concurring in part and dissenting in part.

I.

I join the court’s opinion that the law does not preclude the State’s Legislature from enacting a mid-decade redistricting plan following a court-ordered remedial plan such as the one enacted by the court in Balderas v. State of Texas. I understand the court’s opinion to be limited to that question. I write separately to em*516phasize that it is one question to ask whether the law prohibits a state from enacting a mid-decade redistricting plan. It is quite another to ask whether a state may dictate electoral outcomes by using its Article I authority to thwart the Supreme Court’s mandate that votes cast in a Congressional election be given as nearly equal weight as possible.

I do not read the Court’s decisions in U.S. Term Limits v. Thornton, 514 U.S. 779, 115 S.Ct. 1842, 131 L.Ed.2d 881 (1995) and Cook v. Gralike, 531 U.S. 510, 121 S.Ct. 1029, 149 L.Ed.2d 44 (2001) quite as narrowly as my colleagues. Admittedly, those cases rejected state laws which sought to add to the Qualifications Clause either directly as in Term Limits, or indirectly, through pejorative ballot descriptions as in Cook. From that standpoint they are easier cases than this one to decide. Nonetheless, these decisions underscored that “the Framers understood the Elections Clause as a grant of authority to issue procedural regulations, and not as a source of power to dictate electoral outcomes, to favor or disfavor a class of candidates, or to evade important constitutional restraints.” Term Limits, 514 U.S. at 833-34, 115 S.Ct. 1842; Cook, 531 U.S. at 523, 121 S.Ct. 1029.

Although the state’s authority to issue procedural regulations is broad, at some point a state exceeds the power granted to it by the Elections Clause. Modern technology effectively allows a state to dictate electoral outcomes and to favor or disfavor a class of candidates by enabling extreme partisan gerrymandering. I join the court’s decision that the Elections Clause does not prohibit mid-decade redistricting and note that there may be legitimate state interests advanced by the effort. Whether the present exercise is a “legitimate” state interest may ultimately be resolved by the Supreme Court in Vieth v. Jubelirer, and it seems to me that the Elections Clause would be equally offended by a state’s abuse of its authority regardless of whether such abuse occurred in the beginning, the middle, or the end, of a decade following the release of the census data. It is not the timing of the endeavor that creates problems, it is the fact of it.

As noted by the court in Balderas, partisan gerrymandering reflects a “fundamental distrust of voters.” As in other contexts, extreme partisan gerrymandering leads to a system in which the representatives choose their constituents, rather than vice-versa. I join the court’s judgment on the partisan gerrymandering issue given the high bar set in the equal protection context, but I do not join the court’s discussion of the issue under Section III.B. That discussion implies that extreme partisan gerrymandering is okay as long as one party can reverse the tide when it takes over the statewide offices. Dr. Alford’s report and testimony provides a detailed analysis of the effects of the extreme partisan gerrymander enacted in this case and the dangers of ignoring these issues. As Dr. Alford explains, these dangers include the encouragement of race-based redistricting. At present, we lack the legal precedent to sustain this challenge under either the Equal Protection Clause or the Elections Clause. The present case fails under the standards announced in Davis v. Bandemer and its progeny-even though those standards may or may not ultimately control the disposition of this case, given its timing and that of the arguments in Vieth v. Jubelirer.

I join much of the balance of the opinion. I agree that the plaintiffs have not prevailed on their equal protection claims and I agree that no violation of the Voting Rights Act has been shown with respect to Districts 18 and 30. On the remaining § 2 *517issues, I dissent from the court’s opinion related to the issues surrounding District 23 and I concur in the court’s judgment insofar as it rejects the claims surrounding District 24 and the claims surrounding Districts 1, 2, 4, 9, 10, 11 & 17. I will first examine the claims related to District 23 and then visit the issue of the influence districts.

II.

A.

The state action in this case unlawfully dilutes the strength of the Latino voters residing in former District 23. To that end, the majority errs when it holds that the State may permissibly “trade off’ the rights of minority voters in former District 23 for those in new District 25, a district created to assist the state with its pre-clearance efforts. I would enjoin the operation of Plan 1374C, permit the Legislature to remedy the violations, and, given the amount of the state that would likely be affected by the changes, order that the elections be conducted under Plan 1151C.

B.

For a statute whose purpose “is to prevent discrimination in the exercise of the electoral franchise and to foster our transformation to a society that is no longer fixated on race,” Georgia v. Ashcroft, — U.S. —, —, 123 S.Ct. 2498, 2517, 156 L.Ed.2d 428 (2003), the Voting Rights Act receives a curious interpretation in this case. The State’s heavy reliance on the flexibility given to it by Ashcroft might carry the day with me if Plan 1374C bore any resemblance to the redistricting plan at issue there. It does not.

To begin, the state senate plan at issue in Ashcroft was overwhelmingly supported by the minority legislators. Just the opposite is true in this case. Georgia’s plan sought to maximize the statewide influence of black voters in the State of Georgia. The plan unpacked districts in which blacks constituted a super-majority and (1) maintained the number of majority-minority districts and, at the same time, (2) increased the number of “influence districts” in which blacks would be expected to exert a significant-if not decisive-force in the electoral process. Georgia did so by placing black voters into districts that were likely to lean Democratic in the election returns. Because black voters in Georgia usually supported Democratic candidates, the creation of these influence districts made it likely that blacks could exert more influence in the election process and, correspondingly, increased the likelihood that the elected candidate would be responsive to the black communities’ needs. The distinct but related concepts of substantive versus descriptive representation were at issue.

Georgia argued that, although the plan made the super-majority districts somewhat less “safe,” its plan as a whole increased the political clout of blacks statewide and therefore did not result in a retrogression in African-American voters’ effective exercise of the electoral franchise. Ashcroft held that “[i]n order to maximize the electoral success of a minority group, a State may choose to create a certain number of ‘safe’ districts, in which it is highly likely that minority voters will be able to elect the candidate of their choice. Alternatively, a State may choose to create a greater number of districts in which it is likely-although perhaps not quite as likely as under the benchmark plan-that minority voters will be able to elect candidates of their choice.” Ashcroft, — U.S. at —, 123 S.Ct. at 2511. The Court explained that the extent to which a plan created coalition and influence districts were important considerations in the preclearance inquiry. Ultimately, the Court held that *518the three-judge court, by circumscribing its focus on the unpacked districts, had engaged in too narrow a review of whether the redistricting plan would lead to retrogression at the statewide level. The majority reads Ashcroft to give the state flexibility to comply with its obligations under the Voting Rights Act. This is true, to a point. Ashcroft was a § 5 case and its language certainly permits the states a certain amount of latitude in determining how to maximize minority voter influence. But I do not read Ashcroft-a decision designed to foster minority participation in the political process-to permit the state to dismantle an existing opportunity district for political purposes so long as the loss is made up somewhere else. Nor do I read Ashcroft to make it harder to prove claims under § 2. There was no § 2 question decided in Ashcroft. Indeed, one of the grounds Georgia asserted in support of pre-clearance was the plan’s compliance with § 2 of the Voting Rights Act. It is hard for me to believe that the Supreme Court intended to make it harder to prove a § 2 claim by endorsing a statewide approach that the Court believed could have the effect of maximizing minority influence.

Ashcroft recognizes the value of safe districts, coalition districts, and influence districts as means to increase minority voting opportunities and political influence. Make no mistake about it: Despite the State’s protestations to the contrary, the changes to District 23 under Plan 1374C were not intended to increase minority voter participation either by strengthening the district or “unpacking” the minority voters into adjoining districts to maximize the overall political strength of Hispanics. These changes were designed to crush these minority voters’ participation in the political process.

C.

Under Plan 1000C, in effect after the 1990 census, as well as under its successor, Plan 1151C, District 23 was a protected Latino opportunity district. The district had a Latino citizen voting age population in excess of 50% and was designed to offer Latino voters the opportunity to elect candidates of choice. Under Plan 1000C, the district boasted a Spanish Surname Voter Registration of 53.3%. And under Plan 1151C, that figure was increased to 55.3%. The Balderas court implicitly recognized that District 23 under Plan 1151C was a protected minority opportunity district when it maintained only six Latino majority citizen voting age districts. A review of the statistical package for Plan 1000C reveals that District 23 was one of those six districts. All six (and only those six) had a Spanish Surname Voter Registration in excess of 50%.

As it was configured under Plan 1151C, District 23 offered Latino voters the opportunity to elect their candidate of choice. In the present case, Dr. Gaddie, the State’s expert, testified that District 23 under Plan 1151C performed for Latino voters in 2002 by electing the Latino candidate of choice in 13 out of 15 statewide elections. He also testified that a district which elects the Latino-preferred candidate of choice in such numbers offers Latinos the opportunity to elect their candidate of choice. Dr. Engstrom, the GI Forum plaintiffs’ expert, reached a similar conclusion: He analyzed racially contested statewide elections from 1994 to 2002 and concluded that District 23 under Plan 1151C elected the preferred Latino candidate in 5 out of 8 races and offers Latino voters the opportunity to elect candidates of choice. Dr. Lichtman, called by the Jackson plaintiffs, arrived at the same conclusion.

*519When it enacted Plan 1374C, the State altered the racial composition of District 23 not to increase the likelihood that the Latino community therein would elect a candidate of its choice, but to ensure it would have no practical influence on the congressional election. All of the experts agree that Plan 1374C alters District 23 to the point where it has no hope of functioning as an effective Latino opportunity district. There is no dispute that the State altered District 23 to help re-elect Congressman Henry Bonilla because it predicted that if Latinos continued to constitute a majority of the citizen voting age population in District 23, Congressman Bonilla would ultimately lose. The evidence is that he is not the Latino candidate of choice. The 2002 congressional election foreshadowed the need for this change: Congressman Bonilla received only 8% of the Hispanic vote. Spanish surname voter registration has correspondingly risen during the same time period and, until the Legislature passed Plan 1374C, had grown to approximately 55% for the district.

The State’s solution to this political problem was brutal, yet simple: destroy the opportunity district. The state did so by cracking a cohesive Hispanic community out of Webb County and taking in Anglos from the Texas Hill Country to build a district in which the Hispanic community will not be able to influence the outcome of the election. The majority’s characterization of new District 23 as a Latino “influence district” is therefore in error. An influence district is a district in which the minority population carries enough political weight potentially to be the swing vote in the election and command the attention of the representative. A competitive district is the touchstone. Not so with new District 23. The Hispanic citizenship figures have been reduced by so great a margin that the Hispanics in District 23, although having the ability to control the Democratic primary, will lose the general election every time. The undisputed regression analyses confirm this. Contrary to the majority’s characterization, the district’s very design ensures a lack of competitiveness and a corresponding lack of responsiveness.

There are, however, a total of 359,000 Latinos who continue to reside in new District 23. They object to the State’s dismantling of their opportunity district under § 2. The question presented is whether a state can, consistent with § 2, intentionally dilute a minority group’s voting rights in an existing opportunity district to obtain a partisan advantage while, at the same time, offset the effects by creating a new district in another part of the state. The majority’s answer is that Ashcroft and Johnson v. De Grandy, 512 U.S. 997, 114 S.Ct. 2647, 129 L.Ed.2d 775 (1994), permit this sort of line-drawing. I disagree.

D.

Contrary to the majority, I do not read Ashcroft or De Grandy to encourage the sort of voting rights swap meet put into play by Plan 1374C. The majority effectively reads Ashcroft’s § 5 holding and De Grandy to countenance violations of § 2 in one part of the state, so long as those violations are “offset” by the creation of new minority opportunity districts elsewhere in the state. The State conceded that this was its position in response to questions from the bench during closing arguments. By its terms, of course, Ashcroft says nothing about a state plan which unlawfully dilutes the minority strength in one part of the state and seeks to “offset” that dilution by the creation of a new minority opportunity district in a different part of the state-one of Georgia’s conten*520tions was the plan’s compliance with § 2 required pre-clearance under § 5.

That Ashcroft did not endorse the approach suggested by the State is not surprising. The Court had already answered this question once in De Grandy and again in its Shaw cases. In De Grandy, the state proposed a per se rule that proportionality was a safe harbor, insulating any reapportionment plan which provided proportional representation from a Section 2 challenge. In rejecting that proposal, the Court stated that the state’s argument rested on an “unexplored premise of highly suspect validity: that in a given case, the rights of some minority voters under § 2 may be traded off against the rights of other members of the same minority class.” 512 U.S. at 1019, 114 S.Ct. 2647. Later, in Shaw v. Hunt, 517 U.S. 899, 917, 116 S.Ct. 1894, 135 L.Ed.2d 207 (1996)(Shaw II), the Court made its views even more explicit:

If a § 2 violation is proved for a particular area, it flows from the fact that individuals in this area “have less opportunity than other members of the electorate to participate in the political process and to elect representatives of their choice.” 42 U.S.C. § 1973(b). The vote-dilution injuries suffered by these persons are not remedied by creating a safe majority-black district somewhere else in the State. For example, if a geographically compact, cohesive minority population lives in south-central to southeastern North Carolina, as the Justice Department’s objection letter suggested, District 12 that spans the Piedmont Crescent would not address that § 2 violation. The black voters of the south-central to southeastern region would still be suffering precisely the same injury that they suffered before District 12 was drawn. District 12 would not address the professed interest of relieving the vote dilution, much less be narrowly tailored to accomplish the goal.

Arguing, as appellees do and the District Court did, that the State may draw the district anywhere derives from a misconception of the vote-dilution claim. To accept that the district may be placed anywhere implies that the claim, and hence the coordinate right to an undiluted vote (to cast a ballot equal among voters), belongs to the minority as a group and not to its individual members. It does not.

Although I recognize that “States retain broad discretion in drawing districts to comply with the mandate of § 2,” Shaw II, 517 U.S. at 917 n. 9, 116 S.Ct. 1894 (citing Voinovich v. Quilter, 507 U.S. 146, 156-57, 113 S.Ct. 1149, 122 L.Ed.2d 500 (1993); and Growe v. Emison, 507 U.S. 25, 32-37, 113 S.Ct. 1075, 122 L.Ed.2d 388 (1993)), I do not read the Court’s cases to mean that the “effects” test of § 2, if satisfied, may be defended against by pointing to a political agenda in the affected portion of the jurisdiction and compensation, over the long haul, to other members of the injured group residing elsewhere in the jurisdiction. De Grandy rejected the state’s proposed safe harbor rule precisely because it encouraged what happened in this case. The court said: “... we reject the safe harbor rule because of a tendency the State would itself certainly condemn, a tendency to promote and perpetuate efforts to devise majority minority districts even in circumstances where they may not be necessary to achieve equal political and electoral opportunity.” De Grandy, 512 U.S. at 1019-20, 114 S.Ct. 2647 (emphasis added). New District 25 falls squarely within the scope of this statement.

In this case, the State’s redistricting effort mirrors precisely what De Grandy cautioned against. The fundamental flaw *521in the majority’s approach is that it fails to appreciate that the Latino districts created in Plan 1374C are not in “the same area” as the Gingles districts have been located historically. In the past, the Latino groups have enjoyed participation in the six Gingles districts located in Plan 1000C and in Plan 1151C. Although there are still six Latino opportunity districts in Plan 1374C, those new districts take in quite different geography and population. The State took in, geographically, 10 additional counties and portions of 4 others which span an area of at least 8,500 square miles. The State also took in more than 651,000 additional persons. To accomplish its goals, the state was forced to create new District 25 which reaches from the border in excess of three hundred miles north to include Hispanic population in Travis County.

In its effort to validate Plan 1374C,' the majority first tells us that District 25 was created “[t]o avoid retrogression under Section 5.” Quite correctly, the majority recognizes that the changes to District . 23 resulted in the dilution of the Latino vote therein. But then, when confronted with the language of Shaw and De Grandy which prohibits exactly this sort of tradeoff, the . majority explains that the state enjoys the flexibility under De Grandy to draw Gingles districts in a different way, so long as it creates a total of six (but only six) districts in which Latinos may elect a candidate of choice. But Gingles addresses liability under § 2. No one suggests the State needed to create new District 25 to guard against a potential vote dilution challenge by a resourceful if not aspiring coalition of Latinos living in Travis County and those living in the colonias over 300 miles away near the Texas-Mexico border. At yet • another place in its opinion, the majority implicitly repudiates this very species of § 2 liability when it holds to be non-compact the GI Forum’s demonstration districts — even though all have better compactness scores than new District 15 in Plan 1374C.

The majority blurs the distinction between the State’s flexibility to comply with its procedural obligations under § 5 and its flexibility to comply with its remedial obligations under § 2. The majority misses this distinction because it fails to appreciate the fact that the Latino opportunity districts in Plan 1374C take in different populations and geography than before.1 The majority’s initial reaction to new District 25 was of course correct: the State created new District 25 to “offset” the loss of old District 23 for § 5 preclearance purposes. But the underlying forces driving the need for new District 25 are inconsistent with Ashcroft’s rationale. At issue in Ashcroft was Georgia’s decision to unpack black opportunity districts to increase the statewide political might of the minority voters. Central to Ashcroft’s willingness to provide states with flexibility was the Court’s recognition that the plan was intended (or at least contended to be intended) to increase minority voting strength statewide. Unlike the efforts in Ashcroft, the dilution of Latino voting strength in District 23 was not designed to maximize Latino voting strength statewide, and, as noted, the majority’s characterization of District 23 as continuing on as some sort of an Hispanic influence district is error. Under the majority’s theory of influence districts, we are left to conclude that Congressman Bonilla will be more *522responsive to the Latino community now that they constitute over 10% less of his constituency.

This turns Ashcroft on its head. District 23 is designed so that the Hispanic population therein will not influence the outcome of the elections and mil not tend to make the elected representative more responsive to the community’s needs. The only thing the Hispanic community in District 23 will influence is the total population of the district to meet “one person, one vote” requirements. The state action was designed to increase Henry Bonilla’s chance of electoral success at the expense of District 23’s dissatisfied Hispanic voters. To suggest that those minority voters cannot prevail on a dilution claim under § 2 elevates the State’s political agenda over the individual rights of Latino voters in former District 23. For these reasons, I would reject the State’s efforts to trade the rights of Latinos in old District 23 for those in new District 25.

E.

As stated, it is my opinion that the State violated § 2 when it dismantled old District 23 and replaced it with new District 25. If the majority is correct that the State could draw new District 25 to comply with its Gingles obligations, then the majority is confronted at once with the GI Forum plaintiffs’ dilution claim. These plaintiffs claim that Plan 1374C dilutes the statewide strength of Hispanic voters because it fails to create a seventh Latino majority district. The majority, however, rejects the GI Forum plaintiffs’ claims for several reasons. First, the majority opines that the proposed districts will not be effective. Second, the majority holds that the districts are not compact. Third, the majority determines that De Grandy proportionality tends to show that Plan 1374C will not have the effect of unlawfully diluting the Latino vote. Finally, the majority suggests that Balderas rejected the identical claim. None of these reasons is persuasive.

1.

In this case, if the State is permitted or required to draw Gingles remedial districts into the new areas of the state and in the manner in which it did, the GI Forum plaintiffs assert that the Gingles preconditions suggest that seven majority Latino districts should be created. I agree. The evidence is undisputed that, under Plan 1385C, sponsored by the GI Forum plaintiffs, Latino voters constitute the majority of citizen voting age population in seven congressional districts in South and Central Texas. See Valdespino v. Alamo Heights Independent School Dist., 168 F.3d 848 (5th Cir. 1999)(interpreting Gingles to require consideration of citizen voting age population or “CVAP”); Campos v. City of Houston, 113 F.3d 544, 548 (5th Cir. 1997)(same). These districts are District 15 (63.3% CVAP), District 16 (68.0% CVAP), District 20 (58.0% CVAP), District 23 (57.2% CVAP), District 25 (58.4% CVAP), District 27 (59.9% CVAP), and District 28 (50.3% CVAP).

The majority resists the GI Forum plaintiffs’ demonstration plan because it believes the plan does not provide seven “effective” Latino districts. I disagree. The majority places great emphasis on the fact that the Hispanic CVAP of District 28 in Plan 1385C is only 50.3%. Be that as it may, under the evidence submitted by the GI Forum plaintiffs, the district elected the Latino candidate in 8 out of 8 elections since 1994, and the vote for the Latino candidate in the most recent elections, 2002, exceeded 55% of the total vote in each of the four elections. See, e.g. GI Forum plaintiffs’ Exh. 131; (Tr. 12/16/03 a.m. at 29-30). On balance, District 28 is almost identical in performance to District *52328 in the State’s plan 1374C despite the fact that, in District 1374C, District 28 has an Hispanic CVAP of 56.2%. By contrast, the weakest district in the GI Forum plaintiffs’ demonstration map is District 23, which elected the Latino candidate of choice in 5 out of 8 elections at margins nearly identical to those .which led the Balderas court to conclude that District 23 was a protected minority opportunity district. Each of the remaining Latino districts in Plan 1385C elects the Latino candidates to office in 8 out of 8 elections. Viewed in the light of these election returns as opposed to simply looking at the CVAP content or any other single statistic, all of these districts perform, in terms of winning elections, at least as. well as their counterparts in Plan 1374C.

The majority also relies on the testimony of experts called by other parties, including Dr. Jerry Polinard, an expert witness called by the Valdez-Cox plaintiffs. The majority reads Dr. Polinard’s testimony to suggest that in all cases a higher Hispanic CVAP number is necessary to find an effective Hispanic opportunity district. The majority is wrong. In the first place, as set forth above, Dr. Engstrom’s analysis of the election results tell a different story. Moreover, Dr. Polinard was in fact testifying about the effects of the districts as drawn in Plan 1374C, not Plan 1385C. What Dr. Polinard said with respect to numbers is:

I have no magic number. That’s going to vary-vary by District. I will state the obvious, that if the Spanish surname voter registration percentage goes up, the opportunity goes up.

I think you become comfortable with opportunity districts once you break into those 60 percent ranges.

(Tr. 12/16/03 p.m. at 50-51).

The majority also cites the testimony of the State’s expert with respect to new District 15 to reject as effective District 28 in the demonstration plan. In terms of actual election results, however, new District 15 appears to perform worse than the GI Forum plaintiffs’ proposed District 28, despite the fact that new District 15 has a higher Hispanic CVAP than proposed District 28.

The truth of the matter is that the effectiveness of a minority opportunity district will vary by the district. A expert might find, for example, a need for a greater Hispanic CVAP in. a district ranging from Travis County to the border and covering three or four media markets than would be needed in a district that is somewhat more compact. Under the evidence, all of the districts in Plan 1385C have the numbers to make them effective Latino opportunity districts.

This might be a closer call if the demographic evidence were otherwise. A showing that Latino population growth and voter registration was on the decline might counsel a court to conclude that gains over the short run might be offset by anticipated losses in the end. That is not our record. The evidence establishes that the Latino voter registration as well as overall population growth is rising. Over the course of the decade, the only conclusion to be drawn from this evidence is that Latino voting strength in these districts will become stronger, not weaker. The GI Forum plaintiffs have established that their proposed districts would be effective.

2.

The majority also rejects the GI Forum plaintiffs’ contention that their districts are “compact.” But under the objective compactness scores, several of the GI Forum plaintiffs’ demonstration districts are, on balance, more compact than those in Plan 1374C. Although a few are worse, none of the districts scores worse than District 15 *524in Plan 1374C. Under the demonstration plan, 5 of the proposed districts have better “perimeter to area” scores than their counterparts in Plan 1374C. None of the districts in Plan 1385C have a perimeter to area score as high as District 15 in Plan 1374C. Under the smallest circle measure, three of the demonstration districts score better, one scores the same, and none is, again, as high as District 15 in Plan 1374C.

The majority disregards these scores, saying the comparison “understates the unusual shapes of the proposed demonstration districts.” But it is the majority who overstates the significance of the “unusual” shapes of the districts. Focusing in on the “look” of the map as opposed to the objective scores and population centers, the majority says that District 25 virtually “bisects” District 23 in the demonstration plan. Viewed in the context of a redistricting plan, however, the two districts make perfect sense. District 23 strives to include the border communities of interest along the Rio Grande and, at the same time, retains over 99% of the City of Laredo-a major border population center-in District 23. The portions of District 25 which allegedly “bisect” District 23 are actually the boundaries of Dimmitt County, the total population of which is 10,248. The population of Laredo, 99% retained in District 23, is 176,576. The GI Forum plaintiffs’ demonstration plan is, on balance, more compact than the pertinent portions of Plan 1374C.

3.

The majority’s principal answer to the GI Forum plaintiffs is that Plan 1374C provides Latinos the opportunity to elect congressional candidates in substantial proportion to their share of the relevant population as a whole under Johnson v. De Grandy. I disagree.

a.

In De Grandy, the Court noted:

If the three Gingles factors may not be isolated as sufficient, standing alone, to prove dilution in every multimember dis-tricting challenge, a fortiori they must not be when the challenge goes to a series of single-member districts, where dilution may be more difficult to grasp. Plaintiffs challenging single-member districts may claim, not total submergence, but partial submergence; not the chance for some electoral success in place of none, but the chance for more success in place of some. When the question thus comes down to the reasonableness of drawing a series of district lines in one combination of places rather than another, judgments about inequality may become closer calls.

512 U.S. at 1012-13, 114 S.Ct. 2647.

Under De Grandy, proportionality is one factor to be considered in assessing the totality of circumstances to determine whether unlawful vote dilution has occurred under § 2. Id. at 1022, 114 S.Ct. 2647. “Proportionality” in this sense involves a comparison between (1) the percentage share of legislative districts in which the population of the protected class has a majority and (2) the protected class’s percentage of the relevant population. Proportionality is one factor to be considered, and it does not create a safe harbor precluding § 2 liability when present, nor does it impose per se § 2 liability when absent.

b.

Proportionality does not compel the rejection of the GI Forum plaintiffs’ claims in this case. Because of the procedural posture in De Grandy, the Court reserved the question whether proportionality should be measured on a statewide basis or by focusing on a smaller area of the jurisdiction at issue. De Grandy, 512 U.S. *525at 1022, 114 S.Ct. 2647 (“[w]e have no occasion to decide which frame of reference should have been used if the parties had not apparently agreed in the District Court on the appropriate geographical scope for analyzing the alleged § 2 violation and devising its remedy.”). Other courts have struggled with the question, but many have assessed proportionality by focusing on the nature of the specific Voting Rights Act claims at issue.2

Under the facts of the present case, De Grandy proportionality should be measured by comparing the number of effective Latino congressional districts to the Latino percentage of the relevant statewide population. It is true that the GI Forum plaintiffs propose their Latino districts in the South and West Texas areas; however, it is equally the case that the GI Forum plaintiffs attack the operation of Plan 1374C because of its impact and its overall effect on the Latino voters in the state as a whole. They did not limit their challenge to the effect of Plan 1374C to the areas in South and Central Texas. An integral part of the claim is Plan 1374C’s failure to maintain districts in other parts of the state in which minority communities might play an influential if not outcome determinative role in the general elections to increase the likelihood that the elected candidate will respond favorably to the minority communities’ needs.

The State apparently agreed with this approach until the evidence was in and it became apparent there was a problem under § 2 with 1374C. The State’s pretrial submissions addressed proportionality on a statewide basis. In particular, the State’s trial brief addressed De Grandy proportionality and represented, misleadingly, that proportionality was met under Plan 1374C, because Hispanics “would have a majority of the population in 25% of the districts.” Revised State Defendants’ Trial Brief, filed December 3, 2003, at 39 (emphasis added).3 It was not until after the close of the evidence that the State, or anyone else, contended that proportionality should not be assessed statewide.

There is no risk that assessing proportionality on a statewide basis will lead to an over-representation of Latinos in Congress. For instance, there is no evidence tending to prove that Latinos elsewhere in the state (e.g., the panhandle, Dallas/Fort *526Worth, or East Texas) constitute a reasonably compact effective voting majority such that they could in the future assert a Section 2 violation in those parts of the state and require the creation of yet additional Gingles districts. Such evidence might require the court to assess proportionality on a smaller scale. Rural West, 209 F.3d at 844 (“The State complains that by allowing the plaintiffs to define the frame of reference for their § 2 claim, we will enable future litigants to carve up successively smaller areas of the State until they are able to maximize the number of majority-minority legislative districts-a result not countenanced by the Voting-Rights Act [but] as the district court pointed out, however, the Gingles preconditions operate to prevent just the sort of limit-lessly small ‘reverse gerrymander’ whose specter the State raises here.”). In this case, however, the claim is that Plan 1374C dilutes the statewide voting strength of Latinos. That is the proper geographic scope by which to measure De Grandy proportionality.

c.

So measured, Latinos constitute 32% of the total population and 29% of the voting age population in the State of Texas. De Grandy proportionality thus suggests that Latinos should comprise an effective voting majority, depending on whether one uses voting age or total population figures, in possibly nine or ten such districts. Under either measure, Plan 1374C fails to provide proportional representation to Latinos. Latinos enjoy the ability to elect representatives of their choice in only six districts: Districts 15, 16, 20, 25, 27 and 28. As a result, Plan 1374C fails to provide proportional representation.

The majority concludes that proportionality should be assessed on the basis of the minority group’s percentage of the citizen voting age population. The majority overlooks that this question arises in the context of the apportionment of seats to the United States Congress. The seats are apportioned according to total population. From 1990-2000, the Latino growth in Texas accounted for more than 60% of the state’s total population. The state thus enjoys the benefit of the total Latino population through an increase in its Anglo congressional delegation, but at the same time seeks to restrict the Latino’s opportunity to elect to the fewest number of districts. Latinos are thus counted one way for apportionment purposes, but another when it comes to equal opportunity to participate in the political process. Stated another way, the state seeks the maximum use of the Latino population to gain power but seeks to minimize the sharing of power with the Latinos by using a more restrictive measure.

d.

Even assuming the correctness of the majority’s approach, however, 6 districts out of 7 remains at the low end of rough proportionality. For reasons that are apparent from this record, the majority errs when it rejects the § 2 claims by relying on rough proportionality. I would ascribe far less weight to the proportionality issue given the circumstances under which the present case comes to us. We must remember that the State is having to defend this claim because it made the conscious choice to dismantle a minority opportunity district to thwart the growing Latino dissatisfaction with an incumbent Congressman. Just when it became apparent that District 23 was becoming more effective for the class it was intended to protect, the State intentionally altered it. The State thus placed itself in this position and the related position of having to create a new Latino opportunity district to meet a retrogression concern. Although the majority concludes this action is permitted by De *527Grandy, that case rejected a safe harbor rule precisely because such a rule would encourage states to do what Texas did: create an “offsetting” district where none was necessary and substitute one group’s voting rights for another.

The weight ascribed to proportionality by the majority allows the State to mask its efforts to thwart the policies underlying the Voting Rights Act. To illustrate, suppose that new District 15 in Plan 1374C fails to perform and does not elect a candidate of choice for Latinos. Suppose further that over the course of the next few years, the Latino community mobilizes and puts pressure on the incumbent. To use De Grandy to permit the State at that stage to dismantle District 15 and create a new district somewhere else has a tendency to perpetuate the legacy of discrimination, not to thwart it. Our record, and the State’s action directed toward old District 23, is no different.

4.

Finally, the majority suggests that the Balderas court rejected this identical claim. The Balderas court did not. It is ironic that the State finds comfort in a remedial order that the Legislature rejected as controlling. Although I agree that the State has the authority to enact a redistricting plan, that new plan must stand or fall on its own merit, and not on any language of the Balderas remedial order-an order which must be read in context of Plan 1151C as a whole and which did not purport to address the merits of Plan 1374C. Balderas drew a remedial plan without the benefit of either the controlling citizen voting age population data or the most recent election returns. Those data are now available. Under those figures, seven districts may be drawn in which Latino voters will have the opportunity to elect their candidates of choice to Congress.

Balderas was primarily motivated by the desire to meet the Supreme Court’s holding that a federal court should draw a remedial plan bearing in mind the requirements of the Voting Rights Act. The remedial order was thus forced to straddle the requirements of § 2 and § 5. At that time, prior to Ashcroft, a compelling argument could be made that a reduction in the overall Latino voting strength in any of the particular districts would have worked a retrogression prohibited by § 5. One of the forces driving the Balderas court’s rejection of a seventh Latino district in the Central and South Texas regions was that very concern. But the Attorney General has concluded, at least as to Plan 1374C, that an increase to the relevant geographic and population locations to accommodate six Latino districts, coupled with a corresponding reduction in the strength of District 15, will not lead to a retrogression of the minority group’s effective exercise of the electoral franchise. It is incumbent upon us, therefore, to assess whether, by embracing that increased geographic area and drawing new districts, the State has drawn its districts in a way that will dilute the strength of the Latino vote therein.

The state is trying to have its cake and eat it too. The Balderas court found no vote dilution based on the totality of the circumstances then before it. The Balder-as court expressly noted that its remedial plan increased the Latino voting strength in District 24 and created a minority opportunity district in District 25 and, in doing so, hardly left “a bleak terrain” for minority voting opportunities. Similarly, under Plan 1151C, Latino voters enjoyed substantial influence over the outcome of the elections in, at a minimum, Districts 10, 11 and 17 which influence, in turn, increased the likelihood that the elected candidates from those districts would be responsive to the needs of the minority *528communities. The presence in 1151C of districts in which the candidates are responsive to the needs of the minority community is a relevant consideration under the Zimmer factors. Under the totality of the circumstances, the failure to create additional opportunities did not lead to a dilution in statewide Latino voting strength. Section 2 is concerned with effects, and the State is unable to point to the presence of these minority influence districts to counterbalance its failure to create a seventh Latino majority-minority district. The State cannot rely on the Balderas order to overcome this particular Section 2 challenge any more than the plaintiffs can rely on it to defeat the Legislature’s ability to redistrict mid-decade. Balderas therefore stands as no opposition to the GI Forum plaintiffs’ Section 2 claim.

III.

My opinion that the State violated § 2 when it dismantled old District 23 and replaced it with new District 25 renders it unnecessary to assess whether the “bacon strip” districts violate the principles set forth in Shaw v. Reno and its progeny.4 Restructuring of the South and Central Texas districts is necessary to remedy that § 2 violation. It is doubtful that the court could, consistent with Upham v. Seamon, simply remedy the areas adjoining District 23, without treading on the state’s policy as reflected overall in Plan 1374C. It is possible that any new configuration would cure any Shaw issues. For present purposes, it is enough for me to embrace the wisdom of Dr. Alford’s opinion that any Shaw issues present in South and Central Texas under Plan 1374C resulted from the state’s own action in altering District 23 for political purposes. The ripple effect of changes to this area of the map, however, would likely be felt across a large portion of the state. E.g., Abrams v. Johnson, 521 U.S. 74, 86, 117 S.Ct. 1925, 138 L.Ed.2d 285 (1997). Given the primary schedule, and the evidence in the record about the difficulties of conducting the elections even under the plan as enacted by the State, I would grant the Legislature the opportunity to cure these defects and order the elections to be held under Plan 1151C, a plan that is beyond dispute a legal one.

IV.

This is not an easy case. My colleagues’ concerns with interfering with legislative and political prerogatives are not without force, and I agree that politics motivated many of the decisions involved in the case. As difficult as the case is, however, I am unable to agree with the majority that § 2 was not violated when the state enacted Plan 1374C. I therefore dissent from that portion of the court’s opinion.

V.

As I stated at the outset, I reluctantly concur with the court’s judgment insofar as it addresses the claims related to the dismantling of District 24. It is not so much my agreement with the majority’s assessment of the facts that causes me to do so, but rather it is instead the lack of clear guidance from the Supreme Court or the circuits regarding the extent to which Ashcroft’s recognition of the value of coalitional and influence districts carries over into the § 2 context. At the present time, controlling law compels the conclusion reached by the majority.

The question whether § 2 creates influence dilution liability is a close one-made so by language in Ashcroft recognizing the importance and the value of such districts. My philosophical trouble with the controlling law’s sacrosanct view of the 50% rule *529flows from the fact that the Voting Rights Act endeavors to put minority voters on an equal footing in all aspects of the political process. See generally Stanley Pierre-Louis, The Politics of Influence; Recognizing Influence Dilution Claims Under Section 2 of the Voting Rights Act, 62 U. Chi. L.Rev. 1215, 1224 (1995)(“Indeed, Section 2 refers to open participation for minority voters in the ‘political processes leading to nomination or election’ as well as the opportunity ‘to elect representatives of choice.’ ”) Just as minority voters are not immune from the obligation to “pull, haul, and trade to find common political ground,” neither should a redistricting plan have the effect of operating unequally on a minority group’s ability to engage in these very activities.

Participation in the political process is hard work. It is harder for minority groups who have suffered the legacy of a history of official discrimination. We heard compelling testimony from Deralyn Davis, the Chairman Emeritus of the Texas Coalition of Black Democrats, about just how hard it is for minority voters in the state of Texas. She described the grassroots efforts of her and others to build coalitions, mobilize, and increase the African American influence in the state’s political machine. The efforts did not focus on one location in the state, but extended statewide. The testimony of others was consistent.

Under Plan 1000C, in use until the 2000 census, the black citizen voting age population of District 24 was 20.1%. Under Plan 1151C, the court drawn plan, District 24 has a black voting age population of 21.4%. The electorate in the Democratic primary over the last four election cycles has been roughly 64% African American. Both Dr. Lichtman and the State’s expert, Dr. Gad-die, agreed that current District 24 performs for African Americans because those voters control the primary election. The key to the performance of District 24 is the makeup of the balance of the district. It is a political reality that blacks and Latinos in Texas vote largely Democratic in the general elections. They do so because, at least in Texas, the Democratic candidates are generally more responsive to the concerns of these minority communities. Under Texas’s election scheme, in a Democratic leaning district, the key to a minority group’s ability to elect a candidate of its choice on an ongoing basis is found in the group’s practical ability to “pull, haul, and trade to find common political ground” in the Democratic primary. The ability to nominate in such districts is tantamount to the ability to elect.

The Supreme Court recently reminded us that “the power to influence the political process is not limited to winning elections.” Ashcroft, — U.S. at —, 123 S.Ct. at 2512. Ashcroft noted that “various studies have suggested that the most effective way to maximize minority voting strength may be to create more influence or coalition districts.” Id. at —, 123 S.Ct. at 2512-13 (citing David Lublin, Racial Redistricting and African-American Representation; A Critique of “Do Majority-Minority Districts Maximize Substantive Black Representation in Congress?,” 93 Am. Pol. Sci. Rev. 183, 185 (1999); Charles Cameron, David Epstein & Sharyn O’Halloran, Do Majority-Minority Districts Maximize Substantive Black Representation in Congress?, 90 Am. Pol. Sci. Rev. 794, 808 (1996); C. Swain, Black Faces, Black Interests 193-234 (1995); and Bernard Grofman, Lisa Handley, & David Lublin, Drawing Effective Minority Districts; A Conceptual Framework and Some Empirical Evidence, 79 N.C.L.Rev. 1383 (2001)). If the most effective way to maximize minority voting strength is to create more influence or coalition districts, then the most effective way to minimize *530minority voting strength may be to dismantle those districts in which a minority group has proven successful in its efforts to “pull, haul, and trade to find common political ground.”5

Ashcroft leaves the state with flexibility to choose the method of creating fewer black majority-minority districts in favor of a greater number of influence or coalition districts. Alternatively, a state enjoys the flexibility to strengthen its existing majority-minority districts to ensure those districts will continue to elect candidates of choice of the minority community. In the Dallas/Fort Worth area, Texas chose neither route. Texas identified District 24, in which the minority community played a decisive role in the nomination and election processes and, because that group was not a literal majority and elected an Anglo Democrat, rearranged a cohesive community of those voters not into new District 12 (where they might have an impact on the general election), but into new District 26. This arrangement has the practical effect of eliminating the minority voters’ political influence. But it has a much larger effect: by treating the minority group in this manner, the state action, I fear, will have the effect of destroying the minority group’s hope. It is not accidental that, unlike the plan in Ashcroft, the present plan had overwhelming opposition from the minority legislators. The law’s insistence on a 50% majority shields this move, but at the same time it ignores the political and social reality that there is more to participation in the electoral process than winning elections.

It is no answer that the disparate destruction of such coalitions might be tough to identify and even harder to remedy. Our charge is not to decide just the easy cases: it is to apply and, when necessary, enforce the protections of the Voting Rights Act. The question, difficult as it may be to answer, should be whether the treatment of such minority coalitions disproportionately affects those voters’ ability to participate in the political process. The remedy need not require the protection inviolate of an old district, but it might require the creation of a more competitive one.

The treatment of the minority coalitions in old District 24 was inconsistent with the purposes of the Voting Rights Act. The evidence demonstrates that District 24 under Plan 1151C and Plan 1000C functioned as a district that fostered our progression to a society that is no longer fixated on race. Under this record, the black voters in old District 24 repeatedly nominated and helped to elect an Anglo congressman with an impeccable record of responsiveness to the minority community. To this end, I view the fact that Congressman Frost has not had a primary opponent to reflect favorably on his record. I would credit the testimony of Mayor Ron Kirk and Senator Royce West. Their testimony concerning how District 24 functions for the African American community is persuasive. Although the State, for § 5 purposes, created a new black opportunity district in the Houston area, this is of little consolation to the minority voters in old District 24, particularly the African American community in Tarrant County. The political influence of that minority community has been diluted, understanding, as *531did the court in Ashcroft, that power at the polls and participation in the political process is not always measured by mathematical majorities. Given the current state of the law, however, I join the court’s judgment denying relief on the plaintiffs’ claim that the State violated § 2 by the dismantling of District 24. For the same reasons, I necessarily join the court’s judgment denying relief with respect to the plaintiffs’ claims that the changes to Districts 1, 2, 4, 9, 10, 11 and 17 resulted in cognizable influence dilution claims.

Various voters, members of Congress, the City of Austin, Texas, the GI Forum, the League of United Latin American Citizens (“LULAC”), and voters of Cherokee County challenge congressional redistricting set forth in Plan 1374C, enacted into law by the Texas Legislature on October 12, 2003,1 and precleared by the Department of Justice on December 19, 2003. Plaintiffs2 allege that Plan 1374C is invalid because (1) Texas may not redistrict mid-decade; (2) the Plan unconstitutionally discriminates on the basis of race; (3) the Plan is an unconstitutional partisan gerrymander; and (4) various districts in Plan 1374C dilute the voting strength of minorities in violation of § 2 of the Voting Rights Act.

We hold that Plaintiffs have failed to prove that the State statute prescribing the lines for the thirty-two congressional seats in Texas violates the United States Constitution or fails to comply with § 2 of the Voting Rights Act. We also reject Plaintiffs’ argument that the Texas Legislature lacked authority to draw new districts after a federal court drew them following the 2000 census.

We decide only the legality of Plan 1374C, not its wisdom. Whether the Texas Legislature has acted in the best interest of Texas is a judgment that belongs to the people who elected the officials whose act is challenged in this case. Nor does the reality that this is a reprise of the act of the 1991 State Legislature weigh with the court’s decision beyond its marker of the impact of the computer-drawn map. This extraordinary change in the ability to slice thin the lines brings welcome assistance, but comes with a high cost of creating much greater potential for abuse. Congress can assist by banning mid-decade redistricting, which it has the clear constitutional authority to do, as many states have done. In Texas, the phenomenon is new but already old. The larger lesson of 1991 and 2003 is that the only check upon these grasps of power lie with the voter. But, perversely, these seizures entail political moves that too often dance close to avoiding the recall of the disagreeing voter. We know it is rough and tumble politics, and we are ever mindful that the judiciary must call the fouls without participating in the game. We must nonetheless express concern that in the age of technology this is a very different game.

Part I presents the factual background of the case. Part II addresses whether Texas had the legislative authority to draw new district lines mid-decade. Part III addresses generic claims that challenge the map as a whole, namely, claims of racial discrimination and partisan gerrymandering asserted to be unconstitutionally extreme. Part IV lays out the legal principles governing our analysis of Plaintiffs’ more specific claims. Part V addresses the § 2 vote dilution claims as directed toward the Dallas-Fort Worth area, as well as the other potential influence districts in East and Central Texas. Finally, Part VI addresses the § 2 vote dilution and Shaw claims directed at districts drawn in South and West Texas.

I

The U.S. Census Bureau released the 2000 decennial census in March 2001. As a result of its population growth, Texas was due two additional seats in the House of Representatives, bringing its total to thirty-two. Texas in turn had to draw thirty-two equipopulous districts to account for its additional representation and to meet the constitutional requirement of one man, one vote. Under Texas law, the *458Texas Legislature had the task of drawing the districts.3

Despite the imminency of state primary elections, the 77th State Legislature failed to adopt a redistricting plan. Lawsuits in state and federal court followed. Voters and others requested that the court draw a new map. The Balderas court deferred to state court efforts to adopt a state redistricting plan. When these state court efforts failed, we recognized that the State’s existing congressional districts were unconstitutionally malapportioned and reluctantly accepted the duty to prepare a new, constitutional plan.

Without a baseline state plan in place, the court invited the parties to submit redistricting recommendations. Following a bench trial, the panel applied neutral districting factors and adopted Plan 1151C to govern the State’s 2002 elections. The panel refused suggestions not required by law and rejected policy choices better left to legislative consideration.

Balderas ultimately ordered that Plan 1151C would govern the 2002 congressional elections.4 Certain plaintiffs representing Hispanic voters appealed the decision, arguing that the panel erred by not drawing an additional Hispanic district in the Southwest region of the state. The Supreme Court summarily affirmed.5 As a result of the 2002 elections, the Texas congressional delegation included seventeen Democrats and fifteen Republicans. However, with their newly drawn state districts, legislative Republicans gained control over both houses of the Texas State Legislature, as well as control over all prominent Executive Branch positions.

The Texas Legislature revisited redistricting in 2003. The Legislature was unable to adopt a new plan during the 2003 regular session, in part because Democratic House members, by absenting themselves, denied a quorum. Governor Perry called the Legislature into special session. During the first special session, the House approved a new congressional map, but the Senate failed to do so because its “two-thirds” supermajority rule permitted the Democrats to block a vote. To break the impasse, Lieutenant Governor Dewhurst announced that he would suspend operation of the two-thirds rule in any future special session considering congressional redistricting legislation. Although Democratic legislators again attempted to prevent formation of a quorum, the 78th Legislature ultimately was able to accomplish during its third special session what the 77th Legislature could not: pass a congressional redistricting plan, Plan 1374C.

II

Plaintiffs argue that Texas lacks the power, under either the Constitution or the election statutes, to redraw congressional districts in the middle of the decade. Some Plaintiffs find this limitation implicit in the text of the Elections Clause, while others urge that Congress has affirmatively limited state authority to redistrict by § 2c of Title II.6 A third strain of arguments focuses on the Balderas judgment and asserts either that the judgment collaterally estops the State from enacting a new plan or that the judgment exhausted the State’s authority to redistrict.

Although there are compelling arguments why it would be good policy for states to abstain from drawing district lines mid-decade, Plaintiffs ultimately fail to provide any authority — constitutional, statutory, or judicial — demonstrating that mid-decade redistricting is forbidden in Texas. • In fact, what meager authority we have found seems to allow the states to redraw lines mid-decade, at least where a court drew the existing lines within the *459decade. As we will explain, the Elections Clause of the Constitution grants states broad power to regulate the “time, place, and manner of holding elections for Senators and Representatives.”7 Congress has the power to override state regulations or to impose rules of its own, but it has not chosen to limit redistricting to the period immediately following the release of the decennial census. Judicial decisions, both by the Supreme Court and by district courts throughout the country, have allowed and even invited states to redraw district lines following a court’s action.

Against this backdrop of authority, we cannot agree that either the Constitution or the voting statutes restricts the states to once-a-decade redistricting. We therefore reject the argument that the Texas Legislature had no authority to draw the lines of congressional districts and deny Plaintiffs’ Motion for Summary Judgment.

A

The Constitution of the United States delegates to states the power to develop procedures governing congressional elections by the Elections Clause. It provides:

The times, places and manner of holding elections for Senators and Representatives, shall be prescribed in each state by the legislature thereof; but the Congress may at any time by law make or alter such regulations, except as to the places of choosing Senators.8

This provision delegates to state legislatures both the power and responsibility to redraw congressional voting districts.9 States do not possess this authority as an incident of them sovereignty. Rather, the Elections Clause delegates power to the states in broad terms. While states may only enact “time, places and manner” regulations, the text does not define or otherwise limit the states’ discretion. Nonetheless, Congress may, if it chooses, make regulations governing the “times, places and manner” of holding elections or alter regulations enacted by state legislatures. This reservation to Congress, however, is not a direct limitation on the scope of the states’ authority; rather, it allows Congress to override state election decisions or to enact regulations of its own. Unless and until Congress chooses to act, the states’ power to redistrict remains unlimited by constitutional text.

Plaintiffs would read an implicit, temporal limitation into the text of the Elections Clause, but the argument is empty. The argument is that the Elections Clause allows Congress to pass laws regulating elections “at any time,” but does not explicitly allow states to act at any time. Plaintiffs reason that, by failing to include the phrase “at any time” within the grant of power to states, the Elections Clause implicitly denies that power. Hence, they conclude, the Elections Clause allows states to draw districts only once, immediately after the release of each decennial census.

We are unpersuaded. The argument tortures the text of the Clause, which by its clear terms has no such limitation. That Congress may exercise its power at any time says nothing of the states’ power to enact election regulations, especially when the states are given this authority in terms that suggest no restriction beyond those that Congress may impose. To paraphrase the argument, if the Framers had intended to limit the states’ power in such a specific way, surely they would have done so explicitly. This is just too convenient and tailored. For the first fifty years of our Nation’s history, it was not *460uncommon for representatives to be chosen in statewide, at-large elections;10 states often did not divide into congressional districts. The notion that the Elections Clause somehow embodies an implicit limitation on mid-decade redistricting is therefore anachronistic at best; presumably, it never entered the Framers’ minds. Even today, there is no constitutional requirement that states must necessarily subdivide their territory into districts;11 the requirement that states draw districts is largely a creature of statute.12

Even if the Elections Clause did not give states the power to prescribe election regulations at “any” time, Plaintiffs do not explain why we should read the Clause to allow states to exercise election power only one time after the census or how any such interpretation could find mooring in the text of the Constitution. What Plaintiffs ask us to do, then, is not simply to add a single limitation to the Elections Clause’s grant of power; they ask that we create, out of whole cloth, a detailed scheme for states to exercise their constitutional authority. This we cannot do.

Judicial decisions have implicitly rejected the notion that a state may impose only one redistricting map each decade. While no court has, to our knowledge, explicitly addressed whether states have the power to do so under the Constitution,13 innumerable decisions have either assumed that a state legislature may draw new lines mid-decade or have invited a state to do so after the court has drawn a map in a remedial role.14

*461The Supreme Court has intimated on several occasions that states may redistrict mid-decade following court action. In Upham v. Seamon,15 the Court noted that the parties “urged that because the District Court’s plan is only an interim plan and is subject to replacement by the legislature in 1983, the injury to appellants, if any, will not be irreparable.”16 Similarly, in Branch v. Smith, the Court noted that the district court’s holding that state courts could not constitutionally create redistricting plans, a holding the Court vacated, was not “binding upon state and federal officials should Mississippi seek in the future to administer a redistricting plan adopted by the Chancery Court.” 17 The Court’s most vivid statement on the topic came in Wise v. Lipscomb:

Legislative bodies should not leave their reapportionment tasks to the federal courts; but when those with legislative responsibilities do not respond, or the imminence of a state election makes it impractical for them to do so, it becomes the “unwelcome obligation” of the federal court to devise and impose a reapportionment plan pending later legislative action.18

The Court’s language contemplates that any federal court plan must give way to later legislative redistricting efforts. And, of course, the displaced plan here was judicially crafted. No legislative plan is being displaced.

Given this authority and the broad language of the Elections Clause, we conclude that the Elections Clause itself — the provision in the Constitution that grants states the authority to redistrict — does not limit states to redistricting once per decade, particularly where, as here, the State’s action follows a court-imposed map. If any such limitation is to be found, then it must be found elsewhere.

B

Some Plaintiffs would locate limitations in other clauses in the Constitution, most notably the Census Clause.19 Those Plaintiffs point to two phrases which, we are told, prevent a state from redistricting any time it chooses. The Census Clause provides, in pertinent part:

Representatives ... shall be apportioned among the several States ... according to their respective numbers. The actual Enumeration shall be made within three years after the first meeting of the Congress of the United States, and within every subsequent term' often years, in such manner as they shall by law direct.20

*462Plaintiffs argue that the phrase “according to their respective numbers” and the sentence requiring enumeration every ten years together impose affirmative limitations on the states’ power to redistrict.

We disagree. The Census Clause by its terms applies to the apportionment of seats in the House of Representatives among the states. It ensures that no state is over-represented in the House by linking each state’s delegation to the state’s population. The Clause says nothing about how district lines must be drawn. It is true, of course, that the Census Clause affects the states’ obligation to redistrict. When the census is released every ten years, states are required to redistrict in order to accommodate changes in population and to bring its districts into conformity with the Equal Protection rule of one man, one vote. But the Census Clause does not expressly limit the states’ ability to redistrict more frequently. Indeed, the Census Clause does not mention the states or their power to redistrict,21 and we fail to see how it can limit a power it never references.

Plaintiffs concede that the constitutional text itself is silent regarding repeated reapportionment. They nonetheless insist that the Supreme Court’s decision in U.S. Term Limits, Inc. v. Thornton22 prevents States from “adding” to the Constitution a provision entitling states to redistrict mid-decade. Plaintiffs’ reliance on U.S. Term Limits is misplaced.

U.S. Term Limits was not based on the Elections or Census Clauses, but on the Qualifications Clause of Article I.23 In U.S. Term Limits, the Court rejected an effort by Arkansas to impose term limits on its Representatives because a state cannot add qualifications to those enumerated in the Constitution. The Court noted that, “in certain limited contexts, the power to regulate the incidents of the federal system is not a reserved power of the States, but rather is delegated by the Constitution.”24 “In the absence of any constitutional delegation to the States of power to add qualifications to those enumerated in the Constitution, such a power does not exist.”25 Drawing on this language, Plaintiffs urge that, by undertaking mid-decade redistricting, Texas has in effect “added” to the text of the Elections Clause a power to redistrict intradecenially.

We disagree. As a preliminary matter, Texas has “added” nothing to the text of the Constitution by redrawing its district lines mid-decade. The Elections Clause is a broad grant of authority to the states that is checked only by the power of Congress to make or alter voting regulations. Nowhere in the text of the Elections Clause or in judicial interpretations is there a limitation of the frequency with which states may exercise their power. *463Since the power given the states is broad enough to encompass mid-decade redistricting, it cannot be fairly said — as Plaintiffs assert — that Texas has “added” anything to the Constitution’s text. U.S. Term Limits, by contrast, dealt with a very different situation. It dealt with the Qualifications Clause, a provision that is of a “precise, limited nature.”26 By adding a term limit, Arkansas engrafted a provision into the Qualifications Clause that was not there before. The same cannot be said here. The Qualifications Clause in U.S. Term Limits by its terms gave the states no role to play in setting the qualifications of representatives.27 The Elections Clause is different: it appeals to both state legislatures and Congress to set the “time, manner and place” of holding elections. When a state exercises this authority, it adds nothing to the Constitution.

Fairly viewed, it is Plaintiffs who seek to “add” to the Constitution. They ask us to add an implicit limitation to the Elections Clause that states may prescribe the “times, places and manner” of holding elections only after each decennial census. There is no basis for this addition, either in the text of the Constitution or in court decisions interpreting it. In sum, neither the Census Clause nor the Qualifications Clause limits state power to redraw district lines intradecennially.

C

The Elections Clause grants Congress the power to pass voting regulations or to alter voting regulations enacted by the states. Several Plaintiffs argue that Congress has exercised its power to limit the authority of the states to redistrict.

The most relevant statutory provisions are in Title 2. Section 2a specifies that the President must inform Congress after each decennial census of the population of each state and the corresponding number of representatives each state is entitled to send to the House of Representatives.28 Section 2c requires every state entitled to more than one representative under these census figures to create a number of districts equivalent to the number of representatives it sends to the House.29

Congress has given courts a significant role in redrawing district lines. Should a state legislature not redistrict after the decennial census, for example, a court is empowered to remedy any defects *464in the state’s maps.30 At the same time, the Supreme Court has emphasized that primary responsibility for drawing the lines of congressional districts remains with state legislatures.31 The Court has reaffirmed this principle in a series of decisions constraining federal courts in redistricting cases. The Court, for example, insists that the judiciary defer to legislative districting if the legislative plan meets population equality and racial fairness standards applicable to court-ordered plans.32 In addition, the judicial role is remedial; courts are not to replace valid legislative judgments with their own preferences.33 Absent evidence that a state will fail to perform its redistricting duty in a timely fashion, a federal court can neither obstruct the State’s redistricting efforts nor allow federal litigation to impede it.34 Similarly, courts conducting redistricting are obliged to honor the State’s redistricting traditions.

Plaintiffs assert that § 2c constrains the power of states to redraw district lines at will. Their argument comprises three basic steps. First, in § 2c Congress revoked the power granted to state legislatures by the Elections Clause and delegated a far more limited power. Second, they urge that § 2c allows redistricting once after the decennial census. As a result, they urge that when Balderas imposed Plan 1151C, the judgment effectively “used up” the redistrieting power delegated to the states through § 2c. Under this view, Plan 1374C is invalid because Texas had no power to enact it once Balderas installed Plan 1151C.

We are not persuaded. First, we cannot agree that by passing § 2c, Congress revoked the authority granted states by the Elections Clause. To be sure, § 2c constrains the redistricting decisions that states can make, but it cannot fairly be said to revoke the states’ power. Plaintiffs advance a specie of preemption argument: that by passing legislation that relates in some way to congressional districting, Congress has effectively usurped the entire redistricting field.

This interpretation of § 2c ignores the text of § 2c and misreads the Elections Clause. Section 2c has no language suggesting that Congress is “revoking” the authority granted by the Elections Clause, or even that Congress is “redelegating” a more limited authority. If Congress wishes to revoke the states’ redistricting authority, it must do so clearly.35 More*465over, Plaintiffs’ argument tacitly assumes that any congressional regulation relating to election procedures automatically revokes the broad authority given states under the Elections Clause. The structure of the Clause, however, suggests that the primary source of election regulation is state law, federal law supplementing state procedures or overriding them only when necessary. Reading § 2c for what it is — a congressional regulation imposing a single election requirement on the states — preserves the relative roles of Congress and the states under the Elections Clause.

Second, even if § 2c did somehow revoke and redelegate redistricting authority, we disagree that § 2c would allow redistricting only on the decennium. Plaintiffs base their argument on the text of § 2c, which provides in pertinent part:

In each State entitled ... to more than one Representative under an apportionment made pursuant to the provisions of section 2a(a) of this title [the decennial census], there shall be established by law a number of districts equal to the number of Representatives to which such State is so entitled, and Representatives shall be elected only from districts so established, no district to elect more than one Representative .... 36

Plaintiffs argue that this provision directly links the “time” at which the state must redistrict to the “mode” in which it must exercise its redistricting power. That is, the clause requiring states to establish districts immediately follows the reference to the decennial census.37 Plaintiffs also reference several other federal statutes imposing “time” and “manner” restrictions, apparently in an effort to bolster the notion that federal law restricts a state’s ability to redistrict frequently.

While it is true that states are under an obligation to redistrict after each census, we find nothing in § 2c that limits the frequency with which they may do so. It would have been remarkably easy for Congress to impose such a limitation in the text of § 2c, but it did not. It merely required states with more than one representative to divide their territory into a like number of districts.38 Other courts have similarly failed to find such a limitation in § 2c. As we noted above, numerous courts have either allowed or invited state legislatures to enact redistricting plans in the middle of the decade when a court has previously imposed a plan.39 Section 2c was in force when each of these decisions was handed down.

The final step in Plaintiffs’ § 2c argument is similarly flawed. Plaintiffs conclude that Balderas, by establishing a constitutional redistricting plan, “used up” the state’s constitutional authority to redistrict. Plaintiffs frame their argument by again focusing on the meaning of § 2c— which, they assert, reveals that court redistricting is constitutionally “equivalent” for purposes of Article I, Section 4 to state legislative redistricting.40 We agree that *466court-drawn maps are functionally equivalent to legislative maps, although there are key differences between the two. But our agreement on this point is of little help to Plaintiffs because they fail to persuade that a state cannot redraw district lines after a valid court-imposed plan is in place. That is, Balderas could only “use up” the State’s constitutional authority to redistrict if the state is somehow constrained to draw district lines only once per decade. We have rejected Plaintiffs’ argument that redistricting following a judicially imposed plan is forbidden, and we do not reach Plaintiffs’ Balderas claim.

D

Plaintiffs also assert that Texas “tradition” prevents the State Legislature from redrawing district lines in the middle of the decade. They argue that, under Supreme Court precedent, the State Legislature was bound to follow its traditional redistricting principles in creating a new map. Since Texas does not have either a history or tradition of mid-decade redistricting, Plan 1374C is said to be invalid.

Plaintiffs’ argument misreads Supreme Court precedent. Although the Supreme Court has required courts to use a state’s “districting traditions” when drafting voting maps, the Court has never held that a state legislature is bound to follow its prior districting practices indefinitely. Indeed, “tradition” normally fills a very different role in redistricting suits.41 Plaintiffs cite two Supreme Court decisions, White v. Weiser42 and Branch v. Smith,43 to support their argument, but neither decision holds that states are bound to follow state tradition in drawing maps. In White, the Supreme Court held that federal courts, not state legislatures, must abide state districting traditions; the Court iterated that legislatures, not courts, have “primary jurisdiction” over reapportionment, and reinforced the notion that court intervention in the redistricting process is meant to be minor and remedial.44 Branch v. Smith is also inapposite. In Branch, the Court attempted to reconcile § 2c with the seemingly conflicting requirements of § 2a(c)(5).45 The Court explained that when a “federal court redistricts a State in a manner that complies with that State’s substantive districting principles, it does so ‘in the manner provided by the law there*467of.’ ”46 Plaintiffs cite this passage as evidence that states are constrained by tradition. Presumably, their argument is that since the Court held that a federal court redistricts “in the manner provided by the law [of the state]” only when it follows the state’s districting traditions, so too a state legislature can only redistrict “in the manner provided by the law [of the state]” when it follows the state’s districting traditions. Plaintiffs misread this passage. Branch does not hold that a state must follow its redistricting traditions without deviation. Rather, the passage is focused solely on the power of federal courts. It specifically holds that if a federal court redistricts a state using the state’s traditions and preferences, the court does so “in the manner provided by the law” of the state, such that § 2a(c)(5)’s requirement that at-large districts be used is not invoked.

In any event, it would be illogical to require a state legislature to adhere strictly to the state’s districting principles whenever it undertook to redraw the state’s map. Any such rule would, in effect, freeze the state’s districting traditions in place.47 We can find no reasoned basis for such a rule.48

*468E

Plaintiffs’ final argument focuses on the effect of the Balderas judgment. They argue that Balderas was a final judgment that is binding on the State because it was a party to the proceeding. They argue that the State in enacting Plan 1374C is attempting to avoid the judgment in that case, and that the State is collaterally estopped from contesting the use of Plan 1151C.

The prerequisites for collateral es-toppel are not met here. Issue preclusion has four basic requirements: (1) the issue must be identical to an issue involved in prior litigation; (2) the issue must have been fully and vigorously litigated; (3) the issue must have been necessarily decided in the prior litigation; and (4) special circumstances must not render preclusion inappropriate or unfair.49 The first two requirements are not met here. It is true, as Plaintiffs argue, that the two majority-minority districts — Districts 18 and 30— are involved in both cases, as are several other features of the two plans. But the Balderas court’s tasks were different: it had to bring the district map into line with the equal population rule, while accommodating the two new congressional districts and obeying the Voting Rights Act. The issues are different here. We must first decide whether the State has the constitutional power under the Elections Clause or § 2c to redraw district lines mid-decade. This issue did not arise in Balderas. We also must examine a never-before-considered legislative districting plan, Plan 1374C, and decide whether it passes muster under the Constitution and the Voting Rights Act. We find no merit in Plaintiffs’ collateral estoppel argument.

F

Perhaps the most compelling arguments offered by Plaintiffs against mid-decade redistricting focus on the impropriety — rather than the illegality — of frequent redistricting. A significant portion of Plaintiffs’ arguments raise policy concerns. For example, Plaintiffs argue that frequent redrawing of district lines will undermine democratic accountability and exact a heavy cost on state independence as federal congressional leaders exert their influence to shape state districting behavior.

As persuasive as these arguments may be, they are directed to the wrong forum. If Congress chooses to ban intradecennial redistricting, it has the power to do so under the Elections Clause. We have found no provision in either the U.S. Constitution, federal law, or state law that proscribes mid-decade redistricting, and our mandate ends there.

G

We deny Plaintiffs’ Motions to Dismiss and Motions for Summary Judgment on *469the issue of mid-decade redistricting and collateral estoppel.

III

Turning to the merits, we will first consider Plaintiffs’ claims in their two most sweeping forms. First, Plaintiffs argue that the proposed plan must be set aside in its entirety because it is laced with impermissible racial discrimination against Blacks and Latinos in violation of the Equal Protection Clause of the United States Constitution. Second, they allege that Plan 1374C is an impermissible partisan gerrymander. After addressing these allegations, we will then discuss a kindred but “analytically distinct claim” of racial discrimination under the Equal Protection Clause that in the drawing of various districts the Legislature was predominately motivated by race. We will reject the broad-based claims at the outset and return to these Shaw claims in our consideration of the more focused claims under § 2 leveled against specific districts. That process will amplify the findings that underpin our conclusion that Plaintiffs have failed to prove purposeful racial discrimination.

A

Since Washington v. Davis,50 a claimed denial of Equal Protection has required proof that discrimination was purposeful; differential or adverse impact alone is not sufficient. In Davis, the Supreme Court considered an employment discrimination claim brought under the Equal Protection Clause in the District of Columbia before Title VII was extended to the District. Writing for the Court, Justice White rejected the argument that a party alleging racial discrimination under the Equal Protection Clause could focus solely on the racially differential impact of the challenged state practice.51 He explained that the Court had “never held that the constitutional standard for adjudicating claims of invidious racial discrimination [was] identical to the standards applicable under Title VII,”52 which in certain circumstances allowed the adverse impact upon a protected minority to constitute sufficient proof of a statutory violation. He concluded that the Equal Protection Clause required more; it demanded proof that the challenged state action was intended to be discriminatory. Davis marked only the first step in the Court’s analysis of Equal Protection claims, and in its wake came a range of questions, including questions about the allocations of the burden of proof and about the character of proof demanded by the requirement that racial discrimination be purposeful.

The next term, in Arlington Heights v. Metropolitan Housing Corp.,53 the Court repudiated the Seventh Circuit’s emphasis on the adverse impact of a challenged zoning decision in a Chicago suburb, rather than its purpose. Writing for the majority, Justice Powell — while acknowledging that proof of purpose would seldom be easy — explained that it was not necessary to prove that a decision was motivated by a single concern, or even that a particular purpose was the dominant or primary one. There need only be proof “that a discriminatory purpose has been a motivating factor in the decision.”54 The inquiry, he explained, may start with the impact of the *470legislative act, which while not alone sufficient to prove purpose remains relevant, and continue to the exploration of the act’s history, including any contemporary statements by members of the decision-making body.

Two years later in Personnel Administrator of Mass. v. Feeney,55 the Court upheld a Massachusetts statute granting lifetime preferences to veterans for civil service positions. The Court rejected claims that because few women could qualify, the statute discriminated against women in violation of the Equal Protection Clause. Writing for the majority, Justice Stewart pointed to findings of the district court that the statute had a legitimate purpose — awarding benefits to veterans — and was not a pretext for discriminating against women. He then offered a critical observation, one that proved to be a powerful and enduring feature of Equal Protection Clause jurisprudence:

It would [be] disingenuous to say that the adverse consequences of this legislation for women were unintended, in the sense that they were not volitional or in the sense that they were not foreseeable.

“Discriminatory purpose,” however, implies more than intent as volition or intent as awareness of consequences. It implies that the decision maker, in this case a state legislature, selected or reaffirmed a particular course of action at least in part “because of,” not merely “in spite of,” its adverse effects upon an identifiable group.56

After grappling with the explicit remedial use of race by courts, legislative bodies, and various federal and state institutions, the Court turned to redistricting plans drawn to enhance the opportunity of minorities protected by the Voting Rights Act where race had become more than another of many necessary considerations in line-drawing. Justice Ginsberg put it succinctly in Miller v. Johnson:

Two Terms ago, in Shaw v. Reno, this Court took up a claim “analytically distinct” from a vote dilution claim. Shaw authorized judicial intervention in “extremely irregular” apportionments, in which the legislature cast aside traditional districting practices to consider race alone — in the Shaw case, to create a district in North Carolina in which African-Americans would compose a majority of the voters.57

In short, Miller instructs that we are to engage in a searching review of district lines “predominantly motivated” by race when a state subordinates traditional districting practices to race.

Plaintiffs have not proven their claim of racial discrimination. There is little question but that the single-minded purpose of the Texas Legislature in enacting Plan 1374C was to gain partisan advantage. With the Republican sweep of statewide offices in 2000 came control of *471the Legislative Redistricting Board. The Legislature was initially unable to redraw district lines for either state legislative or congressional seats. The federal courts drew a congressional district plan58 and after one modification held that the plans for the Texas House of Representatives and Senate drawn by the Republican-controlled Board were legal.59 Although the judicial plan for the congressional districts reflected the growing strength of the Republican Party in Texas, with 20 of the 32 seats offering a Republican advantage, the voters in 2002 split their tickets and elected only 15 Republicans. Six incumbent Anglo Democrats were elected by narrow margins in Republican-leaning districts. With Republicans in control of the State Legislature, they set out to increase their representation in the congressional delegation to 22. As we will explain, all that happened thereafter flowed from this objective, with the give-and-take inherent in the legislative process along the way. The result disadvantaged Democrats. And a high percentage of Blacks and Latinos are Democrats.60

The majority of Plaintiffs’ Equal Protection claims focus on District 26 in Plan 1374C, which reflects Republican refusal to preserve Democratic Congressman Martin Frost’s District 24 while at the same time preserving adjoining Republican districts. To remove Congressman Frost, he needed to lose a large portion of his Democratic constituency, many of whom lived in a predominantly Black area of Tarrant County. This group of voters was taken from previous District 24 and grouped with Denton and Cooke Counties, which are north of Tarrant County. Plaintiffs view the protrusion that reaches down to include the Black Democrats as evidence of intentional racial discrimination.

We disagree. That African-Americans in Texas vote overwhelmingly for Democratic candidates and that various political compromises were reached to arrive at the current district lines belie the assertion that Texas intentionally discriminated against the African-American voters. Bob Davis, who assisted the Texas Senate in drawing various plans and submitting them to the Legislative Redistricting Board, credibly testified as to the various political considerations that combined to result in the lines of current Congressional District 26. First, Representative Kent Grusendorf, who served on the House side of the districting committee, wanted his State House District — which covers the city of Arlington — to remain whole.61 Arlington’s western boundary forms most of the eastern edge of District 26’s southern protrusion. Second, the court-drawn map, Plan 1151C, split State House Representative Glenn Lewis’s District 95 into two different congressional districts. Representative Lewis wanted his district to fall completely within one congressional dis*472trict.62 House District 95 now forms the southern tip of Plan 1374C’s District 26, explaining the southernmost boundaries of District 26. Third, Democrats could not be placed in Congressional District 12 to the west because District 12 would then become “far more Democrat and very marginally Republican, if Republican at all.”63 Finally, Representative Phil King, the chairman of the redistricting bill in the Texas House, wanted Parker and Wise Counties to be included completely in Congresswoman Granger’s District 12.64 But if the Tarrant County population fell in District 12, population would need to be taken out, likely from Parker or Wise counties.

So, the net result was the political consequences of putting that territory, either in District 12 or District 6, were not good. And District 26, which was, in the Court Plan, this area in here, adjacent to it, and so it was placed on the District 26 because the political structure of 26 could handle that particular component of the Tarrant County population and still produce Republican results for District 26.65

We find these unchallenged explanations to be credible, and we find that including the large Democratic area of southeast Tarrant County in District 26 was the sole product of political give-and-take by legislative members over their own state districts and the effort to not create another Democratic district. The actions were not taken because of race; they were taken in spite of it.

Plaintiffs’ expert’s testimony supports our conclusion that politics, not race, drove Plan 1374C. The Jackson Plaintiffs’ expert, Dr. John Alford, professor of political science at Rice University, testified that “one would have a very hard time not recognizing that the State has a very strong partisan motivation in this particular map.”66 Representative Phil King testified that the purpose of the plan was to make the congressional delegation more reflective of state voting trends. The ami-cus brief of the Texas House Democratic Caucus and Representatives John Lewis, Chris Bell, Martin Frost, Sheila Jackson Lee, and Nick Lampson filed in Vieth v. Jubelirer in support of Appellants told the Supreme Court that “[t]he newly dominant Republicans ... decided to redraw the state’s congressional districts solely for the purpose of seizing between five and seven seats from Democratic incumbents.” It was clear from the evidence that this assertion is true. Former Lieutenant Gover*473nor Bill Ratliff, one of the most highly regarded members of the Senate and commonly referred to as the conscience of the Senate, testified that political gain for the Republicans was 110% of the motivation for the Plan, that it was “the entire motivation.” 67 He explained that he is leaving the Senate before the expiration of his term in large part out of disappointment at its partisan turn. In the course of the redistricting bill’s passage, Senator Ratliff, a Republican, refused to abandon the two-thirds rule, which does not allow a bill to come to the floor without the support of 21 members, a practice calculated to promote consensus building.

Plaintiffs nonetheless insist that there was racial discrimination along the way in the specific drawing of the lines. We will examine this less sweeping assertion as we examine the particular districts that are alleged to have been drawn in violation of § 2 of the Voting Rights Act, or in defiance of the principles of Shaw v. Reno.

While keenly aware of the long history of discrimination against Latinos and Blacks in Texas, and recognizing that their long struggle for economic and personal freedom is not over, we are compelled to conclude that this plan was a political product from start to finish. The myriad decisions made during its creation were made in spite of, and not because of, its effects upon Blacks and Latinos. To find otherwise would frustrate the fundamentals of Washington v. Davis and inject the federal courts into a political game for which they are ill-suited, and indeed in which they are charged not to participate under the most basic principles of federalism and separation of power. Concluding that the purpose requirement of the Equal Protection Clause was met on these facts would pass redistricting from the state legislatures and redistricting boards to the federal courts. This is not to say that we wholly withdraw, of course. We simply recognize the fundamental decision in Washington v. Davis that federal judges are not legislative players; we are only the guardians of the boundaries. As Justice Ginsburg put it, while “[legislative dis-tricting is highly political business ... [generations of rank discrimination against [minorities] account for [the court’s] surveillance.”68

Having been watchful, we are not persuaded that this most fundamental boundary of the Equal Protection Clause was crossed. In the redistricting arena, an area that has proven most reluctant to yield discernible standards, there are large incentives to reach for the seeming certainty of the Equal Protection Clause’s familiar condemnation of purposeful racial discrimination and draw upon its comforting moral force, rather than confront the task of developing proper standards or concede their ephemeral political character. To our eyes, the certainty is an illusion, and its deployment to heel radical partisan line-drawing by state legislatures is a mistake. And turning to Washington v. Davis’s insistence of purpose, rather than confronting directly the questions now before the Court in Vieth v. Jubelirer, is just an old Texas two-step.69

*474B

We have no hesitation in concluding that, under current law, this court cannot strike down Plan 1374C on the basis that it is an illegal partisan gerrymander. Seventeen years ago, the Supreme Court held in Davis v. Bandemer70 that an excessively political or partisan gerrymander presents a justiciable issue under the Equal Protection Clause. But the Court was unable to settle upon a manageable standard for addressing such claims. It is now painfully clear that Justice Powell’s concern that the decision offered a “ ‘constitutional green light’ to would-be gerrymandered” has been realized.71 Bandemer insisted upon proof of both discriminatory purpose and discriminatory effect, two requirements that are difficult to meet in the courtroom, particularly as they have been interpreted by the lower courts. That the response to this difficulty must be to develop a new standard does not necessarily follow. The question remains how much of a role the judiciary ought to play in policing the political give-and-take of redistricting. It may be the most difficult question, but it is certainly the most important.

When the Supreme Court resolves Vi-eth, it may choose to retreat from its decision that the question is justiciable, or it may offer more guidance on the nature of the required effect. Perhaps the Court will draw on its experience in developing federal common law in the antitrust arena, which draws a fine line between competitive effect and injury to competition.72 We have learned firsthand what will result if the Court chooses to do neither. Throughout this case we have borne witness to the powerful, conflicting forces nurtured by Bandemer’s holding that the judiciary is to address “excessive” partisan line-drawing, while leaving the issue virtually unenforceable. Inevitably, as the political party in power uses district lines to lock in its present advantage, the party out of power attempts to stretch the protective cover of the Voting Rights Act, urging dilution of critical standards that may, if accepted, aid their party in the short-run but work to the detriment of persons now protected by the Act in the long-run. Casting the appearance both that there is a wrong and that the judiciary stands ready with a remedy, Bandemer as applied steps on legislative incentives for self-correction.

There are ameliorations available short of a grand judicial pronouncement, remedies which are perhaps superior. In Texas, redistricting advantages can be overcome through the political process. The exchange of political advantage between the Democrats in 1990 and the Republicans in 2000 demonstrates this reality. If the Democratic party takes the main statewide offices, Democrats can block a state legislative redistricting plan and write their own through the Legislative Redistricting Board. The resulting State Legislature could then redraw the congressional lines.

Even if the partisan gerrymander issue were not justiciable but Congress allowed *475the drawing of new lines only when there was no extant legal plan, and in any event no more often than once in a decade, the picture would likely be quite different. That the limitation would only reach congressional seats and not state legislatures themselves does not mean that its effects would not be larger. As the record in this case makes clear, Congress often plays a large role in state redistricting, not only of congressional districts but also of the state chambers themselves. Members of Congress work to protect their incumbency and to affect the partisan makeup of the House, with keen interest in the election of members of the State Legislature. Accordingly, a rule that the game is played only once per decade could matter a great deal in the real world of politics. It is fair to ask what if Congress had imposed a once-a-decade rule seventeen years ago, even if Bandemer had dismissed the case as presenting a non-justiciable political question.

Our point is that if the judiciary must rein in partisan gerrymanders, limitations that focus upon the time and circumstance of partisan line-drawing and less upon the “some but not too much” genre of strictures offer the best of an ugly array of choices. Drawing upon the Voting Rights Act jurisprudence to give Bandemer teeth may be the worst of choices.

IV

For convenience, we record some of the general principles to which we will refer in addressing Plaintiffs’ challenges to specific districts. Section 2 of the Voting Rights Act of 1965 (as amended) provides:

(a) No voting qualification or prerequisite to voting or standard, practice, or procedure shall be imposed or applied by any State or political subdivision in a manner which results in a denial or abridgement of the right of any citizen of the United States to vote on account of race or color, or in contravention of the guarantees set fort in section 1973b(f)(2) of this title, as provided in subsection (b) of this section.

(b) A violation of subsection (a) of this section is established if, based on the totality of the circumstances, it is shown that the political processes leading to nomination or election in the State or political subdivision are not equally open to participation by members of a class of citizens protected by subsection (a) of this section in that its members have less opportunity than other members of the electorate to participate in the political process and elect representatives of their choice. The extent to which members of a protected class have been elected to office in the State or political subdivision is one circumstance which may be considered: Provided, That nothing in this section establishes a right to have members of a protected class elected in numbers equal to their proportion in the population.73

To prevail on a claim of vote dilution under § 2, a plaintiff must, as a threshold requirement, satisfy the three now-familiar preconditions set forth in Thornburg v. Gingles: (1) a minority group must be “sufficiently large and geographically compact to constitute a majority in a single-member district”; (2) the group must be “politically cohesive”; and (3) sufficient racial bloc voting must exist such that the white majority usually defeats the minority’s preferred candidate.74 *476Gingles withheld deciding whether there could ever be a showing of potential success without a showing that a clear majority could gather in the absence of the accused practice or structure. The lower courts, with the exception of a recent decision by a divided panel of the First Circuit,75 have strictly enforced the 50% rule, including the Fifth Circuit.76 There are powerful reasons to be exacting, as we will explain, but the facts of this case offer no occasion to decide if there is a tolerable deviation from the rule that a minority must demonstrate that, absent an accused practice or structure, it had the potential to elect a candidate of its choice by proof that it could constitute 50% of the district.

Although satisfying the Gin-gles factors is a prerequisite, meeting the three conditions is alone not enough to prevail under § 2. If they are met, the court is to consider the totality of the circumstances, including a searching inquiry into whether the political process is equally open to minority voters. In Johnson v. De Grandy, the Court explained how to evaluate dilution under a single-member districting plan and discussed the extent of a state’s duty to create additional majority-minority districts under § 2.77 In De Grandy, the plaintiffs attempted to establish liability by pointing to a number of places where minority voters had been “cracked” and placed in majority-white districts where their votes would be “sub*477merged” and ineffective. The Court rejected the dilution claim of the Hispanic and African-American voters because the challenged districting plan provided both sets of voters “rough proportionality,” the opportunity to exercise electoral control in a number of districts that roughly corresponded to their share of the relevant population.78 The Court emphasized that proportionality does not provide a complete or mechanical defense to a § 2 suit.79 The Court also made it clear that proportionality is significant in evaluating dilution claims and has become a preeminent measure of fairness in redistricting.80 Now known as “De Grandy proportionality,” dilution may be found to be absent under the totality of the circumstances when the protected minority groups “constitute effective voting majorities in a number of districts ... substantially proportional to their share in the population.”81

In Shaw v. Reno (“Shaw I”), the Court addressed the constitutionality of a district drawn with race as the predominant motivation, as evidenced by a bizarrely-shaped district drawn to augment minority voting strength.82 The Court held that such a district could be challenged, depending on how it was drawn, under the Equal Protection Clause. In Shaw v. Hunt (“Shaw II”), the Court concluded that a racially gerrymandered district would be subject to strict scrutiny, and that compliance with § 2 could justify a racially gerrymandered district only if the remedial district was narrowly tailored toward that end.83 In so holding, the Court held that a remedy for vote dilution in one part of the state, where it was possible to draw an additional, compact majority-minority district, is not narrowly tailored to comply with § 2 if the remedial district is drawn in a different part of the state where polarized voting also exists, but where a compact district cannot be crafted. In so holding, the Court stated:

Arguing, as appellees do and the District Court did, that the State may draw the district anywhere derives from a misconception of the vote-dilution claim. To accept that the district may be placed anywhere implies that the claim, and hence the coordinate right to an undiluted vote (to cast a ballot equal among voters), belongs to the minority as a group and not to its individual members. It does not.84

Similarly, in De Grandy, the Court stated that one reason for rejecting the inflexible safe harbor rule the state advocated in that case, under which no dilution can *478occur as a matter of law if the percentage of single member districts in which minority voters form an effective majority mirrors the minority voters’ percentage of the relevant population, was that it rested on

an unexplored premise of highly suspect validity: that in any given voting jurisdiction (or portion of that jurisdiction under consideration), the rights of some minority voters under § 2 may be traded off against the rights of other members of the same minority class. Under the State’s view, the most blatant racial gerrymandering in half of a county’s single-member districts would be irrelevant under § 2 if offset by political gerrymandering in the other half, so long as proportionality was the bottom line.85

With these basic principles in mind, we now turn to Plaintiffs’ specific claims.

V

We first examine the § 2 challenges to the districts in Central and East Texas, most notably District 24 in the Dallas-Fort Worth area. As a preliminary matter, it bears emphasis that the majority requirement of the first Gingles precondition cannot be met in these districts by summing Black and Hispanic voter populations.86 Plaintiffs cite Brewer v. Ham87 for the proposition that minority groups may be combined to satisfy Gingles’s majority requirement. Brewer, however, allowed for minority combination when the groups vote cohesively. Here, there is no serious dispute but that Blacks and Hispanics do not vote cohesively in primary elections, where their allegiance is free of party affiliation.88 Minority voters must have the potential to elect in the absence of the accused practice or structure if their claim of injury by that practice or structure is to be sustained.89

A

1

Georgia v. Ashcroft90 is the most recent discussion of the factual and legal distinctions between majority-minority, coalition, and influence districts.91 Majority-minority or “safe” districts are voting districts with a majority of minority voters, making it “highly likely that minority voters will be able to elect the candidate of their choice.”92 Coalition districts are voting districts where minority voters “ ‘are able to form coalitions with voters from other racial and ethnic groups, having no need to be a majority within a single district in order to elect candidates *479of their choice.’ ”93 Influence districts are voting districts “where minority voters may not be able to elect a candidate of choice but can play a substantial, if not decisive, role in the electoral process.”94 The elected representatives in influence districts, as a result of the influence of minority voting, take minority interests into account.95

Georgia tested the limits of these types of districts when it redrew its state senate’s voting districts following the 2000 census. The previous map, which was finally precleared after much litigation, included various safe districts. Shifting its strategy, the Democratic-controlled Legislature “unpacked” three of the safe districts and spread the minority voters to create influence and coalition districts.96 The goal was to increase the overall influence of minorities in Georgia politics. The Justice Department challenged the plan as retrogressive because of the reduction of minorities in the previously safe districts. The district court found the plan retrogressive because the change in the three safe districts created less opportunity for minorities to elect the representatives of their choice.

Georgia appealed the decision to the Supreme Court, arguing that preclearance was appropriate because the new map did not harm the minorities’ “effective exercise of the electoral franchise.” Georgia asked the Court to examine the voting plan as a whole so that the increase in voting strength in other influence and coalition districts could offset the decrease in the three previously safe districts. The Justice Department argued that the district court correctly found retrogression based solely on the decrease in minority population in the three previously safe districts.

Georgia’s argument prevailed. The Court held that states are free to choose the best way to avoid retrogression and ensure equal opportunity to minority voters. To determine whether a new map is retrogressive, the Court examined the state as a whole and considered all the relevant circumstances, “such as the ability of minority voters to elect their candidate of choice, the extent of the minority group’s opportunity to participate in the political process, and the feasibility of creating a nonretrogressive plan.”97 The Court accepted that minority interests may be better served by coalition and influence districts rather than safe districts. The Court noted that although safe districts ensure descriptive representation, they also “isolat[e] minority voters from the rest of the state, and risk[ ] narrowing political influence to only a fraction of political districts”98 — that “various studies have suggested that the most effective way to maximize minority voting strength may be to create more influence or coalitional districts,”99 and, critically, Georgia may make that choice. Georgia’s alteration of its safe districts would have been problematic at best in the 1980s and early 1990s, but the Court allowed it as a valid political choice that Georgia might choose to make in an effort to increase minority voting strength — an alternative to its obligation *480under Gingles to draw a safe majority-minority district.

The Court held that the district court erred in focusing too heavily on the decrease of voting power in the previously safe districts, and in ignoring the offsetting influence and coalition districts. In the Court’s view, the ability of minorities to elect their preferred candidates is important, but not dispositive. Despite the views of the Justice Department, the ACLU, and lower federal courts, the Court held that creating safe districts was not the only means of assuring an effective vote for minorities.

While Ashcroft is a § 5 preclearance case addressing the question of retrogression, the Court’s opinion makes plain that safe districts are no longer untouchable. States previously read Gingles as requiring safe districts to ensure the election of minorities by countering racially polarized voting. But Georgia v. Ashcroft makes clear that safe districts are not necessarily required; states may choose to avoid retrogression by creating coalition and influence districts. Georgia decreased the percentage of minority voters in three previously safe districts to such a degree that, in the opinion of the Justice Department, minorities could not elect their candidate of choices. Yet, the addition of coalition and influence districts countered possible retrogression.

2

All parties here rely upon Georgia v. Ashcroft. Plaintiffs argue that Ashcroft effectively overruled Gingles’s first requirement that “the minority group must be able to demonstrate that it is sufficiently large and geographically compact to constitute a majority in a single-member district.” 100 Plaintiffs base this assertion on Ashcroft’s reasoning that a district may provide effective representation to minorities despite the absence of a mathematical majority.101 From this, Plaintiffs conclude that influence districts must be protected under § 2 of the Voting Rights Act.

In response, Texas argues that under Ashcroft redrawing an influence district does not inevitably dilute minority votes under § 2. Texas argues that if majority-minority districts may be altered without running afoul of the Voting Rights Act, then a fortiori an influence district may be altered. Ashcroft provided states with the flexibility to choose the means of complying with the Voting Rights Act, and although the Court opined that coalition and influence districts may be the most effective means of increasing minority influence, they are not required. Ashcroft did not dictate that a state must maximize both majority-minority and influence districts. Indeed, the question whether federal law requires influence districts has been avoided many times,102 reflecting, the State argues, the Court’s wariness of being drawn further into the political arena. Finally, the State notes that the Supreme Court faced the exact question and rejected it when it summarily affirmed a district court’s rejection of the contention “that the first Gingles precondition is not fully applicable” to districts “where a distinct [minority] group cannot form a majority, but they are sufficiently large and cohesive to effectively influence elections, getting their can*481didate of choice elected.”103

Texas’s argument finds support in the language of Ashcroft:

On one hand, a smaller number of safe majority-minority districts may virtually guarantee the election of a minority group’s preferred candidate in those districts .... And while such districts may result in more “descriptive representation” because the representatives of choice are more likely to mirror the race of the majority of voters in that district, the representation may be limited to fewer areas.

On the other hand, spreading out minority voters over a greater number of districts creates more districts in which minority voters may have the opportunity to elect a candidate of their choice. Such a strategy has the potential to increase “substantive representation” in more districts, by creating coalitions of voters who together will help to achieve the electoral aspirations of the minority group. It also, however, creates the risk that the minority group’s preferred candidate may lose.... Section 5 gives States the flexibility to choose one theory of effective representation over the other.104

As we see it, these choices are for the states to make, as long as they avoid other constitutional and statutory violations under the Equal Protection Clause and § 2. Allowing influence districts to meet a Gin-pies-imposed obligation to create a majority-minority district does not mean that a state must create an influence district in the absence of an obligation to create a majority-minority district. We are not persuaded that Texas had the duty in drawing a new map to trace the old lines to avoid any disruption of coalitions. To so conclude would have profound consequences, freezing ephemeral political alliances, which are the bull’s eyes of partisan redistricting. We will turn to a concrete example. The 24th is a Democratic district, and its “coalitions” are simply minority Blacks joining with majority Anglos voting a Democratic ticket in the general election. Plaintiffs’ understandable efforts to freeze this “coalition” by locating some duty under § 2 not to redraw the district is a transparent effort to use race as a shield from a partisan gerrymander when the district itself was a child of identical efforts to gerrymander. As we will explain, under the new plan, Democrats — both Anglo and Black — lose control of the District 24.

B

We turn first to District 24, located in the Dallas-Fort Worth metroplex in an area referred to as the mid-cities. The congressman from the district is Martin Frost, an Anglo Democrat elected in 1978. This area between Dallas and Fort Worth experienced enormous growth from the early 1960’s, leading to the creation of the 24th in 1972. With some iteration, the district has remained in place until the passage of Plan 1374C in 2003. The old 24th touched both Dallas and Fort Worth and included the communities of Duncan-ville and Cedar Hill on its south side. It also included large plants of Bell Helicopter, General Motors, Northrop Grumman, and Lockheed Martin, as well as the Texas Rangers’ baseball stadium and the Six Flags Over Texas theme park.105

*482Frost, then thirty-two years old, ran for the seat two years after it was created but was defeated in the Democratic primary by Dale Milford, the Anglo incumbent. Congressman Frost won the next race and has held the seat since then. Allied with Speaker Wright of Fort Worth, he quickly won leadership positions, including a seat on the Rules Committee. Frost is a major fund raiser for the Democratic party, and he so effectively chaired a redistricting panel in 1991 that he is widely seen as the architect of the redistricting plan of 1991.106 This plan, drawing on the developing computer technology, is cited by political scientists as the shrewdest of the 1990s.107 Congresswoman Eddie Bernice Johnson, who holds a seat in an adjacent largely Black district, testified that Frost drew the 24th for an Anglo Democrat.108

The judicial plan replaced by 1374C left the 24th largely in place. It had a Black voting age population of 21.4% and a Hispanic voting age population of 33.6%. The latter number falls when citizenship (CVAP) is considered, as it must be. The result is that neither Blacks nor Hispanics have 50% of the voters in the district alone or combined, where they constitute only 46.4%. Measured by the statewide races and the Lieutenant Governor’s race in 2002, District 24 is approximately 60% Democrat, although Al Gore carried the district with 54% of the vote and Governor Bush carried the entire state with 59% of the statewide vote and 61% in the Dallas-Fort Worth metroplex. It bears mention that Houston and Dallas-Fort Worth cast 47% of all votes cast in the state.109 Bush carried only 5% of the Black vote but enjoyed 42% of the Hispanic vote and 73% of the Anglo vote.110

Facing the absence of a minority with 50% of the voting age population of the district singly or combined, Plaintiffs contend that the 24th nonetheless meets the first two preconditions of Gingles because it “functions as a fully effective Black opportunity district.” They offer three reasons why we should not adhere to Perez v. Pasadena, Independent School District111 and Valdespino v. Alamo Heights Independent School District,112 which hold that Gingles requires a cohesive group of minority voters comprising a majority of the adult citizen population in at least one proposed single-member demonstration district. First, insisting on a level of 50% is reasonable only when the court is asked to speculate on the potential performance of a minority district, not when an existing district is shown to actually perform for minorities. Second, Plaintiffs point to a divided decision issued only weeks ago by a panel of the First Circuit, Metts v. Murphy, which dispensed with the 50% rule. The panel’s decision, we note, was expressly withdrawn and vacated when the First Circuit voted to take the case en banc.113 Third, Perez and Valdespino involved challenges to non-partisan at-large election *483systems involving only single elections; here, we examine a single-member district-ing system with primary and general elections. Plaintiffs assert that if a minority can “control” the primary election and then find support with a coalition of minorities in the general election, the Gingles requirements are met and the district cannot be modified. Plaintiffs also offer a fourth, seemingly freestanding principle, urging that in any event District 24 cannot be intentionally dismantled because it is an influence or coalition district.

Defendants respond that § 2 requires a showing of strength at a level of at least 50% is a command of § 2, a requirement that is an inevitable byproduct of the statute’s protection of the ability of minorities to elect representatives of their choice. A minority group lacking a majority cannot elect its candidate of choice, and denying the group a separate district cannot be a denial of any opportunity protected by the Act. Rather, such a group can elect their candidate of choice only with the votes of non-majorities through coalitions, but the Voting Rights Act does not protect such political coalitions. Defendants urge that whether a minority can meet the 50% standard is a quite different question from asking “whether a bi-racial coalition of African Americans and ‘crossover’ Anglos can elect a Black-preferred Black candidate. Any such interpretation would render the first Gingles precondition an entirely superfluous subpart of the third Gingles precondition.” Defendants argue further that the Supreme Court has never held that § 2 protects influence districts and such a proposal is foreclosed by the rationale of Georgia v. Ashcroft. Finally, defendants reply that there has been no showing that the 24th would probably elect a Black candidate, pointing to Congressional District 25 in the judicial Plan 1151C in which the demographics are strikingly similar — 21.4% BVAP and 33.6% HVAP — where an Anglo, Bell, defeated Carroll Robinson, the Black candidate of choice, in the 2002 Democratic primary.

Plaintiffs’ arguments are creative, but they ask this court to do more than well-settled law will allow. To the heart of the matter: the contention that § 2 protects District 24 from redrawing asks us to extend § 2’s protection of Blacks and Latinos from vote dilution to the protection of groups whose cementing force is membership and loyalty to a political party. Gin-gles and the cases that followed it have been keenly aware that the defining concepts of Gingles — numbers and cohesion— are critical to its studied effort to confine the limits of the Act to those situations that dilute minorities’ opportunity to vote without protecting coalitions that may be helpful or even essential to the leveraging of their strength.114 Properly confined, *484the Act implements the fundamentals of factions. Unconfined it reaches into the political market and supports persons joined, not by race, but by common view. Serious constitutional questions loom at that juncture.

We are told that Blacks control the Democratic primary with less than 22% of the CVAP because Anglos and Latinos vote either in the Republican primary or not at all, but return home out of party loyalty in the general election. It is argued that this is a Black opportunity district. More accurately, however, it is a strong Democratic district. That there is no cohesion between Black and Latino voters in the primary contests is beyond serious dispute. Black opportunity here lies in coalitions with Anglos who vote with them in the general election for Democrats. Dr. Lichtman’s calculations produce an Anglo crossover rate of 30.75 (unweighted mean). He conceded that in the general election Black turnout will fall to the range of 31, 32 or perhaps 33%, while Anglo turnout will jump into the 60’s, approximately a two to one margin. Just how far this argument departs from the Gingles construct is exposed by the reminder that such a crossover rate has been found to establish the absence of Anglo bloc voting under Gingles’s third precondition as a matter of law.115

The history of the 24th illustrates that Plaintiffs overstate the impact of the Black Democrats’ control of the primaries. The reality is that Frost has not had a primary opponent since his incumbency began. That no Black candidate has ever filed in a Democratic primary against Frost in a district assertedly controlled by Blacks reflects the accuracy of Congresswoman Johnson’s claim that District 24 was drawn for an Anglo Democrat.116 We have no measure of what Anglo turnout would be in a Democratic primary if Frost were opposed by a Black candidate. Plaintiffs’ asserted control of Blacks in the primary rests upon the shaky ground that much of the dominating Anglo Democratic vote does not bother to vote in the primary with Frost filing as an unchallenged Anglo. Its premise is that Frost challenged by a Black candidate would not result in a return to primary voting by the Anglo Democrats. In short, that Anglo Democrats control this district is the most rational conclusion.

Nor is the cohesiveness of this 21.6% black voting age population clear. Plaintiffs’ expert relied heavily upon the high vote in the district for Ron Kirk, an African-American and former popular mayor of Dallas, in his race for the United States Senate. Dr. Lichtman resisted the suggestion that the “friends and neighbors” effect was distorting the results with the counter suggestion that the effect would be offset by Kirk’s opponent in the Democratic primary, Victor Morales. That explanation is not persuasive. Morales is actually from Crandall, a town of 3,000 people some 75 miles away. In the 1998 Attorney General race, Judge Morris Overstreet, a widely known, respected, and distin*485guished lawyer and judge, took 66% of the Black vote in Dallas County and 76% of the Black vote in Tarrant County. Julius Whittier, a Black candidate in the 2002 race for the Texas Court of Criminal Appeals, received 32% of the Anglo vote and 40% of the Black vote. In short, whether Blacks vote cohesively in the primary is far from certain.

That Blacks wield influence in the district is plain. Perhaps recognizing the difficulty of arguing that some specie of a relaxed Gingles construct imposes an obligation to create an influence or coalition district, Plaintiffs offer a fall-back alternative: that as an “influence district,” the 24th cannot be “intentionally” redrawn. But this turns the principle of Georgia v. Ashcroft on its head. As we have explained, Ashcroft gave states greater latitude in complying with the Voting Rights Act. The Court recognized that while Gin-gles in application has resulted in significant gains in elected positions held by minorities, with the keener edge of computer drawn lines, it has also created “safe” white and minority districts, significantly reducing the necessity for the pull and tug to secure support from those with political influence. By locking in safe districts for minorities and Anglos alike— instead of encouraging competition and the pull and tug of politics — this legal regime began to undermine the core idea that the House of Representatives would be the branch of government directly responsive to the people. Ashcroft responds to these and other political realities with greater concern for forces that divide and fraction-alize, especially those that would divide along racial lines. We do not read Ashcroft as fencing even more territory from state legislative reach. Considering that District 24 as a pure influence district is unprotected by § 2, we are persuaded that alterations to it raised questions primarily of § 5, which have been answered by the Department of Justice.

C

Although the bulk of Plaintiffs’ argument focuses on District 24, Plaintiffs also assert that Plan 1374C violates § 2 by modifying minority influence districts in Central and East Texas, specifically districts 1, 2, 4, 9, 10, 11, and 17 from Plan 1151C. According to Plaintiffs, the minority population in these districts will have only a minimal role in the electoral process under the new plan.

However, as the State points out, Plaintiffs never argue — and certainly never prove — that any of these districts satisfies the Gingles preconditions. Even a cursory glance at the population data reveals that none of these districts passes muster under Gingles’s first prong. In none of these districts does the citizen voting age population of any cognizable minority group surpass 22%, and in most the percentage is significantly lower. Indeed, even if we were to follow Plaintiffs’ suggestion and combine the Black and Hispanic citizen voting numbers — despite the lack of evidence of cohesion among these groups117 — the districts would still fail the first Gingles precondition by large margins. The evidence convinces us that these districts are influence districts at best, although we note that these districts vary considerably in the opportunity they afford minorities as a group to play any significant role in the electoral process. The population statistics summarized be*486low bring this conclusion into sharper focus.

_Plan 1151C_Plan 1374C_

Black Hispanic B + H Black Hispanic B + H

District CVAP CVAP CVAP CVAP CVAP CVAP

1_15B_3B_191_1&2_31)_22.1

2_140_52_191_19B_82_27.5

_4_1L5_41_152_103_31)_14.2

9_2L3_9/7_3L0_408_108_63.6

10_1L8_200_308_9/7_121_21,8

11_102_1L6_26.8 4.3_2L9_26.2

17 4.0_145_1&5_100_M_19.8

Plaintiffs admit that these seven districts are at most influence districts, but nonetheless urge that the State, after Ashcroft, is forbidden from altering them. Plaintiffs, however, do not argue that any of these districts is protected individually; they focus instead on these districts as a group and suggest that the State, by redrawing these seven influence districts plus the coalition districts in Districts 23 and 24, has gone too far in limiting minority influence statewide. We cannot agree. As we explained above, the State was under no § 2 obligation to create these districts, and we find that the State labors under no corresponding compulsion to preserve these districts. The allegation that the minority voting strength in these districts has been diluted is in truth no more than a claim that these districts have been drawn to add Republican voting strength to overcome the election advantage that the current Democrat incumbents hold.

VI

We next examine the challenge brought by and on behalf of Latino plaintiffs and intervenors to the impact of the legislative plan on Latino voting strength in South and West Texas. At the outset, it is useful to understand that the difficulties Texas presents to a redistricter are nowhere greater than in the Southern and western part of the State. The area can be described as a huge and rough inverted triangle, beginning in El Paso at the far western corner, extending south and east for hundreds of sparsely populated miles on or near the border to the cities of Laredo in Webb County, McAllen in Hidalgo County, and Brownsville in Cameron County, turning north up to the coastal city of Corpus Christi in Nueces County, then turning west to cover many miles that are lightly populated except for the areas to the south of San Antonio in Bexar County and Austin in Travis County. The sheer size of the land, its irregular shape, and the distribution of the bulk of the population in various pockets of the State are the basics that shape the map before the redistricter even begins. This part of the State also contains the greatest concentrations of Hispanic population. In the State as a whole, Latinos account for approximately 32% of the total population, 29% of the voting age population, and 22% of the citizen voting age population. In the South and West Texas regions at issue here, Hispanics represent 58% of the citizen voting age population. The largest numbers of Hispanics are located in the same areas of South and West Texas as are the large pockets of Anglo population: in El Paso County in the southwest corner; in Webb County, Hidalgo County, and *487Cameron County in South Texas along the border and in the Rio Grande Valley; in Nueces County in the east along the coast; and in Central Texas in Bexar County and Travis County.

The plan the Balderas court approved to account for the population changes documented in the 2000 census, Plan 1151C, placed six congressional districts in South and West Texas, each containing the 651,-620 individuals needed for equal population distribution and the mandate of one-man-one-vote. In Plan 1374C, the Legislature drew seven congressional districts in the same area, each containing the necessary 651,620 individuals. In both plans, six districts have a majority Hispanic citizen voting age population. Of the 44 counties included in the six districts in Plan 1151C and the 58 counties included in the seven districts in Plan 1374C, only the seven counties listed above — El Paso, Webb, Hidalgo, Cameron, Nueces, Travis, and Bexar — have populations above 100,000.

This combination of geography and population distribution fixes certain characteristics of the redistricting map for South and West Texas, reflected in both Plan 1151C and Plan 1374C. Both have a district in the far western corner of the State, in El Paso County, which has a population approaching 680,000. Under both Plan 1151C and Plan 1374C, Congressional District 16 in part of El Paso County has a majority Hispanic citizen voting age population and is an effective “safe” Hispanic opportunity district. Plan 1374C leaves Congressional District 16 essentially unchanged, a decision not challenged in this case. But east of this far western population pocket, the counties are so sparsely populated that the district next to 16— Congressional District 23 — must extend far to the east to reach the numbers of people necessary to satisfy equipopulosity.118 A map drawer must travel east almost 800 miles to reach another county that approaches, much less exceeds, 100,-000 souls: Webb County, at the western edge of the southern tip of Texas.

In Plan 1151C, Webb County, with 193,-117 people, is kept entirely in Congressional District 23; in Plan 1374C, Webb County is divided between Congressional District 23 and Congressional District 28, directly to the east. Whether Webb County is divided or not, the districts that begin in the relatively narrow and relatively densely populated southern part of Texas, which includes the Rio Grande Valley, must extend north to gather enough population to satisfy equality among the districts. Both Plan 1151C and Plan 1374C share this characteristic. Plan 1151C has two “strip” districts that begin in South Texas and travel to the north toward the center of the State to gather the requisite number of people.119 Plan 1374C has three “strip” districts that begin in South Texas; each of those districts follows the same north-south path and has the same *488shape as in Plan 1151C, but is narrower and longer to accommodate three districts rather than two. In both Plan 1151C and Plan 1374C, the redistricter traveling north largely avoided the area that is Congressional District 20, which includes San Antonio in Bexar County and is a “safe” Hispanic district.

The map drawer defining the district in the southeastern corner of the State must also begin in the Rio Grande Valley and proceed north to include enough people to satisfy equipopulosity. The map drawer need not travel as far north as in the “strip” districts to find the necessary population for this southeastern coastal district, however, because Nueces County, which contains Corpus Christi, has over 300,000 people.

Both Plan 1151C and Plan 1374C exhibit similar features, resulting from this combination of geography and population distribution. The district that begins just east of El Paso County must be large and must run east from far West Texas, stretching deeply into Central and South Texas. The districts that begin in far South Texas must run north in “strip” fashion into Central Texas. The district that begins in the southern tip of Texas and travels up the coast must also proceed north.

Against this backdrop, we examine the record as to the effects of the legislative plan in South and West Texas on Latino voting strength.

A

In Plan 1151C, the court drew six districts in South and West Texas, each with a majority of the Latino citizen voting age population. As noted, two of those districts, Congressional District 16, consisting primarily of part of the city of El Paso, and Congressional District 20, consisting primarily of part of the city of San Antonio, are not at issue in the change from Plan 1151C to Plan 1374C and are not challenged in this suit. Both were, and are, effective Hispanic opportunity districts. One of the districts in Plan 1151C, Congressional District 23, had a bare majority of Hispanic citizen voting age population and had not performed consistently as a Hispanic opportunity district. Congressional Districts 28, 15, and 27 made up the remainder of the districts in South and West Texas, each with a majority Hispanic citizen voting age population and a majority of the Spanish-surnamed registered voters, and each performing as effective Hispanic opportunity districts. Each of the six districts in South and West Texas under Plan 1151C was reliably Democratic in both congressional and other elections, with the exception of Congressional District 23. That district has since 1992 elected a Hispanic Republican to Congress, Henry Bonilla.

In Plan 1374C, the Legislature sought to apply to South and West Texas its primary partisan goal of increasing the likelihood that Republican candidates would be elected to Congress, while avoiding violations of the Voting Rights Act, the Equal Protection Clause, and the mandate of one-man-one-vote. The record presents undisputed evidence that the Legislature desired to increase the number of Republican votes cast in Congressional District 23 to shore up Bonilla’s base and assist in his reelection. The evidence showed that Bonilla had lost a larger amount of Hispanic support in each successive election. In 2002, Bonilla attracted only 8% of the Latino vote. In order to make Congressional District 23 more Republican, the map drawers extended the district north to take in largely Republican and Anglo areas in the north-central part of the State, including Bandera, Kerr, and Kendall counties. That change added approximately 101,260 people to Congressional District 23. The *489legislative plan moved the district line at the eastern edge to divide the southern border city of Laredo, in Webb County. That change resulted in 99,776 individuals, who were more than 90% Hispanic in voting age population and 86.5% Democratic in voting according to the 2002 statewide election data, being placed in the adjacent district, Congressional District 28. Although Congressional District 23 still had a majority of Hispanics — 55.1%—and a bare majority of Hispanics of voting age— 50.9% — -it no longer had a majority of citizen voting age Hispanics. In the reconfigured Congressional District 23 in Plan 1374C, Hispanics accounted for 46% of the citizens of voting age, and only 44% of the registered voters had Spanish surnames. By contrast, Congressional District 23 in Plan 1151C had a 57.5% Hispanic citizen voting age population and 55.3% Spanish-surnamed registered voters. To avoid retrogression under § 5, the State created another district in South and West Texas, in which Hispanics were a clear majority of the citizen voting age population. The State then had to adjust the population distributions to avoid inequality among the districts.

To accomplish the first goal, Plan 1374C added a third district to the two already long and relatively narrow districts that covered the bottom of the inverted triangle of South Texas and extended north. To maintain the requisite population numbers, each of these three districts had to extend farther north than the two districts had in Plan 1151C. The three districts under Plan 1374C ran from the population pockets near the border north through sparsely-populated areas to reach the pockets of population in the central part of the State, south and east of San Antonio and Austin. The reconfigured districts each added counties in the process.120 The two preexisting districts — Congressional Districts 28 and 15 — maintained a majority Hispanic citizen voting age population and a majority of Spanish-surnamed registered voters in Plan 1374C. The newly created District — Congressional District 25 — also had a majority Hispanic citizen voting age population and a majority of Spanish-sur-named registered voters. The district that runs along the eastern border of South Texas, Congressional District 27, similarly maintained a majority Hispanic citizen voting population and a majority of Spanish-surnamed registered voters. As a result, Plan 1374C had seven congressional districts in South and West Texas, six with a majority of Latino citizen voting age population that are, as explained further below, effective Hispanic opportunity districts, and one that is a Hispanic influence district. Plan 1151C had six congressional districts in South and West Texas with a majority of Latino citizen voting age population, one of which was not an effective Hispanic opportunity district, but was moving in that direction.

Plaintiffs raise a number of challenges to the redrawn districts in South and West Texas. The GI Forum Plaintiffs claim that Plan 1374C is deficient for the same reason its predecessor, Plan 1151C, was deficient: an additional Latino majority citizen voting age population district should be drawn to achieve compliance with § 2 of the Voting Rights Act. Plaintiffs claim that the legislative plan dilutes *490Latino voting strength, in violation of § 2. Finally, Plaintiffs claim that in deciding where to draw the lines, the map drawers were predominately driven by ethnicity, in violation of the Equal Protection Clause. Plaintiffs assert that in drawing the excessively long “bacon-strip” districts, Congressional Districts 15, 25, and 28, the Legislature subordinated the traditional redistricting criteria of compactness and respect for communities of interest and political divisions to the need to include sufficient numbers of Latino voters to create majority-minority districts, in a manner forbidden by Shaw v. Reno (Shaw I)121 and its progeny.

The State responds by arguing that there have been no developments since the Balderas panel issued its opinion to call into question that court’s holding that § 2 did not require an additional Latino majority citizen voting age population district in South and West Texas. The State argues that the Supreme Court’s decision in De Grandy v. Johnson122 strongly supports its position because the number of effective Latino majority citizen voting age population districts in the relevant area is more than roughly proportional to the Latino citizen voting age population in that area and as proportional to the Latino citizen voting age population in the State as the Gingles requirements support. The State argues, and presented evidence to show, that the changes made to the South and West Texas districts in Plan 1374C resulted from the politically motivated decision to make Congressional District 23 more Republican and improve the reelection chances of the Hispanic Republican incumbent, Congressman Henry Bonilla. The State presented evidence that it could not achieve its political end without splitting Laredo and Webb County to remove reliably Democratic voters; that it added a new district, Congressional District 25, which, with Congressional Districts 15, 28, and 27, provides the same number of effective Latino opportunity districts as did Plan 1151C and meets the requirements of De Grandy proportionality; and that ethnicity did not predominate in the numerous decisions involved in the placement of the district lines in Congressional Districts 28, 15, 25, and 27.

The claims that Plaintiffs make as to this part of the State do not raise the questions of coalition districts in which Anglos, Hispanics, and African-Americans are all present, although no minority group is a majority of the relevant population. Rather, Plaintiffs make claims that are more familiar in Voting Rights Act litigation. They claim that in South and West Texas, they are a minority meeting the Gingles requirements; that in one district, they have been “cracked” and submerged into an Anglo majority, such that they cannot elect candidates of their choice; and that as to the rest of the districts in the area, although they are a majority of the citizen voting age population, a functional examination reveals that their ability effectively to elect their candidates of choice has been weakened. The fact that the dilution claim is familiar, however, does not make the analysis simple. As the Supreme Court stated in De Grandy, “Plaintiffs challenging single-member districts may claim, not total submergence, but partial submergence; not the chance for some electoral success in place of none, but the chance for more success in place of some. When the question thus comes down to the reasonableness of drawing a series of district lines in one combination of places rather than another, judgments *491about inequality may become closer calls.”123

B

The GI Forum Plaintiffs allege that they are entitled under § 2 to an additional district in South and West Texas in which Latinos are a majority of the citizen voting age population and can effectively elect candidates of their choice. In both Plan 1151C and Plan 1374C, six districts in South and West Texas have a majority of Latino citizens of voting age. The GI Forum Plaintiffs present a demonstration district, Plan 1385C, that shows an additional Latino citizen voting age majority district in South and West Texas. The record contains variations of proposed dis-tricting plans that meet the same goal. These Plaintiffs, however, have presented no convincing basis to reject the Balderas holding that § 2 did not require an additional district in South and West Texas after the 2000 census.

The GI Forum Plaintiffs argue that since Balderas, additional data has become available distinguishing the citizen voting age population by race and ethnicity, making it easier to establish that an additional Gingles district can be drawn in South and West Texas. Plaintiffs presented demonstration plans to the Balderas panel that they claimed showed the feasibility of drawing an additional Hispanic citizen voting age population majority district in the area. The Balderas panel found that Plaintiffs had failed to prove that § 2 required the creation of an additional Latino citizen voting age majority congressional district in South and West Texas. Plaintiffs appealed that finding; the Supreme Court summarily affirmed. The additional census data Plaintiffs present does not alter the validity of that finding.124

The GI Forum Plaintiffs assert that they have met the Gingles criteria because their demonstration plan, 1385C, creates seven districts with a Hispanic citizen voting age population above 50%. The record, however, does not show that their demonstration plan would satisfy Gingles. Plan 1385C proposes districts that are more unusually shaped than in either Plan 1151C or Plan 1374C.125 The demonstra*492tion plan Plaintiffs present underscores the difficulty of drawing seven, rather than six, Latino opportunity districts that meet the Gingles requirements in the vast geography and irregularly distributed population of South and West Texas.

Even if this court were to assume that the finding in Balderas rejecting the claim that an additional Hispanic citizen voting age population majority district should be drawn in South and West Texas was entitled to no weight, and even if this court were to overlook the Cingles problems reflected in the proffered demonstration plan, the GI Forum Plaintiffs have failed to make the necessary showing under § 2. This court recognizes that Plaintiffs have established racially polarized voting and a political, social, and economic legacy of past discrimination. But any examination of the totality of the circumstances beyond Cingles must include proportionality. Plaintiffs argue that because Latinos represent 22% of the citizen voting age population of the State, seven out of the thirty-two districts in the State should be drawn to produce Latino citizen voting age majorities. The State responds by noting that Latinos comprise 58% of the citizen voting age population in South and West Texas and constitute the majority of the citizen voting age population in six out of the seven districts in that region, exceeding rough proportionality for the area.126 The *493State also argues that even considering all of Texas as the relevant area for measuring proportionality, an examination of the totality of the circumstances does not lead to the conclusion that an additional Latino majority citizen voting age population district must be drawn in South and West Texas.

The Supreme Court has explained that “proportionality,” a word not expressly used in § 2, involves a comparison between (1) the percentage share of legislative districts in which the population of the protected class has a majority and (2) the protected class’s percentage share of the “relevant population.” 127 Proportionality is an important aspect in evaluating “equality of opportunity, [but is] not a guarantee of electoral success for minority-preferred candidates.” 128

The Supreme Court has not resolved what geographic frame of reference should be used to analyze proportionality, whether it is by district, county, region, or state. In DeGrandy, “the plaintiffs ... passed up the opportunity to frame their dilution claim in statewide terms”; the Court applied the Gingles factors and analyzed proportionality as limited to “Hispanics in the Dade County area.”129 In this case, however, Plaintiffs do not argue that the Gingles criteria justify the creation of' any additional Latino citizen voting age majority districts outside South and West Texas. Despite the presence of large numbers of Hispanics elsewhere in the State, and despite the presence of racially polarized voting throughout the State, no plaintiff asserts that the Gingles criteria would permit an additional Latino citizen voting age population majority district anywhere but in South and West Texas. Again, the combination of geography and population distribution quickly explain why. The areas in which Latinos account for 50% or more of the voting age population are confined to El Paso County in far West Texas, surrounded by counties of vast space and little population; Presidio County in West Texas along the border, which has only 7,304 people arid is surrounded by vast' space with little population; and the South Texas area from Maverick County to the Rio Grande Valley. The Latino population in the rest of the State is numerous, but dispersed over large areas.130 Lower courts that have analyzed “proportionality” in the De Grandy sense have been consistent in using the same frame of reference for that factor and for the fac*494tors set forth in Gingles.131 If South and West Texas is the only area in which Gingles is applied and can be met, as Plaintiffs argue, it is also the relevant area for measuring proportionality. Because the Supreme Court has not yet provided precise guidance on the proper standard for assessing proportionality, however, we also examine proportionality on a statewide basis.

Under the legislative plan, 1374C, as under the court-imposed plan it replaces, 1151C, six out of the seven districts in South and West Texas are Latino citizen voting age majority districts. Given the fact that Latinos comprise 58% of the citizen voting age population in South and West Texas, proportionality is satisfied as to that area.

As noted, the Supreme Court has not made clear what geographic unit is the relevant area to measure proportionality. De Grandy did not discuss the role of a district such as Congressional District 29 in the Houston area of Harris County, or Congressional District 23 in West Texas, in which Latinos constitute approximately 47% of the citizen voting age population, less than a majority but large enough to constitute an influence district, in assessing proportionality as part of the totality of the circumstances. Plaintiffs are correct in their calculation that six districts in which Latinos hold a majority of citizen voting age population, out of the thirty-two districts that comprise Texas, do not equate to arithmetic proportionality between the number of Latino majority-minority districts and the Latinos’ percentage of the citizen voting age population in the State. De Grandy emphasizes, however, that the inquiry is not merely mathematical.132 Rather, De Grandy requires an examination of whether the totality of circumstances includes rough proportionality between the number of effective majority-minority districts that can be drawn meeting the Gingles factors and the minority members’ share of the relevant population.133

One of the districts that Plaintiffs would create in their demonstration plan, proposed Congressional District 28 has a Hispanic citizen voting age population of only 50.3%, and five of the seven districts have a Hispanic citizen voting age population that is below 60%. Plaintiffs vigorously criticize the legislative plan, 1374C, in part *495because three of the districts it creates— Congressional Districts 15, 25, and 28— have Hispanic citizen voting age population numbers that are below 60%. Plaintiffs’ own experts and argument reminded this court that because of the lower turnout of Latino voters, a low majority of the Hispanic citizen voting age population does not produce an effective Latino opportunity district.134

Under Plan 1151C, Congressional District 23 was not an effective minority opportunity district, despite the fact that it had a 57.4% Latino citizen voting age majority. Plaintiffs specifically criticize Congressional District 15 in Plan 1374C, which has a Hispanic citizen voting age population of 58.5% — reduced from 69.3% in Plan 1151C — as weakened to the extent that it is better classified as a Latino influence district rather than a Latino opportunity district, although other witnesses contradicted this characterization.135 The legislative plan has one fewer district in which Latinos constitute a majority of the citizen voting age population than the demonstration plan. The legislative plan and the demonstration plan, however, have the same number of districts in which Latinos equal or exceed 55% of the citizen voting age population. The legislative plan has one more district in which the Hispanic citizen voting age population exceeds 60% *496than the demonstration plan.136

The GI Forum Plaintiffs have shown neither that seven districts can be drawn, meeting the threshold Gingles requirements, that have a majority of Hispanic citizen voting age population, nor that all such districts, if they could be drawn, would function effectively as Latino opportunity districts. We reject the claim of the GI Forum Plaintiffs that § 2 demands a seventh Hispanic citizen voting age population majority district in South and West Texas. Part of the analysis of the dilution claim, to which we now turn, will explain why we believe the six Hispanic citizen voting age population majority districts drawn in Plan 1374C are effective Hispanic opportunity districts.

C

It is undisputed that Plan 1374C eliminated Congressional District 23 as a district with a Latino majority citizen voting age population for the political purpose of increasing Republican voters in the district and shoring up the reelection chances of the Republican incumbent. To avoid retrogression, Plan 1374C added a new Latino majority citizen voting age population district to the east and reconfigured the existing districts. Plaintiffs complain that the weakening of Congressional District 23, preventing it from continuing to move toward becoming an effective opportunity district, results in dilution in violation of § 2. Plaintiffs argue that the dilution cannot be addressed by the creation of Congressional District 25 as a Latino opportunity district. Plaintiffs also complain that even if dilution is to be measured on a broader basis than a single district, Plan 1374C is dilutive because the reconfiguration of Congressional Districts 15, 28, and 27 weakens Latino voting strength in these districts.

Congressional District 23 is unquestionably not a Latino opportunity district under Plan 1374C. The map drawers divided Webb County, which is 94% Latino. This change removed reliably Democratic voters in Laredo out of Congressional District 23. Under Plan 1374C, the Hispanic citizen voting age population is 46%, reduced from 57.5% in Plan 1151C; the percentage of Spanish-surnamed registered voters is 44% under Plan 1374C, reduced from 55.3% in Plan 1151C. Even as configured under Plan 1151C, however, Congressional District 23 did not perform as an effective opportunity district. The GI Forum Plaintiffs’ regression analysis of election data showed that Latino candidates of choice were elected in five out of eight racially contested elections from 1994 to 2002.137 The record showed, however, that Latino voters in Congressional District 23 were *497voting against the Republican candidate Bonilla by larger margins in each successive election. Latino voting strength in Congressional District 23 is, unquestionably, weakened under Plan 1374C. The GI Forum regression analysis of election data for Congressional District 23 under Plan 1374C showed that the Hispanic candidates of choice won only one out of eight racially contested elections from 1994 to 2002.138 Dr. Gaddie’s regression data showed that, in Congressional District 23 under Plan 1374C, Hispanic candidates of choice won seven out of seven statewide primaries, but none of the six Hispanic candidates of choice in general elections carried the district.139

Plaintiffs claim that the map drawers impermissibly reduced Hispanic voting strength in Congressional District 23 to pursue their political end. Plaintiffs urge that the political end could have been achieved by leaving Webb County whole and redrawing Congressional District 23 to put Bonilla’s home in Congressional District 11, where he could have run in an open seat. The State points out that this approach would not have served either the political goal of incumbency protection or of increasing Republican districts, since Congressional District 11 already reliably elects Republicans in congressional and other elections. The change to Congressional District 23 served the dual goal of increasing Republican seats in general and protecting Bonilla’s incumbency in particular, with the additional political nuance that Bonilla would be reelected in a district that had a majority of Latino voting age population — although clearly not a majority of citizen voting age population and certainly not an effective voting majority.

Plaintiffs argue that the admittedly political decision to draw the line through the city of Laredo, placing the largely Democratic, Hispanic voters in Congressional District 28 to the east and “stranding” 359,000 Hispanics in Congressional District 23, where they will be submerged in a Republican, Anglo majority district, violates § 2. De Grandy, however, rejected the claim of separating heavily Hispanic neighborhoods, fragmenting them in certain districts and packing them into others, as a basis for liability, even if the Gingles factors are met, because the totality of the circumstances included a showing that Hispanic voters had effective voting majorities in the relevant area in rough proportion to the their voting age population.140

Typically, vote dilution claims address redistricting schemes that take a minority group whose members have the potential to comprise a numerical majority in a geographically compact district and disperse the group across two or more districts, thereby precluding the minority members from constituting an effective majority in either one.141 Plan 1374C, however, does not produce such an effect in South and West Texas. Instead, Plan 1374C changes the district boundary in a way that divides Democratic Hispanic voters in Webb County, keeping half in Congressional District 23, in which Republican voting strength is increased and Hispanic voting *498strength is weakened, and placing half in Congressional District 28, in which Latinos have a clear majority of the citizen voting age population, even after a new district, 25, is added as a majority Latino citizen voting age population district. The Legislature also reconfigured Congressional Districts 15, 28, and 27 in ways that both achieved equal population distribution and preserved the majority Latino citizen age voting populations in each of those districts.

Plaintiffs rely on Shaw v. Hunt (Shaw II) 142 to argue that the State cannot “offset” the reduction of Latino voting strength in Congressional District 23 by creating a new Latino majority citizen voting age district, Congressional District 25, because to do so would “swap” the voting rights of citizens in Congressional District 23 for the rights of those in new Congressional District 25. In Shaw II, the Court held that the creation of a remedial district in one area of a state, where polarized voting existed but where a compact majority-minority district could be not drawn, could not compensate for the failure to place a remedial district in another area where it was possible to draw a compact majority-minority district and where polarized voting also existed.143 The Court held that a bizarrely shaped district “somewhere else in the State” does not remedy the “vote-dilution injuries suffered” by minority voters residing where the remedial district could have been drawn.144 The Court rejected a district drawn in the center of the state to remedy the dilution injury of voters on the southeastern side of the state, not because there were differences in polarization, but because a remedial Gingles district could not be drawn in the part of the state that the map drawers selected.145 In this case, by contrast, in South and West Texas, six Gingles Latino citizen voting age population majority districts could be and were drawn.146 The Legislature did not place the lines of those districts in parts of the State where the Gingles requirements were not satisfied; the problem present in Shaw II is not involved in this case.

Analysis of the Gingles factors shows that the Latino population is sufficiently numerous and distributed as to support the creation of a number of effective districts in South and West Texas with a majority of Latino citizen voting age population. The Gingles districts in South and West Texas could be drawn in different ways, within the constraints of geography and population distribution. Both Plan 1151C and Plan 1374C have six districts with a majority of Latino citizens of voting age. The choices available to the State included the lines of Plan 1374C, which created a new remedial district that had not been present under the previous plan, or the lines of Plan 1151C, keeping Congressional District 23 as a bare majority Hispanic citizen voting age population district, but not drawing another district with a stronger Latino citizen voting age popu*499lation majority. Most of the Plaintiffs argue that the State should be compelled to return to the lines of Plan 1151C for South and West Texas. Plaintiffs do not argue that the State should both retain the lines of Congressional District 23 under Plan 1151C and add a new majority Hispanic citizen voting age population district along the lines of Congressional District 25. Indeed, there is neither sufficiently dense and compact population in general nor Hispanic population in particular to support such a configuration. But to say that the State could have retained the lines of Congressional District 23 drawn under Plan 1151C and not created a third district based in the Rio Grande Valley with a majority of Hispanic citizen voting age population is different from saying that the State was obligated to make that choice. Sha/iu II does not preclude the State from choosing where and how to draw majority-minority districts in areas where Gingles is satisfied. The states retain broad remedial power to choose where and how to draw remedial districts.147 As the Court noted in Shaw II, even if a violation of § 2 is shown, it does not confer on individual plaintiffs “the right to be placed in a majority-minority district” because “[sjtates retain broad discretion in drawing districts to comply with the mandate of § 2.”148

To examine whether Plan 1374C imper-missibly dilutes the votes of Latinos in South and West Texas, this court must conduct a “comprehensive ... canvassing of relevant facts” making up the totality of the circumstances, including proportionality, to determine whether the districts drawn under Plan 1374C in South and West Texas are effective Latino opportunity districts.149 Congressional District 23 is clearly not a minority opportunity district; Congressional Districts 16 and 20 do clearly provide effective Latino citizen voting age population majorities. The disputes center on whether Congressional Districts 15, 25, 28, and 27 as drawn in Plan 1374C will be effective Latino opportunity districts.

Some of the relevant data from the voluminous record is as follows:

Congressional District % HCVAP % SSR (2002)

1151 15 69.3 67.0

1374 15 58.5 56.7

1151 25 18.6 15.2

1374 25 55.0 55.6

1151 28 61.4 59.6

1374 28 56.2 54.3

1151 27 63.5 61.6

1374 27 60.4 58.0

The parties also presented detailed regression analyses of election data, performed by different experts.150 The regression analyses confirmed that in Congressional Districts 15, 25, 28, and 27 under Plan 1374C, the Latino candidate *500of choice won in nearly every primary and runoff examined. The GI Forum’s regression analysis of election data reveals that under Plan 1374C, Latinos elected their candidate of choice in eight out of eight racially contested elections from 1994 to 2002 in every one of the six Latino majority citizen voting age population districts, including Congressional District 15.151 Dr. Gaddie’s report similarly concluded that the Latino candidate of choice won in seven of seven primaries and five of six general elections in Congressional District 15.152 In Congressional District 25, Dr. Gaddie concluded that Hispanic candidates of choice carried seven of seven statewide primaries and six of six statewide general elections involving candidates of different ethnicity or race, and that in the 2002 contested statewide Democratic primaries and statewide general elections, Hispanic candidates of choice prevailed in five of six primaries and all fifteen general elections.153 In Congressional District 27, Dr. Gaddie concluded that Hispanic candidates of choice won seven of seven statewide primaries and five of six statewide general elections involving candidates of different ethnicity or race.154 In Congressional District 28, Dr. Gaddie concluded that Hispanic candidates of choice prevailed in all seven statewide primaries and all six statewide general elections involving candidates of different ethnicity or race.155

Although Dr. Lichtman expressed concern when testifying for Plaintiffs that Congressional District 15 and Congressional District 25 might be marginal, or moving toward a “toss-up,” he acknowledged on cross-examination that all four of the districts at issue, 15, 25, 27, and 28, are Hispanic opportunity districts, defined as a district in which Hispanic voters have an effective opportunity to control outcomes in both primary and general elections.156 Dr. Lichtman expressed concern that in Congressional District 25 under Plan 1374C, while Hispanic voters would control the primaries, Congressman Lloyd Dog-gett, the incumbent in Congressional District 10 under Plan 1151C and an Anglo Democrat, might run and win the general election without a majority of Hispanic *501votes.157 Dr. Lichtman nonetheless agreed that Congressman Doggett has had “overwhelming Latino support” in his current congressional district.158

Several witnesses expressed particular concern over Congressional District 15, which under Plan 1374C has a Latino citizen voting age population of 58.5% and a Spanish-surnamed registered voter population of 56.7%. Although these numbers are lower than the levels that had been present in Congressional District 15 under Plan 1151C, the regression analyses showed that the Hispanic-preferred candidate won every primary and runoff election studied, and lost in only one general election out of six.159 Both Dr. Lichtman and Dr. Engstrom testified that Congressional District 15 was a minority opportunity district, although not as strong as it had been under Plan 1374C.160 In contrast, Dr. Polinard testified that given this reduction in strength, Congressional District 15 was no longer an effective or performing opportunity district, merely an influence district.161

In Plan 1151C, Congressional District 27 included all of Cameron County along the border, but only part of Nueces County to the north. In Plan 1374C, Congressional District 27 includes all of Nueces County but divides Cameron County. The result did not significantly change the relevant data. In Plan 1374C, Congressional District 27 has a Hispanic citizen voting age population of 60.4% and a Spanish-sur-named registered voter population of 58%, reduced from a Hispanic citizen voting age population of 63.6% and a Spanish-sur-named registered voter population of 61.6% under Plan 1151C.162 As noted, the regression analyses showed that Hispanic candidates of choice won seven of seven statewide primaries and five of six statewide general elections involving candidates of different ethnicity.163 Dr. Lichtman testified that Congressional District 27 under 1374C is a minority opportunity district.164 Dr. Engstrom testified that Congressional District 27 in Plan 1374C is a district in which Hispanics have the opportunity to elect their preferred candidate.165

Congressional District 28 under Plan 1374C includes part of Laredo in South Texas and extends north to the relatively heavily populated areas of Central Texas. The demographic figures show that the district maintains a decisive Hispanic citizen voting age population majority, although reduced from the levels present under Plan 1151C. In Congressional District 28 under Plan 1374C, the Hispanic citizen voting age population is 56.2% and the Spanish-surnamed registered voter population is 54.3%, slightly less than the 61.5% of Hispanic citizen voting age population and 59.6% Spanish-surnamed registered voter population under Congressional District 28 in Plan 1151C.166

*502At trial, witnesses familiar with the areas covered by Congressional Districts 15, 25, 27, and 28 and with the difficulties faced by candidates supported by Latino voters expressed concern as to whether these districts would be effective as Latino opportunity districts.167 The witnesses testified that the size of the districts; the fact that they include several media markets; and the fact that they combine communities along the border with communities in Central Texas, with diverse needs and interests, could make it more difficult for thinly financed Latino-preferred candidates to achieve electoral success and to provide adequate and responsive representation once elected.168 These concerns bear on the extent to which the new districts are functionally effective Latino opportunity districts, important to understanding whether dilution results from Plan 1874C.169

The “bacon-strip” districts — Congressional Districts 15, 25, and 28 — are unquestionably long. Congressional District 25 is 300 miles long and 77% of its total population comes from either end of the district; 39% from Travis County in Central Texas and 38% from Hidalgo County at the southern end. Congressional District 15 is also 300 miles long, from Cameron County in South Texas to Bastrop County in Central Texas, and its population is also concentrated at either end. Congressional District 28 runs nearly 300 miles from Hays County south to Zapata County on the border with Mexico. Districts that began in the same area in Plan 1151C were also long and relatively narrow, extending from the border of South Texas to Central Texas, although not as far. The record does not disclose the number of media markets covered in the districts under Plan 1151C, but it is clear that several of the areas joined in districts under Plan 1374C were also joined together in districts under Plan 1151C. It is also clear that adding a third district based in the Rio Grande Valley made each of the three districts extend north and include residents of both the Rio Grande Valley and Central Texas. While Latino residents in both areas generally have lower socio-economic indicators than Anglos, the evidence at trial showed that the needs and interests of Latino communities on the South Texas border are different from the needs and interests of Latino communities in Central Texas. The issue is whether these features mean that the newly-configured districts dilute the voting strength of Latinos.

At trial, the court was greatly assisted by the testimony of elected officials from the districts at issue. These witnesses, including Congressman Hinojosa, representing Congressional District 15; State Representative Aaron Peña, who resides in Congressional District 15 and represents a district located in Congressional District 25; State Representative Jim Solis from Cameron County in Congressional District 27; and County Judge Ramon Garcia and County Commissioner Ricardo *503Godinez from Hidalgo County, were most concerned that the districts would no longer be dominated by the border cities in the Rio Grande Valley. Instead, the districts would also contain a number of constituents from northern, less impoverished communities. The witnesses testified that the size and diversity of the newly-configured districts could make it more difficult for the constituents in the Rio Grande Valley to control election outcomes. For example, witnesses testified that Congressional District 27 under Plan 1374C includes more affluent and higher-turnout voting areas in Nueces County, in contrast to Congressional District 27 in Plan 1151C, which included only half of Nueces County, and expressed concern that Latinos from the border might have difficulty in maintaining control over election outcomes because of low turnout.170 The district, however, has a Latino citizen voting age majority population of 60.4%, and 58% of the registered voters have Spanish surnames. Witnesses consistently testified that Congressional District 27 under Plan 1374C is a district where Hispanics have an effective opportunity to elect their preferred candidate, and the regression data supports this conclusion.171 As noted, witnesses testified that Congressional District 25, while a more effective Latino opportunity district than Congressional District 23 had been in Plan 1151C, might elect Lloyd Doggett, a well-known and well-financed Anglo Democrat from Central Texas, to Congress, but acknowledged that many Hispanics support this Anglo Democrat.172 The legal test of an effective opportunity district is not whether a Hispanic candidate is elected each time, but whether a Hispanic-preferred candidate has an effective opportunity to be elected.173 Witnesses testified that Congressional Districts 15 and 25 would span colonias in Hidalgo County and suburban areas in Central Texas, but the witnesses testified, and the regression data show, that both districts are effective Latino opportunity districts, with the Hispanic-preferred candidate winning every primary and general election examined in District 25 and, in District 15, winning every primary and runoff election studied and losing in only one general election out of six.174 Witnesses testified that Congressional District 28 had been, and would continue to be, an effective Hispanic opportunity district, confirmed by the regression data showing that Hispanic candidates of choice won in every primary and general election examined.175

A close examination of the voluminous record and the testimony at trial leads to a finding that Plan 1374C does not impermissibly dilute the voting strength of Latinos in South and West Texas in violation of § 2. The newly-configured Districts 15, 25, 27, and 28 cover more territory and travel farther north than did the corresponding districts in Plan 1151C. The districts combine more voters from the cen*504tral part of the State with voters from the border cities than was the case in Plan 1151C. The population data, regression analyses, and the testimony of both expert witnesses and witnesses knowledgeable about how politics actually works in the area lead to the finding that in Congressional Districts 25 and 28, Latino voters will likely control every primary and general election outcome; in Congressional Districts 15 and 27, Latino voters will likely control every primary outcome and almost every general election outcome; and in Congressional District 23, Latino voters will likely control every primary outcome but not the general elections. The fact that Hispanic-preferred candidates will have to campaign for votes not only from Latinos in the Rio Grande Valley but also from Latinos in Central Texas in Congressional Districts 25, 28, and 15 does not mean that Latinos have suffered an impermissible dilution of voting strength in those districts. The fact that Latinos from the Rio Grande Valley in Congressional District 27 are combined with voters from Nueces County, while maintaining a decisive citizen voting age population and Spanish-surnamed registered voter majority, as well as consistent electoral control, does not mean that Latinos have suffered an impermissible dilution of voting strength in that district.176

As the Court stated in Gingles, “[T]he relative lack of minority electoral success under a challenged plan, when compared with the success that would be predicted under the measure of undiluted minority voting strength the court is employing, can constitute powerful evidence of vote dilution.” 177 Whether compared to Plan 1151C or to the GI Forum Plaintiffs’ demonstration plans, Plaintiffs have not shown an impermissible reduction in effective opportunities for Latino electoral control or in opportunities for Latino participation in the political process. Section 2 guarantees minority voters an effective opportunity to exercise electoral control, not overwhelming electoral majorities or guarantees of success. The totality of facts and circumstances, including those pointing to proportionality, as well as past and predicted election outcomes and evidence as to the likely functioning of the newly-configured districts, does not show a violation of § 2 in South and West Texas under Plan 1374C.

Plaintiffs argue that taking a broad view of the státe-wide effect of Plan 1374C, dilution results because of the loss of the “influence districts” in other parts of the State in which Latino voters played a role, specifically, Congressional District 24 in Dallas and Tarrant County and Congressional District 10 in Travis County. The State responds in part by pointing to the strengthening of Congressional District 29 in Harris County as a Latino influence district. But Plaintiffs have not urged that additional Gingles Latino majority-minority districts can be placed in those areas; they have not identified a wrong in *505those areas under § 2 that would justify requiring the State to effect a remedy. As the Court explained in Growe, “Unless [the Gingles prerequisites] are established, there neither has been a wrong nor can be a remedy.”178 Plaintiffs have not met the burden of showing that § 2 forbids the State from making the political choices that primarily shaped the changes reflected in Plan 1374C because the effects of those changes impermissibly diluted the opportunity of Latinos to participate in the political process.

In De Grandy, the Court emphasized that one reason it would not embrace “bottom line” proportionality as a “safe harbor” was that it could make “the most blatant racial gerrymandering in half of a county’s single-member Districts ... irrelevant under § 2 if offset by political gerrymandering in the other half.”179 Plaintiffs argue that the reduction of Congressional District 23 from an emerging effective Hispanic opportunity district for political purposes was political gerrymandering and that the creation of six majority Hispanic citizen voting age population districts in the remainder of West and South Texas resulted from racial gerrymandering. We turn to an examination of this claim.

D

Plaintiffs contend that in Plan 1374C, the map drawers used ethnicity to an unlawful degree and drew districts that reflect a derogation of the traditional dis-tricting values of compactness, respect for political lines, and communities of interest. The Supreme Court has phrased the test for a plaintiff bringing such a Shaw claim in different ways. In 1995, in Miller v. Johnson,180 the Court struck down a Georgia districting plan on the ground that race had been “the predominant factor motivating the Legislature’s decision to place a significant number of voters within or without a particular district.”181 That conclusion will generally follow, the Court explained, where “the legislature subordinated traditional race-neutral districting principles ... to racial considerations.”182 In her concurrence in Bush v. Vera,183 Justice O’Connor reaffirmed that test, but explained it in slightly different terms: “[S]o long as they do not subordinate traditional districting criteria to the use of race for its own sake or as a proxy, States may intentionally create majority-minority districts, and may otherwise take race into consideration, without coming under strict scrutiny.... Only if traditional districting criteria are neglected and that neglect is predominantly due to the misuse of race” is the district presumptively unconstitutional.184 States are not required to ignore race; indeed, the Supreme Court has acknowledged that states will always be aware of race when they draw district lines.185 The factor of race or ethnicity may be considered in the process as long as it does not predominate over traditional *506race-neutral districting principles.186 The fact that race is given consideration in the process and the fact that majority-minority districts are intentionally created do not suffice to trigger strict scrutiny.187

Plaintiffs bear the burden of proving that race was the predominate factor in the Legislature’s districting decisions.188 That burden is significant.189 As the Court stated in Miller, “To invoke strict scrutiny, a plaintiff must show that the State has relied on race in substantial disregard of customary and traditional dis-tricting practices.... [Application of the Court’s standard helps achieve Shaw’s basic objective of making extreme instances of gerrymandering subject to meaningful judicial review.”190

Courts analyzing Shaw claims frequently examine: (a) district shape and demographics; (b) statements made by legislators and their staff; and (c) the nature of the data used to determine whether race played an excessive and unjustifiable role in the redistricting process.191 The parties presented evidence of two widely accepted measures of compactness, “smallest circle” and “perimeter to area” scores.192 The compactness scores of the challenged districts in Plan 1374C are as follows:

_CD 23 CD 28 CD 25 CD 15 CD 27

Smallest Circle_3/3_5b_8.5 ;_6b_3.1

Perimeter to Area_51_5/7_9b_1L6_51.

The scores for the corresponding districts in Plan 1151C are:

_CD 23 CD 28 CD 25 CD 15 CD 27

Smallest Circle_42_3/7_5b_3.1

Perimeter to Area_61_51_8b_41

The general measures of compactness do not lead to a conclusion that the districts in Plan 1374C are bizarre in a Shaw sense. Dr. John Alford, an expert witness for the Jackson Plaintiffs, testified that he did not find the perimeter to area scores for Congressional Districts 15, 25, and 28 in Plan 1374C to be troublesome on their own.193 Neither the smallest circle nor *507perimeter to area scores of Plan 1374C approach those of districts so bizarrely and irregularly drawn that courts interpret their creation as “an effort to segregate the races for purposes of voting, without regard for traditional districting principles.” 194

Compactness scores cannot be considered in isolation; the issue is whether a district's shape, as measured by compactness scores, provides evidence of a constitutional violation when considered in relation to the geography and population distribution in the relevant part of the State.195 Texas has vast geographical areas with widely dispersed population; except for areas around major cities, the State is a challenge for any redistricter who cherishes compactness as a value.196 For example, Congressional District 23 in both Plans 1151C and 1374C extends approximately 800 miles along the border. In Plan 1151C, Congressional District 17 in the north-central part of the State includes 36 counties. Under Plan 1151C, Congressional District 13 in North Texas is larger than a number of states, spanning 40,000 square miles and 43 counties, but many with fewer than three thousand people.197

Texas geography and population dispersion limit the availability of district compactness in the southern and western regions of the state. Under Plan 1374C, the population densities in Congressional Districts 28, 25, and 15, both Anglo and Hispanic, are highest in the Valley and in Central Texas, separated by relatively sparsely populated areas. The high-density population pockets necessary to achieve the one-man-one-vote requirement are situated at either end of the elongated “bacon-strip” shaped districts. As a result, the boundaries of such “strip” districts in Plan 1374C must reach into Central Texas to obtain the requisite number of people, whether Anglo or Latino. For example, in Congressional District 15, the “dumbbell” district, 90% of the population is in the northern and southern tips. Congressional District 15 was already a “strip” district in Plan 1151C; in Plan 1374C, legislators elongated it to add people because part of the lower half was used to create new Congressional District 25. The smallest circle measure of compactness for the southern and western districts in Plan 1374C, examined in relation to the geography and population, reflect the sheer size and population distribution of the area, rather than a calculated stretch to find voters of a particular ethnic makeup.198

The perimeter to area scores for Plan 1374C similarly do not reflect the predominance of ethnicity in drawing the district lines. Perimeter to area calculations are useful for determining whether the map drawers used convoluted lines to bring *508members of one racial group into a district while excluding members of another racial group. An examination of Congressional Districts 25, 28, 15, and 27, in relation to the population distribution within and without the district lines, does not reveal lines precisely drawn to include Hispanics and exclude Anglo voters, to ensure Hispanic citizen voting age majority districts, but rather lines drawn to include enough voters to meet the one-man-one-vote constitutional requirement. An examination of the record, including the maps that show the relationship of Hispanic voting age population to district lines in Congressional Districts 15, 25, and 28, does not show that the districts were extended north in a determined search for Hispanics, in contrast to the way in which the districts in Miller, Bush, and Shaw twisted or spanned diverse areas to reach every available minority of voting age population.199 Neither the districts’ shape nor their shape in relation to ethnic demographics and population densities provides circumstantial evidence of forbidden racial gerrymandering.

Plaintiffs criticized the Legislature for “cracking” Webb County in a way that removed a large number of Latino voters from Congressional District 23 and placed them in Congressional District 28. The State presented undisputed evidence that the Legislature changed the lines of Congressional District 23 to meet the political purpose of making the district more Republican and protecting the incumbent, Congressman Bonilla. Plaintiffs agree that the primary purpose of this change was political and concede that there is a strong correlation between Latino and Democratic voters. Dr. Lichtman stated in his report that the change in the Webb County district boundary affected an area of voters who were more than 90% Hispanic in voting age population and 86.5% Democratic in voting, according to the 2002 statewide elections.200 Plaintiffs nonetheless fault the Legislature for drawing a line that separated Hispanic voters, recognizing that they were also reliably Democratic voters, to achieve the political purpose.201

*509The Supreme Court has recognized that a legislature “must have discretion to exercise the political judgment necessary to balance competing interests,” and that courts must “exercise extraordinary caution in adjudicating claims that a State has drawn district lines on the basis of race.” 202 In this ease, the State has articulated a political reason for the districting decision, protecting a Republican incumbent. The Supreme Court has recognized that incumbency protection can explain a state’s decision to depart from other traditional districting principles as well as, or even better than, race.203 In Theriot v. Parish of Jejferson,204 the Fifth Circuit rejected a racial gerrymander claim where “the inclusion or exclusion of communities was inexorably tied to issues of incumbency.” 205 The decision to move the southeastern edge of Congressional District 23 to the west affected voters whose ethnicity and political partisanship voting achieved strong correlation. Under such circumstances, caution requires this court to defer to the Legislature’s discretion.206 As the Supreme Court has stated, “a jurisdiction may engage in constitutional political gerrymandering, even if it so happens that the most loyal Democrats happen to be Black Democrats and even if the state were conscious of that fact.” 207 The geography and population distribution, in relation to the lines drawn in Congressional Districts 15, 23, 25, 27, and 28, do not provide support for Plaintiffs’ claim of impermissible racial gerrymandering.

Plaintiffs’ primary evidence of a deviation from traditional districting principles is that Congressional Districts 23, 25, 28, 15, and 27 do not observe political lines or respect communities of interest. However, credible testimony from the State’s witnesses demonstrated that factors at the heart of traditional districting criteria, including political goals, predominately influenced the numerous decisions embodied in the location of each district line.

In response to Plaintiffs’ criticisms, the State provided credible race-neutral explanations for Plan 1374C’s county cuts, city divisions, and linking of border and Central Texas communities. The legislative motivation for the division of Webb County between Congressional District 23 and Congressional District 28 in Plan 1374C was political. The State provided evidence *510that in dividing Webb County, the map drawers in part used the interstate highway as a district boundary, deviating where necessary to achieve population balance.208 Plaintiffs faulted the inclusion of part of Comal County in Central Texas in Congressional District 28, arguing that legislators drew the line between Congressional District 21 and Congressional District 28 to place the heavily Hispanic part of Comal County in Congressional District 28. The State provided evidence of a political reason for this county split: to make Congressional District 21 more safe for its Republican incumbent, who was concerned about the effect of including Democratic-leaning voters in his district because of changes to what had been Congressional District 10 in Plan 1151C. Map drawers placed reliably Democratic voters from Co-mal County into Congressional District 28 to make up for the inclusion in Congressional District 21 of Democratic voters from Travis County resulting from the changes to former Congressional District 10. The State also presented evidence that the part of Comal County placed in Congressional District 28 corresponded to a local election boundary.209

Plaintiffs also faulted the inclusion of a heavily Hispanic part of Hays County, also in Central Texas, in Congressional District 28. The State presented evidence that this decision resulted from the Legislature’s desire to keep Texas State University, located in Hays County, out of Congressional District 21, which contains the University of Texas at Austin.210

Plaintiffs criticized the split that placed part of Cameron County into Congressional District 15. The State responded with credible evidence that Congressional District 15 extends into Cameron County in order to keep the city of Harlingen together in a single district. Plaintiffs castigated the division of Hidalgo County between Congressional District 15 and Congressional District 25. The State introduced testimony that Congressman Hinojosa asked that the cities of Mission and Edinburg be kept whole, which could be accomplished only by splitting Hidalgo County.211 Plaintiffs criticized the Legislature for extending Congressional District 27 to include all of Nueces County and part of San Patricio County; under Plan 1151C, only half of Nueces County was included in Congressional District 27. The State introduced evidence that a state senator on the Redistricting Commission, Senator Luna, wanted to keep Nueces County and San Patricio County undivided in a single district. Although that was not feasible because of the need to achieve population balance, the Legislature was able to keep the port of Corpus Christi, located in Nueces County, in a single district with the port communities of San Patricio County.212 The Legislature’s primary map drawer, Bob Davis, and the sponsor of the redistricting bill in the State House of Representatives, Representative Phil King, provided credible testimony that the numerous decisions embodied in the location of each district line combined the broad political goal of increasing Republican seats with local political decisions that are the most traditional of districting criter*511ia.213

The Supreme Court has focused on the specific features of a district that cause it to depart from compactness in assessing whether those features are justifiable on traditional districting grounds. In Bush, as here, the State attempted to justify its decisions on the traditional ground of partisan politics.214 In that case, the Court found it significant that “the maps reveal that political considerations were subordinated to racial classification in the drawing of many of the most extreme and bizarre district lines.”215 The Bush Court found as clear evidence of the subordination of politics to race the fact that the State was willing, in the course of extending a noncompact tentacle engulfing a pocket of African-American voters, to include a large number of Republican voter tabulation districts in a congressional district ostensibly designed to maximize Democratic power.216 In the present case, by contrast, the evidence did not reveal such internal inconsistencies. Instead, the Legislature’s line-drawing decisions were primarily driven by the political goal of increasing Republican strength in Congressional District 23; the related political goal of ensuring that newly created Congressional District 25, as well as'reconfigured Congressional Districts 15, 28, and 27, did not weaken Republican strength in adjacent Congressional District 21; a number of localized political considerations; and the overall need for a balanced population distribution.217 Although the Legislature clearly intended to create a majority Latino citizen voting age population district in Congressional District 25 and maintain Congressional Districts 15, 28, and 27 as majority Latino citizen voting age population districts, the evidence does not show that it subordinated all other traditional districting criteria to a scheme in which ethnicity predominated.218 To the contrary, the evidence shows that many of the lines were drawn for such reasons as balancing population, keeping certain cities or areas intact in a district, and satisfying requests from state or federal legislators to keep certain areas together, or place universities in different districts.219 The evidence also shows that the lines did not make twists, turns, or jumps that can be explained only as efforts to include Hispariics or exclude Ang-los, or vice-versa. Plan 1374C does not reveal a subordination of traditional dis-tricting principles to ethnic considerations.

Plaintiffs urge that a derogation of traditional districting principles in Plan 1374C *512is evidenced by the fact that it links disparate communities in single districts, compromising the ability of the most impoverished and needy areas to obtain necessary representation. Plaintiffs complain that the new plan places border towns with large Hispanic populations in the same districts as Central Texas areas with large Hispanic populations, with ethnicity as their only common characteristic.

According to the Supreme Court, manifestations of a community of interest include, “for example, shared broadcast and print media, public transport infrastructure, and institutions such as schools and churches.” 220 The district challenged in Shaw I was clearly not driven by communities of interest; the Court described the district as one that “winds in snakelike fashion through tobacco country, financial centers, and manufacturing areas” in an attempt to gather up “enclaves of black neighborhoods” and “even towns are divided.”221 And in Miller, because the challenged district reached out to include African-American pockets in several entirely separate urban communities, linked by narrow corridors to a sparsely populated and wholly rural core, the “social, political and economic makeup of the Eleventh District [told] a tale of disparity, not community.” 222

The “bacon-strip” districts of Plan 1374C, Congressional Districts 15, 25, and 28, include disparate communities of interest. Plaintiffs presented evidence of differences in socio-economic status, education, employment, health, and other characteristics between Hispanics who live near Texas’s southern border and those who reside in Central Texas. However, like the deviations from compactness, these deviations from the traditional dis-tricting principle of protecting communities of interest were necessary to meet the constitutional requirement of equal district populations in an area in which population is clustered at the southern tip and the center of the State. Plaintiffs’ evidence has not demonstrated that the linking of disparate border and Central Texas Hispanic communities was caused by the factor of ethnicity, rather than the factors of geography and population distribution and the need to achieve equipopulosity.223 Given the presumption in favor of legislative integrity,224 we cannot say that the combination of communities of interest here provides circumstantial evidence of the predominance of ethnicity.

The record does not present evidence of statements by legislators or staff supporting the claim that ethnicity predominated in the redistricting process. To the contrary, the emails, statements, and other communications from those involved in the process reveal that politics predominated.225 Similarly, the data the Legislature used in the districting process does not support a claim of unwarranted reliance on ethnicity to make the line-drawing decisions. In Bush, the Court condemned the use of census blocks as the fundamental unit for a redistricting plan because the *513use of such units allows a state to focus too heavily on race in drawing district lines.226 In the present ease, by contrast, the State presented undisputed testimony that the map drawers examined race at the block level in the South and West Texas districts on only a few occasions in order to avoid splitting minority communities.227

The Supreme Court has made it clear that a legislature may give consideration to race and ethnicity in the districting process, as long as those factors do not predominate and districting decisions are explainable on grounds other than race.228 The districting decisions involved in Plan 1374C are best explained by Texas’s geography and population distribution and its Legislature’s predominately political intent. The record in this case does not show that ethnicity predominated or that the South and West Texas district boundaries in Plan 1374C cannot be explained except by ethnic considerations. Plaintiffs have not met their significant burden of demonstrating racial gerrymandering.229

VII

Congressional District 18 is a historically significant African-American majority district in Houston, Harris County, Texas. Its first representative was Congresswoman Barbara Jordan. Jordan was elected at the creation of the district in 1972 as one of the two first African-Americans elected to Congress from the South since Reconstruction and the first African-American elected to Congress from Texas. Congressional District 18 is currently represented by Congresswoman Sheila Jackson Lee. Congressional District 30 is also a historically significant district, the first African-American majority district in Dallas, Dallas County, Texas. Congresswoman Eddie Bernice Johnson represents Congressional District 30. Both Congresswomen Jackson Lee and Johnson provided helpful testimony in this case, adding their voices to those emphasizing that racial polarization and appeals to racial prejudice persist in Texas politics and political campaigns.

Plaintiffs Jackson Lee and Johnson, supported by the NAACP and other plaintiffs, criticize Plan 1374C as diluting the African-American voting strength in Congressional Districts 18 and 30. As to Congressional District 18, Plaintiffs claim that in order to create Congressional District 9 as an additional African-American opportunity district in Houston, Plan 1374C moved high turnout African-American voters out of Congressional District 18 and placed them in Congressional District 9. In Congressional District 18, according to Plaintiffs, those voters were replaced with what the NAACP characterized as “transient and Hispanic voters who are noted for their low participation in elections.”230 Plaintiffs criticize Congressional District 30 in Plan 1374C on a similar ground. Plaintiffs assert that the plan removes from Congressional District 30 a predominately Anglo area from Irving, Texas and replaces it with an area of largely Hispanic voters from Congressional District 24 in Plan 1151C. Congresswoman Johnson criticized the effects on Congressional District 30 as removing an economically strong area from the district and combin*514ing areas of African-American and Hispanic voters, setting up a competition between them that could dilute African-American voting strength. Recognizing that both districts retain a strong African-American majority in Plan 1374C, Plaintiffs rely- on projections of the relative growth of the Latino population in the years 2005 and 2010 to support the argument that dilution is threatened, if not present.

The evidence that Plaintiffs present does not support their claim of dilution, present or threatened. Congressional Districts 18 and 30 are effective African-American opportunity districts under Plan 1374C.231 Plaintiffs’ own population and voting data support this conclusion. In Plan 1374C, Congressional District 18 has an African-American voting age population of 40.3% and an African-American citizen voting age population of 48.3%. This is only slightly below the figures that were present in Plan 1151C, which had a Black voting age population of 42.1% and a Black citizen - voting age population of 49.8%. Plaintiffs argue that Plan 1374C increases the number of Hispanics and reduces the number of high-turnout African-American voters. The Hispanic citizen voting age population in Congressional District 18 under Plan 1374C is 19.6%, only slightly higher than it was under Plan 1151C (17.9%). And the percentage of Spanish-surnamed voters in Congressional District 18 under Plan 1374C in 2002 is 16%, only slightly higher than the 14.2% under Plan 1151C.232 Plaintiffs’ comparisons of voting patterns in the areas removed from and added to Congressional District 18 under Plan 1374C support the conclusion that the exchange does not dilute the strength of African-American voters in the district. In the U.S. Senate primary for 2002, which pitted African-American candidate Ron Kirk against Latino candidate Victor Morales, African-American candidate Kirk won by 78% in the area removed from the district, and 76% in the area added to the district.233 In the 2002 general election, the African-American Democratic candidate received 78.8% of the vote in the area removed from the district and 65.8% in the area added to the district.234 These changes are too slight to support the claim that the strength or status of Congressional District 18 as an effective African-American opportunity district, that has and will reliably elect the African-American candidate of choice, is diluted.

The projections of the population changes between 2000 and 2010 by Dr. Richard Murray, expert witness for Plaintiffs Jackson Lee and Johnson, do not alter this conclusion.235 Indeed, Dr. Murray’s projections shed little light on the issue because while he projected the growth of African-American and Hispanic voting age population, he simply did not include any projections of Hispanic citizen voting age population. The failure to include this information makes his projections interesting but of little value to assessing any trend toward future dilution in the relevant period.

As to Congressional District 30, Congresswoman Johnson remained confident that the district would continue to send her to Congress and to elect other African-American candidates of choice. *515Again, the figures support her conclusion. In Congressional District 30, under Plan 1374C, the Black citizen voting age population is 50.6%, while the Hispanic citizen voting age population is 15.6%. This is only slightly more than the Black citizen voting age population (48.6%) and only slightly more than the Hispanic citizen voting age population (14.3%) in Plan 1151C. And in Congressional District 30 under Plan 1374C, the Spanish-surnamed voters in 2002 accounted for only 12.5% of registered voters. The comparison of vote patterns verifies the continuing strength of Congressional District 30 as an effective African-American opportunity district in Plan 1374C. In the 2002 Democratic primary runoff for the Senate, the African-American candidate of choice, Ron Kirk, received 64.2% of the vote in the area removed from Congressional District 30 and Victor Morales, the Latino candidate of choice, received 35.8% of the vote.236 In the area added to Congressional District 30, Ron Kirk received 61.9% of the vote and Victor Morales received 38.1% of the vote.237 Again, the reduction in the votes received by the African-American candidate of choice is too slight to support any inference of dilution. And in the 2002 general election for the Senate, in the area removed from the district, Ron Bark received 39% of the vote, while he received 61.9% of the total vote in the area added to the district.238 Dr. Murray’s population projections are no more useful in understanding the likely African-American voting strength of Congressional District 30 in 2005 and 2010 than they were for Congressional District 18.239 Again, Dr. Murray neglects to include any analysis of the likely Hispanic citizen voting age population in the district in the future. He con-eludes that Congressional District 30 is an effective opportunity district for African-Americans now and offers no basis for believing that its status will change before the next decennial census.

The record, in short, offers no support for the dilution claim as to Congressional Districts 18 and 30. Rather, the arguments underscore that there is no cohesion between African-American and Latino voters in primary elections, particularly those that pit an African-American candidate against a Latino candidate. Although the congresswomen representing these districts ask this court to create districts that are almost exclusively African-American and to create a separate majority Latino citizen voting age population district in Dallas, there is no Gingles analysis to support this request and the record does not suggest that one could be provided. Plan 1374C does not violate § 2 of the Voting Rights Act.

VIII

For the foregoing reasons, we deny all relief requested by Plaintiffs. Judgment will be entered for Defendants.