.... ” Id. at 844. The court then relaxed the prescribed standards for reapportionment to a 14% maximum deviation and up to 30 split counties, opining that such a plan would be safe from attack, “although admittedly it will not have the de minimus presumption that a 10% plan would have.” Id. The court admitted, however, that it could not guarantee that such a plan would pass federal Constitutional muster: “the suit-risk factor is a matter for Legislative discretion.” Id.

The legislature complied, producing a plan with 13.95% total deviation that split 30 counties. A challenge in federal court followed. In Murmy v. Crowell, No. 3:84-0566 (M.D.Tenn. March 8, 1985), a three-judge court determined that the new plan constituted a prima facie case for violation of the one person, one vote standard, since it produced a total variance greater than 10%. The court then considered in great detail the rational underpinnings of the state policy of minimizing the breaking of county lines, and concluded that the plan reasonably furthered that rational policy. There was no mention in the opinion from the bench of any alternate plan presented to the court which might have diminished either the number of county boundaries violated or the total deviation from the one person, one vote standard. Computer technology used in the drafting of plans was not so advanced in the early 1980s as today, and no plan coming within the 10% total deviation figure was presented to the Murmy court.

III.

Against this backdrop, we examine the case before us. Defendants rely heavily on the Lockert II and Murray cases. They contend that the Tennessee Supreme Court carefully balanced the State’s rational interest in keeping county lines intact, as mandated by the state constitution, against the federal Constitutional requirement of compliance with one person, one vote, and came up with a limit of 14% total variance and 30 split counties. When the Murray court approved the plan, according to defendants, they necessarily affirmed the federal Constitutional validity of the numerical limits set by the Lockert II court, creating a “safe harbor” that should protect the plan before us.

This argument misses the point of the three-part test. It fails under a careful reading of the Lockert II opinion itself. The Lockert II court expressly reaffirmed its holding from Lockert I, 631 S.W.2d 702 (Tenn.1982), that “the State’s constitutional prohibition against crossing county lines must be enforced insofar as.is possible and ... any apportionment plan adopted must cross as few county lines as is necessary to comply with federal constitutional requirements.” 656 F.2d at 838. Defendants unquestionably have failed under this standard. First, even if we were to accept without question a total population variance of 14%, raising the de minimus threshold from the 10% level established by the U.S. Supreme Court, plaintiffs have presented a plan that would not exceed this level and would split fewer counties than Chapter 836. Defendants have failed to obey the mandate of the state’s highest court to enforce Tennessee’s constitutional prohibition on splitting county lines “insofar as is possible.”

Second, Lockert II stressed the paramount nature of federal Constitutional requirements throughout: “This Court is not persuaded ... by defendants’ arguments that we should sanction a single county line violation not shown to be necessary to avoid a breach of federal constitutional requirements,” 656 S.W.2d at 839; “the State Constitution must yield [to avoid a total variance of 21.9%]”, id.; “only one fractional segment of each county [Knox and Davidson] need be detached and joined with other counties to reduce the over population deviation to limits acceptable under federal law”, id. at 841 (emphasis supplied).

Finally, nowhere in the Lockert II opinion does the court purport to establish an absolute numerical standard, applicable in all redistricting contexts. On the contrary, the opinion sets forth in great detail the factual findings of the chancellor below concerning the population deviations for particular districts and the counties from which they were formed, under both the challenged state plan *451 and alternative plans, 656 S.W.2d at 842-43. Each of these findings necessarily was based on population figures from the 1980 census, figures that are no longer either accurate or relevant. The guidelines imposed by the Lockett II court when it directed the legislature to try again necessarily were limited to the particular circumstances of the case. The very paragraph in which the court approved a 14% total variance begins with the limiting words, “Our interpretation of the proof in this record is that it may be very difficult to keep the total deviation in either body below 10% and remain close to the limits of State violations set by the Chancellor. ...” 656 S.W.2d at 844 (emphasis supplied). It is true, as defendants point out, that the Lockert II court loosened the standards imposed by the court below of 10% deviation and 25 split counties. But as the passage just quoted indicates, there was some question as to whether such a plan would be possible on the evidence in the record.

Neither may defendants rely on the Murray decision. That opinion concentrated almost completely upon the rational basis for the state’s policy of not breaking county lines; there was no discussion concerning why the challenged plan reasonably advanced that goal. The Murray court’s approval of the legislative redistricting plan at issue in that case was limited to the record before it, and does not control the outcome of this case. No alternate plan was presented staying within the 10% variance figure and breaking fewer than the 30 counties that Lockert II suggests should be the applicable state policy-

There is further justification for revisiting the 10%-deviation, 30-eounty issue. Redistricting methodologies have changed dramatically in the last decade. The legislature’s new $400,000 computer system replaced a system of pencil, paper and pocket calculator, allowing a drop in the time required to produce plans that may satisfy the one person, one vote requirement. As the Supreme Court noted in Karcher v. Daggett, 462 U.S. 725, 733, 103 S.Ct. 2653, 2659-60, 77 L.Ed.2d 133 (1983), “rapid advances in computer technology and education during the last two decades make it relatively simple to draw contiguous districts of equal population and at the same time to further whatever secondary goals the State has.” This is even more true today. Witnesses in this consolidated case testified to preparing state-wide plans in a matter of hours.

Defendants contend that our ruling imposes upon them a new standard, akin to a “least restrictive means” analysis. As long as someone, somewhere can come up with a “better” plan than the legislature enacts, i.e., one that splits one fewer county or has a population variance one tick lower, then the state’s plan may be stricken as unconstitutional. This is not what we have in mind.

We are not disturbed by the prospect of the State being subjected to pressure to obey both the U.S. Supreme Court directive to “make an honest and good faith effort to construct districts ... as nearly of equal population as is practicable,” Mahan, 410 U.S. at 324-25, 93 S.Ct. at 985 (quoting Reynolds, 377 U.S. at 577, 84 S.Ct. at 1390), and its own highest court’s mandate to “cross as few county lines as is necessary to comply with federal constitutional requirements,” Lockert II, 656 S.W.2d at 838. To the extent that this is a hardship, it is one imposed by the two documents defendants are sworn to uphold, the state and federal Constitutions.

We do not interpret the U.S. Supreme Court’s decisions to require that the legislature devote every last iota of its scarce time and resources toward drawing the best plan possible. The Supreme Court’s three-tiered classification scheme for evaluating state apportionment plans provides some leeway. As long as the total population variance is less than 10%, and there is no showing of bad faith on the part of the legislature, the plan will be presumed valid under the one person, one vote standard. Had Chapter 836 produced a variance of 9.9% instead of 13.9%, and assuming no allegation of bad faith, we would be far less interested in alternative plans with lower variances. The Murray court required only that the apportionment plan reasonably further the state’s rational interest, and even this level of scrutiny may be avoided as long as the 10% threshold is not crossed. It is patently unreasonable to *452 justify a 14% variance on the basis of not splitting county lines when fewer counties may be split while simultaneously decreasing that variance below the 10% goal.

We echo the Murray court’s finding that Tennessee’s constitutional provision against splitting counties is a rational state policy to be considered in its apportionment legislation. As the Lockert II court pointed out, however, this policy must yield to the one person, one vote doctrine, as well as to the requirements of the Voting Rights Act. 656 S.W.2d at 844. We turn now to fashioning an appropriate remedy.

IV.

For the reasons stated, we hold that Chapter 836 of the 1992 Tennessee Public Acts violates the one person, one vote doctrine under the Equal Protection Clause, and is therefore unconstitutional. We enjoin defendants from conducting any further elections pursuant to its district boundaries. Defendants must prepare and submit a new apportionment plan that satisfies the one person, one vote standard. Any upward deviation from a total population variance of 10% must be justified by a rational state interest, such as to comply with the state’s prohibition on splitting counties or to prevent the dilution of minority voting strength. Plaintiffs have demonstrated that it is possible to construct a plan that has a population variance of less than 10%, splits only 27 counties. A proposed plan from the legislature that exceeds a population variance of 10% will be invalid if it does not reasonably further rational state interests such as those recognized in Lockert II at least as well as plaintiffs’ alternative plan.

The legislature should prepare and submit a new plan by December 24, 1993. This will allow the parties and court approximately two months to review the plan prior to March 19, 1994, the minimum time needed prior to the May 19, 1994 filing deadline for the public and potential candidates to receive notice of the district lines which will form the new apportionment plan. We are confident defendants will be able to meet the deadline we have imposed. The Supreme Court has “repeatedly held that redistricting and reapportioning legislative bodies is a legislative task, which the federal court should make every effort not to preempt.” Wise v. Lipscomb, 437 U.S. 535, 539, 98 S.Ct. 2493, 2497, 57 L.Ed.2d 411 (1978). We recognize that “judicial relief becomes appropriate only when a legislature fails to reapportion according to federal constitutional requisites in a timely fashion after having had an adequate opportunity to do so.” Reynolds, 377 U.S. at 586, 84 S.Ct. at 1394. In the unlikely event defendants fail to enact a constitutional apportionment plan, we will be faced with the unhappy task of devising one, perhaps by “using the best parts of ... proposed plans.” Id.

Accordingly, IT IS SO ORDERED.

ORDER OF THREE-JUDGE COURT

The defendants have moved to alter or amend the Court’s prior decision in this case of September 15, 1993, pursuant to Rule 59 of the Fed.R.Civ.P., on three grounds. The Court denies the first two grounds but grants the third ground as follows:

1. The Court denies the first and second grounds of the defendant’s motion to alter or amend its prior decision of September 15, 1993. Neither of these two grounds was raised in the prior proceedings. The questions now presented were therefore waived. Moreover, it is clear from the decision of the Court on September 15, 1993, that the Court has placed no significant limitation on the authority or duty of the General Assembly to comply with the Voting Rights Act. The Court’s actions should not be interpreted as a modification of the General Assembly’s duty to comply with the Act or an instruction as to its duties under the Act. No such questions under the Act, or with respect to a claim of a census undercount, were presented to the Court at trial or decided by the Court.

2. The Court grants the third ground of the defendant’s motion to alter or amend its prior decision of September 15, 1993, and hereby changes the deadline by which the defendants must enact a plan from December 24, 1993, to January 25, 1994.

Accordingly, it is so ORDERED.