Alexander Green, et al., Plaintiff-Intervenors, v. ABBOTT (2025)

MEMORANDUM OPINION AND ORDER GRANTING PRELIMINARY INJUNCTION

In August 2025, the State of Texas enacted a new electoral map to govern elections for the U.S. House of Representatives (the “2025 Map”). Claiming that the 2025 Map is racially discriminatory, six groups of Plaintiffs (the “Plaintiff Groups”) ask the Court to preliminarily enjoin the State from using the 2025 Map for the 2026 elections.

For the reasons explained below, the Court PRELIMINARILY ENJOINS the State from using the 2025 Map. The Court ORDERS that the 2026 congressional election in Texas shall proceed under the map that the Texas Legislature enacted in 2021 (the “2021 Map”).

I. INTRODUCTION

“The way to stop discrimination on the basis of race is to stop discriminating on the basis of race.”2

—Chief Justice John Roberts

The public perception of this case is that it's about politics. To be sure, politics played a role in drawing the 2025 Map. But it was much more than just politics. Substantial evidence shows that Texas racially gerrymandered the 2025 Map. Here's why.

Earlier this year, President Trump began urging Texas to redraw its U.S. House map to create five additional Republican seats. Lawmakers reportedly met that request to redistrict on purely partisan grounds with apprehension. When the Governor announced his intent to call a special legislative session, he didn't even place redistricting on the legislative agenda.

But when the Trump Administration reframed its request as a demand to redistrict congressional seats based on their racial makeup, Texas lawmakers immediately jumped on board. On July 7, Harmeet Dhillon, the head of the Civil Rights Division at the Department of Justice (“DOJ”), sent a letter (“the DOJ Letter”) to the Governor and Attorney General of Texas making the legally incorrect assertion that four congressional districts in Texas were “unconstitutional” because they were “coalition districts”—majority-non-White districts in which no single racial group constituted a 50% majority. In the letter, DOJ threatened legal action if Texas didn't immediately dismantle and redraw these districts—a threat based entirely on their racial makeup. Notably, the DOJ Letter targeted only majority-non-White districts. Any mention of majority-White Democrat districts—which DOJ presumably would have also targeted if its aims were partisan rather than racial—was conspicuously absent.

Two days later, citing the DOJ Letter, the Governor added redistricting to the special session's legislative agenda. In doing so, the Governor explicitly directed the Legislature to draw a new U.S. House map to resolve DOJ's concerns. In other words, the Governor explicitly directed the Legislature to redistrict based on race. In press appearances, the Governor plainly and expressly disavowed any partisan objective and instead repeatedly stated that his goal was to eliminate coalition districts and create new majority-Hispanic districts.

The Legislature adopted those racial objectives. The redistricting bill's sponsors made numerous statements suggesting that they had intentionally manipulated the districts’ lines to create more majority-Hispanic and majority-Black districts. The bill's sponsors’ statements suggest they adopted those changes because such a map would be an easier sell than a purely partisan one. The Speaker of the House also issued a press release celebrating that the bill satisfactorily addressed DOJ's “concerns.” Other high-ranking legislators stated in media interviews that the Legislature had redistricted not for the political goal of appeasing President Trump nor of gaining five Republican U.S. House seats, but to achieve DOJ's racial goal of eliminating coalition districts.

The map ultimately passed by the Legislature and signed by the Governor—the 2025 Map—achieved all but one of the racial objectives that DOJ demanded. The Legislature dismantled and left unrecognizable not only all of the districts DOJ identified in the letter, but also several other “coalition districts” around the State.

For these and other reasons, the Plaintiff Groups are likely to prove at trial that Texas racially gerrymandered the 2025 Map. So, we preliminarily enjoin Texas's 2025 Map.

II. BACKGROUND

A. The Law Governing Racial Discrimination Challenges to Redistricting Plans

Because “racial discrimination in voting ․ cannot coexist with democratic self-government,” federal law provides various avenues for challenging an electoral map as racially discriminatory.³ There are at least three avenues to do so.

1. Racial Gerrymandering

First, a plaintiff can bring a racial-gerrymandering claim under the Fourteenth and Fifteenth Amendments.⁴ A racial-gerrymandering claim alleges that the “State, without sufficient justification,” has “separat[ed] its citizens into different voting districts on the basis of race.”⁵ The plaintiff “must prove that the State subordinated race-neutral districting criteria ․ to racial considerations,” such that race was “the predominant factor motivating the legislature's decision to place a significant number of voters within or without a particular district.”⁶

2. Intentional Vote Dilution

Second, a plaintiff can bring an intentional vote-dilution claim, which is “analytically distinct from a racial-gerrymandering claim and follows a different analysis.”⁷ An intentional vote-dilution claim alleges that the State has “enacted a particular voting scheme as a purposeful device to minimize or cancel out the voting potential of racial or ethnic minorities.”⁸ Intentional vote dilution violates both the Constitution ⁹ and Section 2 of the Voting Rights Act (“VRA § 2”).¹⁰ To prevail on an intentional vote-dilution claim, “the plaintiff must show that the State's districting plan has the purpose and effect of diluting the minority vote.”¹¹

3. Effects-Based Vote Dilution (“Gingles” Claims)

Both of the first two claims require the plaintiff to prove that the Legislature acted with some sort of unlawful intent.¹² To supplement these intent-based causes of action, Congress amended VRA § 2 to enable plaintiffs to challenge electoral maps based on their racially dilutive effects alone.¹³

To prevail on an effects-based vote-dilution claim under VRA § 2, a plaintiff must satisfy what are known as the three “Gingles” preconditions.¹⁴ The first and second Gingles preconditions are both defined with reference to a “minority group”: Precondition #1 asks whether a “minority group [is] sufficiently large and geographically compact to constitute a majority in a reasonably configured district” that the Legislature could have drawn, while Precondition #2 asks whether “the minority group ․ is politically cohesive.”¹⁵

Critical to this case, the law governing how to define the requisite “minority group” has shifted over time. From 1988 to 2024, a Fifth Circuit case, Campos v. City of Baytown, permitted Gingles claimants to define the “minority group” as a coalition of two or more races.¹⁶ Campos thus permitted plaintiffs to satisfy the Gingles prerequisites by showing that it would be possible to draw a “coalition district”—a district in which no single race constitutes more than 50% of the voting population, but in which the total minority CVAP exceeds 50% in the aggregate.¹⁷ To avoid the possibility that a court might invalidate their districting plans under Campos, legislatures sometimes needed to preemptively enact maps that contained one or more coalition districts.

In 2024, however, the en banc Fifth Circuit overruled Campos in Petteway v. Galveston County.¹⁸ Petteway holds that “Section 2 of the Voting Rights Act does not authorize separately protected minority groups to aggregate their populations for purposes of a vote dilution claim.”¹⁹ To satisfy Gingles’s 50% threshold, a plaintiff in this Circuit must now prove that a single racial group could constitute a numerical majority in the plaintiff's proposed district—not a coalition of two or more racial groups.²⁰

Petteway changed the applicable standard only for effects-based vote-dilution claims under VRA § 2 and Gingles.²¹ Petteway did not modify the legal standards governing intentional vote-dilution claims or racial-gerrymandering claims under the Constitution because no such claims were before the en banc court.²²

Furthermore, Petteway holds only that “Section 2 does not require” legislatures “to draw precinct lines for the electoral benefit of” multiracial coalitions.²³ Petteway nowhere implies that legislatures must deliberately avoid drawing coalition districts—or that a legislatively drawn map that happens to contain one or more coalition districts is somehow unlawful.²⁴ This point is critical to this case.

4. Partisan Gerrymandering

In Rucho v. Common Cause, the Supreme Court ruled that partisan gerrymandering claims aren't cognizable in federal court.²⁵ Subject to legal restrictions that exist in some states, but not in Texas,²⁶ it is not illegal for a legislature to enact a redistricting plan with the purpose of favoring one political party over another.²⁷ When a plaintiff brings race-based gerrymandering claims, “partisan motivation [acts] as a defense, not a jurisdictional bar.”²⁸ These principles will likewise prove critically important below.²⁹

B. The 2021 Map

In 2021—four years before the Legislature enacted the 2025 Map challenged here—the State redrew its congressional map to account for population shifts in the 2020 census.³⁰ Four of the 2021 Map's congressional districts (“CDs”) are especially relevant here.

The first is CD 9, in the Houston area:

Although CD 9 was majority non-White under the 2021 Map, no single racial group constituted a 50%+ majority by CVAP. The district was 45.0% Black, 25.6% Hispanic, 18.1% White, and 9.3% Asian.³¹

The second relevant district is CD 18, also in the Houston area:

Like CD 9, CD 18 was majority non-White under the 2021 Map, with no single racial group constituting a 50%+ majority. The district was 38.8% Black, 30.4% Hispanic, 23.4% White, and 5.3% Asian.³²

The third relevant district is CD 29, also in the Houston area:

Unlike CD 9 and CD 18, the 2021 version of CD 29 was a single-race majority district—specifically, majority-Hispanic. By CVAP, the 2021 configuration of CD 29 was 63.5% Hispanic, 18.4% Black, 13.7% White, and 3.2% Asian.³³

The fourth relevant district is CD 33, in the Dallas/Fort Worth area:

Like CD 9 and CD 18, the 2021 version of CD 33 was majority non-White, with no single racial group constituting a 50%+ majority by CVAP. The district was 43.6% Hispanic, 25.2% Black, 23.4% White, and 5.7% Asian.³⁴

When the Legislature enacted the 2021 Map, the Fifth Circuit had not yet decided Petteway.³⁵ Because the 2021 versions of CDs 9, 18, and 33 were more than 50% non-White, with no single racial group constituting a numerical majority by CVAP, those districts were coalition districts.

The sponsor of the bill that became the 2021 Map, Senator Joan Huffman, stated repeatedly that the mapmakers did “not look[ ] at any racial data as [they] drew” the 2021 Map.³⁶ Instead, they based the district boundaries exclusively on race-neutral considerations like partisanship.³⁷ The plan that the mapmakers drew on partisan grounds appeared to also satisfy VRA § 2 as Campos interpreted it, so the Legislature passed the map.³⁸

If we take Senator Huffman at her word,³⁹ then any coalition districts that ended up in the 2021 Map were a coincidental by-product of the Legislature's decisions to draw district lines based on race-neutral considerations like partisanship. In other words, there's no evidence in the preliminary-injunction record that the Legislature purposefully drew coalition districts that it wouldn't have otherwise drawn based on concerns that a court would otherwise invalidate the 2021 Map under VRA § 2 and Campos.⁴⁰ Thus, there's no indication that the 2021 Legislature placed a thumb on the scale in favor of minority coalitions based on a now-discredited interpretation of § 2.

C. Calls to Redistrict for Political Purposes

Beginning in February or March 2025, and continuing in earnest in April and May, Republicans met with contacts in the White House to discuss the prospect of Texas redrawing its congressional map.⁴¹ On June 9, 2025, the New York Times published an article reporting that “President Trump's political team [was] encouraging Republican leaders in Texas to examine how House district lines in the state could be redrawn ahead of next year's midterm elections to try to save the party's endangered majority.”⁴² Contemporaneous press coverage indicated that partisan—rather than racial—motivations were behind the White House's redistricting push.⁴³

By all appearances, however, Republican lawmakers didn't have much appetite to redistrict on purely partisan grounds—even at the President's behest. The same New York Times article reported that “[t]he push from Washington ha[d] unnerved some Texas Republicans, who worr[ied] that reworking the boundaries of Texas House seats to turn Democratic districts red by adding reliably Republican voters from neighboring Republican districts could backfire in an election that is already expected to favor Democrats.”⁴⁴ “Rather than flip the Democratic districts,” Texas lawmakers feared that “new lines could endanger incumbent Republicans.”⁴⁵ At an emergency meeting in the Capitol shortly before the New York Times article was published, “congressional Republicans from Texas professed little interest in redrawing their districts.”⁴⁶

Perhaps due to this apparent lack of interest, when the Governor announced on June 23, 2025, that he was calling a special legislative session to address various issues, redistricting was not among them.⁴⁷ As far as some influential members of the Legislature were aware, the prospect of redistricting in 2025 was just a rumor.⁴⁸ In fact, at the bench trial this Court held on the 2021 Map in May–June 2025,⁴⁹ when counsel asked Senator Huffman whether “the Texas Legislature might be considering redrawing the [c]ongressional [d]istricts” as the New York Times had reported just one day earlier, Senator Huffman unequivocally responded: “They are not.”⁵⁰

D. The DOJ Letter

Instead, what ultimately spurred Texas to redistrict was a letter that DOJ sent to the Governor and the Texas Attorney General on July 7, 2025.⁵¹ The DOJ Letter exhorted Texas to redistrict for a very different reason than the political objectives mentioned in the New York Times article. Because the letter is critical to our analysis, we reproduce it here in full:

Re: Unconstitutional Race-Based Congressional Districts TX-09, TX-18, TX-29 and TX-33

Dear Governor Abbott and Attorney General Paxton,

This letter will serve as formal notice by the Department of Justice to the State of Texas of serious concerns regarding the legality of four of Texas's congressional districts. As stated below, Congressional Districts TX-09, TX-18, TX-29 and TX-33 currently constitute unconstitutional “coalition districts” and we urge the State of Texas to rectify these race-based considerations from these specific districts.

In Allen v. Milligan, 599 U.S. 1, 45, 143 S.Ct. 1487, 216 L.Ed.2d 60 (2023), Justice Kavanaugh noted that “even if Congress in 1982 could constitutionally authorize race-based redistricting under § 2 for some period of time, the authority to conduct race-based redistricting cannot extend indefinitely into the future.” 599 U.S. 1, 143 S.Ct. 1487, [sic] (Kavanaugh, J., concurring). In SFFA v. Harvard, the Supreme Court reiterated that “deviation from the norm of equal treatment” on account of race “must be a temporary matter.” 600 U.S. 181, 228, 143 S.Ct. 2141, 216 L.Ed.2d 857 (2023). When race is the predominant factor above other traditional redistricting considerations including compactness, contiguity, and respect for political subdivision lines, the State of Texas must demonstrate a compelling state interest to survive strict scrutiny.

It is well-established that so-called “coalition districts” run afoul the [sic] Voting Rights Act and the Fourteenth Amendment. In Petteway v. Galveston County, 111 F.4th 596 (5th Cir. 2024), the en banc Fifth Circuit Court of Appeals made it abundantly clear that “coalition districts” are not protected by the Voting Rights Act. This was a reversal of its previous decision in Campos v. City of Baytown, 840 F.2d 1240 (5th Cir. 1988). In Petteway, the Fifth Circuit aligned itself with the Supreme Court's decision in 52

Bartlett v. Strickland, 556 U.S. 1, 129 S.Ct. 1231, 173 L.Ed.2d 173 (2009), and determined that a minority group must be geographically compact enough to constitute more than 50% of the voting population in a single-member district to be protected under the Voting Rights Act. See also Thornburg v. Gingles, 478 U.S. 30, 106 S.Ct. 2752, 92 L.Ed.2d 25 (1986). Opportunity and coalition districts are premised on either the combining of two minority groups or a minority group with white crossover voting to meet the 50% threshold. Neither meets the first Gingle's [sic] precondition. Thus, the racial gerrymandering of congressional districts is unconstitutional and must be rectified immediately by state legislatures.

It is the position of this Department that several Texas Congressional Districts constitute unconstitutional racial gerrymanders, under the logic and reasoning of Petteway. Specifically, the record indicates that TX-09 and TX-18 sort Houston voters along strict racial lines to create two coalition seats, while creating TX 29, a majority Hispanic district. Additionally, TX-33 is another racially-based coalition district that resulted from a federal court order years ago, yet the Texas Legislature drew TX-33 on the same lines in the 2021 redistricting. Therefore, TX-33 remains as a coalition district.

Although the State's interest when configuring these districts was to comply with Fifth Circuit precedent prior to the 2024 Petteway decision, that interest no longer exists. Post-Petteway, the Congressional Districts at issue are nothing more than vestiges of an unconstitutional racially based gerrymandering past, which must be abandoned, and must now be corrected by Texas.

Please respond to this letter by July 7, 2025, and advise me of the State's intention to bring its current redistricting plans into compliance with the U.S. Constitution. If the State of Texas fails to rectify the racial gerrymandering of TX-09, TX-18, TX-29 and TX 33, the Attorney General reserves the right to seek legal action against the State, including without limitation under the 14th Amendment.53

It's challenging to unpack the DOJ Letter because it contains so many factual, legal, and typographical errors. Indeed, even attorneys employed by the Texas Attorney General—who professes to be a political ally of the Trump Administration ⁵⁴ —describe the DOJ Letter as “legally[ ] unsound,”⁵⁵ “baseless,”⁵⁶ “erroneous,”⁵⁷ “ham-fisted,”⁵⁸ and “a mess.”⁵⁹

The gist of the letter, though, is that DOJ is urging Texas to change the racial compositions of CDs 9, 18, 29, and 33. From the premise that Petteway forbids a plaintiff from proposing a coalition district for purposes of an effects-based vote-dilution claim under VRA § 2,⁶⁰ DOJ leaps to the conclusion that whenever a legislature enacts a map that happens to contain one or more coalition districts, that legislature has necessarily and unconstitutionally engaged in “racial gerrymandering.”⁶¹ The remedy for such racial gerrymandering, according to DOJ, is to change the offending districts’ racial makeup so that they no longer qualify as coalition districts.⁶²

That reading of Petteway is clearly wrong. Nowhere in Petteway does the Fifth Circuit hold that merely having a coalition district in an electoral map is per se unconstitutional.⁶³ The Petteway court had no occasion to opine about the constitutionality of coalition districts. Instead, the en banc court remanded the case to the district court to consider the plaintiffs’ constitutional claims in the first instance.⁶⁴

Nor could Petteway stand for such a proposition. That would contradict the Supreme Court's admonition that “the Constitution does not place an affirmative obligation upon the legislature to avoid creating districts that turn out to be heavily ․ minority.”⁶⁵ Rather, the Constitution “simply imposes an obligation not to create such districts for predominantly racial, as opposed to political or traditional, districting motivations.”⁶⁶ Thus, even though federal courts in this Circuit can no longer force a legislative body to create a coalition district under VRA § 2, that doesn't prohibit such a body from voluntarily creating a coalition district for political or other race-neutral reasons.⁶⁷

The Supreme Court's plurality opinion in Bartlett v. Strickland ⁶⁸ further reinforces this point. Even if VRA § 2 doesn't require a legislature to create a particular type of district, VRA § 2 and the Constitution don't prohibit the legislature from drawing that type of district. Nor is it lawful for a legislature to purposefully target such districts for destruction.⁶⁹ Bartlett involved a slightly different type of district ⁷⁰ —a “crossover district,” in which the minority population “make[s] up less than a majority of the voting-age population,” but “is large enough to elect the candidate of its choice with help from voters who are members of the majority and who cross over to support the minority's preferred candidate.”⁷¹ Much like Petteway would subsequently hold with respect to coalition districts, Bartlett held that a plaintiff may not satisfy the Gingles preconditions by proposing a crossover district.⁷² Thus, legislatures need not create crossover districts to avoid violating VRA § 2.⁷³

Critically, however, the Bartlett Court emphasized that its “holding that § 2 does not require crossover districts” did not address “the permissibility of such districts as a matter of legislative choice or discretion.”⁷⁴ The Supreme Court cautioned that Bartlett “should not be interpreted to entrench majority-minority districts by statutory command, for that, too, could pose constitutional concerns.”⁷⁵ The Court stressed that “States that wish to draw crossover districts are free to do so where no other prohibition [against such districts] exists.”⁷⁶ But the Bartlett Court also admonished that if a State “intentionally drew district lines in order to destroy otherwise effective crossover districts, that would raise serious questions under both the Fourteenth and Fifteenth Amendments.”⁷⁷

Fifteen years after Bartlett, Petteway determined that all the same legal considerations that apply to crossover districts apply equally to coalition districts.⁷⁸ To underscore the point, the Fifth Circuit took the Bartlett opinion, replaced each instance of the word “crossover” with “coalition,” and pronounced that the opinion's logic remained sound.⁷⁹

Performing Petteway’s word-replacement exercise with the above-quoted passages from Bartlett yields the following propositions: Petteway’s “holding that § 2 does not require [coalition] districts” has no bearing on “the permissibility of such districts as a matter of legislative choice or discretion.”⁸⁰ “States that wish to draw [coalition] districts are free to do so where no other prohibition exists.”⁸¹ “And if there were a showing that a State intentionally drew district lines in order to destroy otherwise effective [coalition] districts, that would raise serious questions under both the Fourteenth and Fifteenth Amendments.”⁸² Those propositions directly contradict the DOJ Letter's assertion that coalition districts are per se “unconstitutional”—as well as its argument that Texas can and must “rectify” any coalition districts that exist in the 2021 Map.⁸³

Besides those legal errors, the DOJ Letter also contains factual inaccuracies. Most egregiously, the letter lumps CD 29 in with CDs 9, 18, and 33 as examples of “coalition districts” that Texas must “rectify.”⁸⁴ As DOJ realizes halfway through the letter, however,⁸⁵ CD 29 was not a coalition district under the 2021 Map; it was a majority-Hispanic district.⁸⁶ Nothing in Petteway has any bearing on single-race-majority districts like CD 29,⁸⁷ so Petteway doesn't provide any legal basis to attack CD 29's racial composition.

All that said, DOJ might have had a decent argument if there were evidence that the Legislature intentionally drew the 2021 Map to include coalition districts that the Legislature wouldn't have otherwise drawn. As noted above, however, the preliminary-injunction record reveals no such thing. Again, nothing in the current record indicates that the Legislature drew the 2021 Map with an eye toward creating coalition districts. We thus presume that any coalition districts that ended up in the 2021 Map were coincidental by-products of the Legislature applying race-neutral redistricting criteria like partisanship.⁸⁸ There's consequently no indication that the Legislature would have drawn its maps differently if Petteway had been the governing law in 2021 instead of Campos.

Legally and factually, DOJ had no valid argument that the Legislature should restore the House map to some preexisting racial equilibrium since Petteway supplanted Campos. Far from seeking to “rectify ․ racial gerrymandering,”⁸⁹ the DOJ Letter urges Texas to inject racial considerations into what Texas insists was a race-blind process.

But what about DOJ's assertion that “TX-33 is [a] racially-based coalition district that resulted from a federal court order years ago”?⁹⁰ If a court forced Texas to draw CD 33 as a coalition district based on Campos’s discredited interpretation of VRA § 2, can't the Legislature redraw that district now that VRA § 2 no longer requires coalition districts?

The short answer is that this is another one of the DOJ Letter's many inaccuracies. It's true that CD 33 traces its lineage to a court-ordered map that a different three-judge panel of this Court imposed in 2012 when the State couldn't get its own map precleared under VRA § 5.⁹¹ It's also true that the three-judge panel based CD 33's boundaries partly on racial considerations.⁹² The challengers in the VRA § 5 preclearance proceedings had raised potentially viable claims that the Legislature had intentionally discriminated when drawing CD 33, and the panel configured CD 33 to address that concern.⁹³

But it's not true that the 2012 panel drew CD 33 as a “racially[ ] based coalition district” based on a now-overruled interpretation of VRA § 2.⁹⁴ Because the panel was “unable to conclude” that the plaintiffs were “likely to succeed on their § 2 claims premised upon coalition districts,” the panel said it would have been “inappropriate to intentionally create a coalition district on the basis of race or otherwise intentionally unite populations based on race.”⁹⁵ Thus, in its order imposing the court-drawn map, the panel emphasized that its configuration of CD 33 was “not a minority coalition district and was not drawn with the intention that it be a minority coalition district.”⁹⁶ In a subsequent order issued five years later, the panel again reiterated that “CD 33 was not intentionally drawn as a minority coalition district under § 2. Rather, it was created to remedy the alleged intentional discrimination (cracking) claims” raised in the VRA § 5 preclearance proceedings.⁹⁷

While it might be accurate to say that CD 33 ultimately became a coalition district based on its electoral performance and racial composition,⁹⁸ DOJ's implication that the Legislature purposefully drew CD 33 as a “racially-based coalition district” based on pre-Petteway law is demonstrably false.⁹⁹ Because the prior three-judge panel didn't force Texas to draw CD 33 as a coalition district under VRA § 2, nothing about Petteway’s subsequent reinterpretation of § 2 casts any doubt on CD 33's legality.

Even if the three-judge panel had drawn CD 33 as a coalition district based on VRA § 2 and Campos, CD 33's lines changed when the Legislature redistricted in 2021, as the blue arrows on the following maps reflect:

To the extent the DOJ Letter accuses the Legislature of “dr[awing] TX-33 on the same lines” as the 2012 court-drawn map “in the 2021 redistricting,”¹⁰⁰ that is also factually inaccurate.

The DOJ Letter is equally notable for what it doesn't include: any mention of partisanship.¹⁰¹ Had the Trump Administration sent Texas a letter urging the State to redraw its congressional map to improve the performance of Republican candidates, the Plaintiff Groups would then face a much greater burden to show that race—rather than partisanship—was the driving force behind the 2025 Map. But nothing in the DOJ Letter is couched in terms of partisan politics.¹⁰² The letter instead commands Texas to change four districts for one reason and one reason alone: the racial demographics of the voters who live there.¹⁰³

E. The Governor Adds Redistricting to the Legislative Agenda Immediately After Receiving the DOJ Letter

Though the Trump Administration's plea to redistrict for political reasons failed to gain any immediate traction,¹⁰⁴ the Administration's demand that Texas redistrict for racial reasons achieved quick results.¹⁰⁵ On July 9, 2025—just two days after the DOJ Letter ¹⁰⁶ —Governor Abbott issued a proclamation adding the following item to the agenda for the upcoming special legislative session: “Legislation that provides a revised congressional redistricting plan in light of constitutional concerns raised by the U.S. Department of Justice.”¹⁰⁷ The Governor shared—or, at minimum, wanted the Legislature to take legislative action to address—DOJ's “concerns” that CDs 9, 18, 29, and 33 were “unconstitutional” because of their racial makeup.¹⁰⁸

Like the DOJ Letter, the Governor's proclamation contains no request that the Legislature revise the congressional map for partisan purposes.¹⁰⁹ Here too, if the Governor had explicitly directed the Legislature to amend the congressional map to improve Republican performance, the Plaintiff Groups would then face a higher burden to prove that the motivation for the 2025 redistricting was racial rather than political.¹¹⁰ Instead, by incorporating DOJ's race-based redistricting request by reference, the Governor was asking the Legislature to give DOJ the racial rebalancing it wanted—and for the reasons that DOJ cited.

Contemporaneous media interviews reinforce that the Governor was asking the Legislature to redistrict for racial rather than partisan reasons. When asked during an August 11, 2025, press interview whether his decision to add redistricting to the legislative agenda was motivated by President Trump's demand for five additional Republican seats, the Governor demurred and insisted that the real impetus for redistricting was Petteway:

MR. TAPPER: The Texas Tribune reports that in June you told Texas Republicans delegation [sic] of Congress that you were reluctant to add redistricting to the legislative agenda in Austin. The Tribune says that President Trump then called you to discuss redistricting, and you agreed to put it on the special session agenda.

Would you have gone forward with redistricting if President Trump had not personally got involved and asked you to do this?

GOVERNOR ABBOTT: To be clear, Jake, this is something that I have been interested in for a long time.

First of all, I have been involved in redistricting litigation for more than 20 years now.

Second, one thing that spurred all this is a federal court decision that came out last year, by the way, a case that was filed by Democrats. The federal court decision that came out last year said that Texas is no longer required to have coalition districts. And as a result, we had drawn maps with coalition districts in it. Now we wanted to remove those coalition districts and draw them in ways that, in fact, turned out to provide more seats for Hispanics. For example, four of the districts are predominantly Hispanic. It just coincides it's going to be Hispanic Republicans elected to those seats.

One thing that's happened in the state of Texas is the Hispanic community, a lot of it, have [sic] decided they are no longer with the Democrats who believe in open border policies, who believe in going against our law enforcement, who believe that men should play in women's sports. And they instead align with Republicans.

What we want to do is to draw districts that give those Hispanics and African Americans in the state of Texas the ability to elect their candidate of choice.

MR. TAPPER: But that's not really—I mean, you are doing this to give Trump and Republicans in the House of Representatives five additional seats, right? I mean, that's the motivation, is to stave off any midterm election losses.

GOVERNOR ABBOTT: Again, to be clear, Jake, the reason why we are doing this is because of that court decision, Texas is now authorized under law that changed that was different than in 2021 when we last did redistricting. Under new law, as well as new facts that served us in the aftermath of the Trump election, showing that many regions of the state that historically had voted Democrat that were highly Hispanic now chose to vote Republican and vote for Trump as well as other Republican candidates. Districts where the electorate voted heavily for Trump, they were trapped in a Democrat congressional district that have every right to vote for a member of congress who is a Republican. We will give them that ability.111

When given an opportunity to publicly proclaim that his motivation for adding redistricting to the legislative agenda was solely to improve Republicans’ electoral prospects at President Trump's request, the Governor denied any such motivation.¹¹² Instead, the Governor expressly stated that his predominant motivation was racial: he “wanted to remove ․ coalition districts” and “provide more seats for Hispanics.”¹¹³ The fact that the racially reconfigured districts would happen to favor Republicans was, to paraphrase the Governor's own words, just a fortuitous coincidence.¹¹⁴

In other press statements around the same time, the Governor similarly stated that his motivation for directing the Legislature to redistrict was to eliminate coalition districts ¹¹⁵ —not for political reasons like appeasing President Trump.¹¹⁶ And the Governor consistently used language suggesting that he viewed the map's improved Republican performance not as an end in itself, but as a coincidental by-product of the plan's goal of increasing the number of majority-Hispanic districts.¹¹⁷

F. The Texas Attorney General's Response to the DOJ Letter

At the same time the Governor was announcing the 2025 Map's racial objectives to the press, the Attorney General of Texas was saying the opposite. Just two days after the Governor added redistricting to the legislative agenda based on DOJ's “constitutional concerns,”¹¹⁸ the Attorney General sent DOJ a response to its letter.¹¹⁹ That response said essentially the same thing we say above ¹²⁰ —that the change in law effected by Petteway cast no doubt on the legality of the 2021 Map, since there's no indication that the 2021 Legislature drew any coalition districts for legal-compliance reasons that it wouldn't have drawn anyway for race-neutral reasons like partisanship.¹²¹ Although the Attorney General doesn't say so explicitly, the purpose behind his letter appears to have been to refocus the redistricting dialogue toward permissible considerations like partisanship, politics, and traditional districting criteria—and away from legally fraught considerations like race.¹²²

If that was the letter's purpose, it didn't work. The Governor continued to declare publicly that Petteway was the impetus for the 2025 redistricting, and that Texas's reason for redistricting was to change the map's racial characteristics by eliminating coalition districts and increasing the number of majority-Hispanic districts.¹²³ And the Legislature proceeded to do just that.

G. The Legislature Enacts the 2025 Map

Ultimately, the 2025 Map did all but one of the things that DOJ and the Governor expressly said they wanted the Legislature to do.

1. CD 9

First, the Legislature eliminated CD 9's status as a coalition district by making it a district in which a single racial group (Hispanics) are just barely a majority by CVAP (50.3%).¹²⁴ By doing so, the Legislature simultaneously satisfied not just DOJ's command that Texas convert CD 9 from a coalition district to a single-race-majority district, but also the Governor's goal of increasing the number of majority-Hispanic districts in the State. The Legislature reached that outcome by reconfiguring CD 9's boundaries so radically that only 2.9% of the people who were in CD 9 under the 2021 Map remain in the district under the 2025 Map:¹²⁵

2. CD 18

The Legislature likewise eliminated CD 18's status as a coalition district—another one of the “asks” in DOJ's Letter ¹²⁶ —by making it just barely a majority Black district (50.5%).¹²⁷ The Legislature did so primarily by importing large numbers of predominantly Black voters from CD9.¹²⁸

3. CD 29

Perhaps perplexed by DOJ's request to “rectify” CD 29's status as a “coalition” district when it wasn't actually a coalition district,¹²⁹ the Legislature eliminated CD 29's status as a majority-Hispanic district. Under the 2025 Map, CD 29's Hispanic CVAP drops from 63.5% to 43.3.¹³⁰ Here too, the Legislature achieved that result by radically reconfiguring the district's boundaries ¹³¹ to remove various Latino communities.¹³²

4. CD 33

There is, admittedly, one thing that DOJ requested that the Legislature didn't do: eliminate CD 33's status as a coalition district.¹³³ Under both the 2021 Map and the 2025 Map, CD 33 remains majority non-White.¹³⁴ Nevertheless, the district—like CDs 9, 18, and 29—is completely reconfigured and unrecognizable when compared to the old CD 33:

5. Other Districts Converted to Single-Race-Majority Districts (CDs 22, 27, 30, 32, and 35)

In keeping with the spirit of DOJ's request, the Legislature also eliminated five coalition districts that DOJ didn't mention.¹³⁵

First was CD 22. Under the 2021 Map, CD 22 was just shy of being a majority-White district (49.2%).¹³⁶ The remaining 50.8% was made up of voters of various other races, making the district majority-non-White.¹³⁷ Thus, at least with respect to its racial composition (though maybe not with respect to its electoral performance),¹³⁸ the 2021 version of CD 22 could have been described as a coalition district. The 2025 Map increased CD 22's White CVAP to 50.8%, thereby making it just barely a single-race-majority district:¹³⁹

Second was CD 27. No single race constituted a majority in the 2021 version of CD 27 either; the electorate was split relatively equally between Hispanics (48.6%) and Whites (44.1%), with voters of other races constituting the remainder.¹⁴⁰ Here too, CD 27 could be described as a coalition district with respect to its racial composition, even if it might not be so described with respect to its electoral performance.¹⁴¹ The 2025 Map increased CD 27's White CVAP to 52.8% while decreasing Hispanic CVAP to 36.8%—thereby making CD 27 another new single-race-majority district:¹⁴²

Third was CD 30. Under the 2021 Map, CD 30 was a coalition district: it was majority non-White by CVAP, with no single racial group constituting more than 50% of eligible voters.¹⁴³ The 2025 Map converts CD 30 to a single-race-majority district by making it just barely majority-Black (50.2%):¹⁴⁴

Fourth was CD 32. Although Whites constituted a plurality of eligible voters (43.9%) under the 2021 version of CD 32, it was nevertheless a majority-non-White coalition district.¹⁴⁵ The 2025 Map radically reshapes the boundaries of CD 32 and converts it to a single-race-majority district by making it 58.7% White:¹⁴⁶

The final district was CD 35, which was also a coalition district.¹⁴⁷ The 2025 Map converts CD 35 to a single-race-majority district by making it just barely majority Hispanic (51.6%):¹⁴⁸

In sum, the 2025 Map:

(1) fundamentally changed the racial character of three of the four districts identified in the DOJ Letter, and dramatically dismantled and left unrecognizable all four districts;

(2) eliminated seven total coalition districts;

(3) created two new bare-majority-Hispanic districts, while eliminating an existing strongly majority-Hispanic district identified in the DOJ Letter; and

(4) created two new bare-majority-Black districts.

H. The Plaintiff Groups’ Preliminary Injunction Motions

Immediately after the Texas Senate passed the 2025 Map on August 23, 2025—and, indeed, before the Governor even signed the bill ¹⁴⁹ —the Plaintiff Groups moved to preliminarily enjoin the State from using the 2025 Map for the upcoming U.S. House elections.¹⁵⁰ The Plaintiff Groups’ theory of the case is that:

(1) DOJ unlawfully demanded that Texas “redraw certain congressional districts because of their multiracial majority status”;

(2) “In response, Governor Abbott called the Texas Legislature into a Special Session specifically to eliminate the coalition and majority minority districts identified by DOJ”; and

(3) “Over the course of the redistricting process ․ the Governor, DOJ, and multiple Texas legislators repeatedly, publicly, and explicitly stated that Texas was redistricting to eliminate multiracial majority districts.”151

The Plaintiff Groups thus claim that Texas's actions in the 2025 redistricting amount to unconstitutional racial gerrymandering.¹⁵² Altogether,¹⁵³ the Plaintiff Groups challenge the following districts on racial-gerrymandering grounds: CDs 9, 18, 22, 27, 30, 32, 33, and 35.¹⁵⁴

The State Defendants, by contrast, insist that the motives underlying the 2025 redistricting were exclusively partisan and political ¹⁵⁵ —not racial.¹⁵⁶ According to the Defendants, the Legislature enacted the 2025 Map solely to satisfy President Trump's demand that Texas create five more Republican seats in the U.S. House of Representatives ¹⁵⁷ and counteract threatened partisan gerrymanders in Democrat states.¹⁵⁸

To resolve the preliminary-injunction motions, the Court held a nine-day hearing from October 1–10, 2025, at which the parties introduced voluminous documentary and testimonial evidence. Having now carefully reviewed that evidence and the applicable caselaw, the Court rules as follows.

III. DISCUSSION

A. The Legal Standard for Obtaining a Preliminary Injunction

To obtain a preliminary injunction, the Plaintiff Groups must show:

(1) “a likelihood 159 of success on the merits” of their claims;

(2) “a likelihood of suffering irreparable harm if an injunction is not granted;”

(3) “that the balance of equities tips in their favor;” and

(4) “that an injunction would serve the public interest.”160

“In considering these four prerequisites, the court must remember that a preliminary injunction is an extraordinary and drastic remedy which should not be granted unless the movant clearly carries the burden of persuasion.”¹⁶¹

B. The Plaintiff Groups Have Demonstrated a Likelihood of Success on the Merits of Most of their Racial-Gerrymandering Claims

For the reasons explained below, the Plaintiff Groups have successfully shown a likelihood of success on their racial-gerrymandering challenges to CDs 9, 18, 27, 30, 32, and 35.¹⁶² Because that alone suffices to preliminarily enjoin the 2025 Map—and given the short timeframe the Court had to write this complex and record-intensive opinion—the Court will not address the Plaintiff Groups’ intentional vote-dilution claims at this time.¹⁶³

1. Applicable Procedural Standards

The “likelihood of success on the merits” factor is “the most important.”¹⁶⁴ To demonstrate a likelihood of success on the merits, the Plaintiff Groups don't need to prove that they're definitely going to win at the trial on the merits; they need only prove that they're likely to win at trial.¹⁶⁵

The exact quantum of evidence that a plaintiff must present to satisfy the likelihood-of-success factor varies from case to case.¹⁶⁶ The Fifth Circuit applies a “sliding scale” approach, whereby a plaintiff who makes a strong showing on the other three preliminary injunction factors bears a lesser burden on the likelihood-of-success requirement (and vice versa).¹⁶⁷ “Where the other factors are strong,” the movant need only show “some likelihood of success on the merits” to obtain a preliminary injunction.¹⁶⁸

To preview our conclusions below, the Plaintiff Groups have made a very strong showing on the irreparable-injury factor ¹⁶⁹ and a compelling showing on the balance-of-equities and public-interest factors.¹⁷⁰ Under the Fifth Circuit's sliding-scale approach, the Plaintiff Groups need to show more than just “some likelihood of success on the merits” to obtain a preliminary injunction, but not much more.¹⁷¹

2. Applicable Substantive Standards

“To assess the likelihood of success on the merits,” we must “look to standards provided by the substantive law.”¹⁷²

A plaintiff asserting a racial-gerrymandering claim may “make the required showing through direct evidence of legislative intent,”¹⁷³ such as “a relevant state actor's express acknowledgement that race played a role in the drawing of district lines,”¹⁷⁴ “circumstantial evidence of a district's shape and demographics, or a mix of both.”¹⁷⁵ The court must “make a sensitive inquiry into all circumstantial and direct evidence of [the legislature's] intent” to determine whether “race ․ drove [the challenged] district's lines.”¹⁷⁶

Although a plaintiff pressing a racial-gerrymandering claim need not prove that the enacted map has racially dilutive effects,¹⁷⁷ there are several other significant obstacles that a racial-gerrymandering plaintiff must surmount. First, in a state like Texas—where race and partisan affiliation are closely correlated ¹⁷⁸ —“a map that has been gerrymandered to achieve a partisan end can look very similar to a racially gerrymandered map.”¹⁷⁹ Again, though, partisan-gerrymandering claims aren't cognizable in federal court.¹⁸⁰ So, to prevail on a racial-gerrymandering claim, “a plaintiff must disentangle race from politics by proving that the former drove a [challenged] district's lines.”¹⁸¹

Second, the mere fact that a legislature was aware of a particular district's racial demographics when it made its districting decisions doesn't necessarily mean that the legislature engaged in illegal racial gerrymandering. “Redistricting legislatures will ․ almost always be aware of racial demographics[,] but it does not follow that race predominates in the redistricting process.”¹⁸² Thus, litigants and courts must be mindful of “[t]he distinction between being aware of racial considerations and being motivated by them.”¹⁸³

Finally—and most importantly—“federal courts must exercise extraordinary caution in adjudicating claims that a State has drawn district lines on the basis of race.”¹⁸⁴ “The Constitution entrusts state legislatures”—not federal courts—“with the primary responsibility for drawing congressional districts.”¹⁸⁵ “Federal-court review of districting legislation” thus “represents a serious intrusion on the most vital of local functions.”¹⁸⁶

Aside from those federalism concerns, federal courts must also be mindful that “[e]lectoral districting is a most difficult subject for legislatures” and that “the States must have discretion to exercise the political judgment necessary to balance competing interests.”¹⁸⁷ Courts must therefore “be sensitive to the complex interplay of forces that enter a legislature's redistricting calculus.”¹⁸⁸

For those reasons, courts must “presum[e] that the legislature acted in good faith” when devising and enacting a redistricting plan.¹⁸⁹ When “confronted with evidence that could plausibly support” either a racial or a non-racial motivation for a legislature's action, “district courts [must] draw the inference that cuts in the legislature's favor.”¹⁹⁰

“If a plaintiff can demonstrate that race drove the mapping of district lines, then the burden shifts to the State”¹⁹¹ “to prove that its race-based sorting of voters serves a ‘compelling interest’ and is ‘narrowly tailored” to that end.’ ”¹⁹² The Court will expound on those requirements below.¹⁹³

3. Direct Evidence of Racial Gerrymandering

The direct evidence here is strong. In conjunction with the circumstantial evidence discussed below,¹⁹⁴ the direct evidence indicates that the Plaintiff Groups have more than some likelihood of prevailing on their racial-gerrymandering claims at trial.

a. DOJ Asked Texas to Engage in Unlawful Racial Gerrymandering

By directing Texas to “separate its citizens into different voting districts on the basis of race,” DOJ directed Texas to engage in racial gerrymandering.¹⁹⁵ The letter asserts—incorrectly ¹⁹⁶ —that CDs 9, 18, and 29, and 33 are unlawful because they happen to be coalition districts.¹⁹⁷ That is, the districts are objectionable to DOJ solely because of their racial composition.¹⁹⁸ Although the letter doesn't specify how DOJ wants Texas to “rectify” and “correct[ ]” the listed districts,¹⁹⁹ there's only one way to remedy a district whose only “objectionable” characteristic is that no single racial group constitutes a 50% majority by CVAP: redraw it so a single racial group constitutes a 50% majority by CVAP.²⁰⁰ We therefore interpret the DOJ Letter as imposing a 50% racial target for Texas to meet when redrawing its districts.

Our interpretation—that DOJ commanded Texas to meet a 50% racial target—is consistent with the map the Legislature ultimately passed. As discussed, the Legislature took two of the three true coalition districts mentioned in the DOJ Letter and increased their CVAP figures to just barely over 50%: CD 9 (50.3% Hispanic); CD 18 (50.5% Black).²⁰¹

Supreme Court precedent establishes that when:

(1) a relevant political actor “purposefully establishe[s] a racial target” that voters of a single race “should make up no less than a majority” of the voting population; and

(2) the Legislature “follow[s] those directions to the letter, such that the 50%-plus racial target ha[s] a direct and significant impact on [the districts’] configuration,”

a court may permissibly conclude “that race predominated in drawing” those districts.²⁰² DOJ and the Governor did the first of those things. The Legislature did the second.

b. The Governor's Actions Suggest a Predominantly Racial Motivation

i. The Governor's Proclamation

By explicitly referring to DOJ's “constitutional concerns” in his proclamation,²⁰³ the Governor:

(1) endorsed DOJ's erroneous view that Petteway required the Legislature to fundamentally change the targeted districts’ racial character;204 and

(2) exhorted the Legislature to redistrict for the same racial reasons that DOJ gave in its letter.

The DOJ Letter is dated July 7. On July 9, the Governor issued a proclamation adding redistricting to the legislative agenda to advance DOJ's racial objectives. This close temporal proximity undermines the State Defendants’ position that the motivation for the 2025 redistricting was political rather than racial. Lawmakers initially showed little appetite to redistrict when the Trump Administration pressed the State to redistrict for exclusively partisan reasons.²⁰⁵ What triggered the redistricting process was the Administration reframing the request in exclusively racial terms.²⁰⁶

ii. The Governor's Contemporaneous Press Statements

In his contemporaneous press statements, the Governor framed his objectives for the 2025 redistricting in slightly different terms than the DOJ Letter. Governor Abbott said that Petteway permitted Texas to “remove ․ coalition districts” from the congressional map, and that this provided an opportunity for the Legislature to replace those coalition districts with majority-Hispanic districts, as opposed to single-race-majority districts more generally.²⁰⁷ That was fortuitous, according to the Governor, because many Hispanic voters had recently “decided they're no longer with the Democrats who believe in open border policies, who believe in going against our law enforcement[,] who believe that men should play in women's sports[,] and they instead align with the Republicans.”²⁰⁸ The purpose behind the 2025 redistricting was to “take the people who were in those coalition districts”—specifically, “Hispanics and [B]lacks”—and place them “in districts that really represent the voting preference[ ] of those people who live ․ in Texas.”²⁰⁹

That's a stark admission. The Governor wanted Texas to “use[ ] race as a basis for separating voters into districts.”²¹⁰ According to the Governor, the 2025 Map's modus operandi was to:

(1) specifically target Hispanic and Black voters based on the assumption that Texan voters of color—especially Hispanics—now trend Republican;211

(2) take those voters out of their existing districts; and

(3) place those voters into new districts—all because of their race.

That's tantamount to using “race ․ as a proxy for political characteristics” and “stereotyp[ing]” voters based on race.²¹² “[D]istricting decisions that rely on stereotypes about racial voting are constitutionally suspect.”²¹³ As the Supreme Court has explained, “[w]hen the State assigns voters [to particular districts] on the basis of race, it engages in the offensive and demeaning assumption that voters of a particular race, because of their race, think alike, share the same political interests, and will prefer the same candidates at the polls.”²¹⁴

At the same time, Governor Abbott consistently rejected the idea that Texas was redistricting to fulfill President Trump's demand for additional Republican districts.²¹⁵ The Governor “subordinated race-neutral districting criteria” like partisanship “to racial considerations.”²¹⁶ Race—not politics—was “the predominant factor motivating the ․ decision to place a significant number of voters within or without a particular district.”²¹⁷

c. The Motives of State and Federal Executive Branch Actors Aren't Automatically Imputable to the Legislature

The mere fact that the federal and state executive branches told the Legislature to engage in racial gerrymandering is not dispositive. “[L]egislators who vote to adopt a bill are not the agents of the bill's sponsor or proponents,” as “legislators have a duty to exercise their [own independent] judgment” when crafting and passing legislation.²¹⁸ What ultimately matters is the Legislature's motivation for devising and enacting the 2025 Map—not the motivations of political actors outside the legislative branch.²¹⁹ The unlawful motivations of DOJ and the Governor “do not become those of the [Legislature] as a whole unless it is shown that a majority of the [Legislature's] members shared and purposefully adopted (i.e., ratified) the [Governor and DOJ's] motivations.”²²⁰

The Northern District of Florida's recent decision in Common Cause Florida v. Byrd illustrates this point. There, the Governor of Florida proposed a congressional districting map that eliminated a district that elected Black voters’ candidates of choice.²²¹ The Florida Legislature ultimately enacted that map.²²²

The district court assumed without deciding that the Governor had “acted with some unlawful discriminatory motive in creating and proposing the redistricting map that was ultimately enacted into law.”²²³ Even assuming that “the Governor was motivated in part by racial animus,” however, the plaintiffs also needed to “prove that the Florida Legislature itself acted with some discriminatory purpose when adopting and passing the Enacted Map”²²⁴ —such as by introducing “evidence that the [legislators] themselves agreed with the discriminatory motives,” or that they passed the map “for the purpose of giving effect to the [Governor's alleged] discriminatory motives.”²²⁵ Because “not one legislator said or did anything to suggest ․ that any legislator voted for the Enacted Map because they shared or intended to effectuate any racially discriminatory motive on the Governor's part,” the plaintiffs failed to prove “that the Legislature acted with race as a motivating factor in passing the Enacted Map.”²²⁶

d. Legislators’ Statements

This case is very different from Common Cause Florida. Direct evidence in the preliminary-injunction hearing shows that key legislators in the 2025 redistricting process had the same racial objectives as DOJ and the Governor.

i. Speaker Burrows

When the Texas House passed the 2025 Map, the Speaker of the House, Representative Dustin Burrows, issued a press release favorably announcing that the House had just “delivered legislation to redistrict certain congressional districts to address concerns raised by the Department of Justice and ensure fairness and accuracy in Texans’ representation in Congress.”²²⁷ This press release publicly announces that high-ranking legislators honored and followed the instruction in the Governor's proclamation to redistrict for the racial reasons cited in the DOJ Letter.²²⁸ The Speaker's press release also undermines other legislators’ assertions (discussed below) that the DOJ Letter did not influence the Legislature during the 2025 redistricting process.²²⁹

In the same press release, the Speaker also praised the House for “deliver[ing] the legal, remedied maps Texas voters deserve.”²³⁰ Speaker Burrows shared DOJ's erroneous view that the 2021 Maps were illegal because they contained coalition districts and that the Legislature needed to “remedy” that defect by extirpating those districts.

To be sure, the press release is also peppered with statements that could suggest a partisan motive. Speaker Burrows celebrates that “the new map ․ secures Republican representation in Congress.”²³¹ For that reason, the press release does not establish by itself that race predominated over partisan concerns during the 2025 redistricting cycle. But the press release is not the only direct evidence of racial motivation in the record.

ii. Representative Oliverson

In contemporaneous interviews and press releases, several other high-ranking legislators espoused that the Legislature's motivation for redistricting was not to fulfill President Trump's demand for more Republican congressional seats, but rather to eliminate coalition districts as DOJ requested. In an August 6, 2025, interview with National Public Radio (“NPR”), the Chair of the Texas House Republican Caucus, Representative Tom Oliverson, said the following:

AILSA CHANG: ․ So this congressional map. It's being redrawn after your party already drew it in 2021. And one of the main objections to what you all are doing is that Texas Republicans are doing this only because President Trump asked you to do so.

Let me just ask you directly. Is that true? Are you redoing this map now specifically because of the [P]resident's request?

REP. TOM OLIVERSON: No, we are not. And in fact, the first conversations that I heard about and had myself regarding redistricting began before the legislative session began in January as a result of a court case where a federal appeals court basically rejected the idea of the coalition districts as being consistent with the Voting Rights Act.232

Another stark admission: the desire to eliminate coalition districts drove the 2025 redistricting—not pressure from President Trump to redistrict for partisan gain.

iii. Representative Toth

In a press interview following the 2025 Map's enactment, Representative Steve Toth similarly insisted that the motive behind the 2025 redistricting was not to achieve political gains, but rather because DOJ had commanded Texas to redistrict in response to Petteway:

JOHN SOLOMON: ․ [Y]ou pointed out something important here, which is that the storyline Democrats and their liberal friends like to say is, oh, this is being done by Texas for gerrymandering and for political gain in the [2026] election. But in fact, the Justice Department required the state to do this because there were appellate court rulings that said Texas was out of compliance with the current law. So, this isn't actually gerrymandering. This was actually required to be done, right?

STEVE TOTH: It was required of us to do it in ․ response to Petteway to get compliant.233

Like the Governor, Speaker Burrows, and Representative Oliverson, Representative Toth shared DOJ's erroneous legal position that Petteway affirmatively required Texas to eliminate coalition districts. He therefore shared and adopted DOJ's racial objective of erasing coalition districts from the map. Representative Toth's statements reinforce that “Justice Department pressure led the State to act based on an overriding concern with race.”²³⁴

iv. Chairman Hunter and His Joint Authors

Further evidence that race was a key factor motivating the 2025 redistricting comes from Chairman Todd Hunter's statements and exchanges with other legislators on the House floor.²³⁵ Because Chairman Hunter introduced and championed the bill that ultimately became the 2025 Map,²³⁶ we consider his and his joint authors’ statements to be more probative of the full Legislature's intent than those of other legislators.²³⁷

Chairman Hunter introduced a redistricting bill on July 30, 2025, during the first special legislative session.²³⁸ With certain changes, the Legislature would ultimately pass Chairman Hunter's bill in the second special session.²³⁹ In his August 1, 2025, layout of that bill,²⁴⁰ Chairman Hunter volunteered—without prompting from any other legislator ²⁴¹ —that “four of the five” new Republican districts proposed by the bill were “majority[-]minority Hispanic CVAP districts.”²⁴² Chairman Hunter likewise volunteered, again without prompting:²⁴³

(1) that the introduced map increased the total number of majority-Hispanic 244 and majority-Black 245 congressional districts; and

(2) the CVAP statistics for the majority-Hispanic 246 and majority-Black 247 districts in the introduced plan.248

Taken by themselves, those factual statements about the bill's racial statistics do not imply anything more than mere awareness of race, which is not actionable.²⁴⁹ Chairman Hunter could have had an innocuous reason to preemptively mention the districts’ racial characteristics in his layout—namely, to stave off the criticism that opposing legislators had made during the previous redistricting cycle, which was that he didn't have certain racial data ready in response to legislators’ questions.²⁵⁰ These statements alone do not clear the presumption of legislative good faith.²⁵¹ But the combination of these statements with Chairman Hunter's additional direct evidence overcomes that presumption.

Chairman Hunter's floor statements and exchanges with other legislators suggest that he and the bill's joint authors viewed the plan's racial numbers not merely as raw statistical facts, but as selling points of the bill. After Chairman Hunter's layout,²⁵² a Republican legislator and one of the bill's joint authors, Representative Katrina Pierson,²⁵³ engaged in a colloquy with Chairman Hunter about the proposed plan's racial makeup. The purpose of that exchange was apparently to elicit for the legislative record that, by increasing the number of majority-Black districts, the bill would improve representation for voters of color, thereby addressing concerns about minority representation raised earlier in the legislative process.²⁵⁴ One of the bill's other joint authors, Representative David Spiller,²⁵⁵ likewise engaged in a colloquy with Chairman Hunter. In this colloquy, Representative Spiller emphasized that the proposed map increased the number of majority-Black and majority-Hispanic districts to rebut opponents’ arguments that the map was “racially motivated” and “race negative.”²⁵⁶ Chairman Hunter himself said multiple times during the process that it was “important [for other legislators] to note that four of the five new [Republican] districts [were] majority[-]minority Hispanic CVAP districts.”²⁵⁷ He said it was “good,” “great,” and a “strong message” that those four districts were majority-Hispanic.²⁵⁸ Chairman Hunter also made value-laden statements indicating that he thought his map's racial numbers were “better” and “improv[ed]” over the 2021 Map.²⁵⁹

The joint authors also repeatedly invoked Petteway. Chairman Hunter referred again and again to Petteway as one of the main impetuses for the 2025 redistricting.²⁶⁰ He said that he and his joint authors had “redrawn the congressional map” based on Petteway’s “clarification” that “Section 2 does not require [the Legislature] to draw coalition districts.”²⁶¹ He likewise commented that Petteway had given the Legislature a new “justification ․ to look at redistricting” since the 2021 Map's enactment.²⁶² And he indicated that his proposed map had taken into account Petteway’s holding that “there's not a requirement” to have coalition districts.²⁶³ That all suggests that the mapdrawers purposefully manipulated the districts’ racial demographics to convert coalition districts into single-race-majority districts.

Chairman Hunter's exchanges with Representative Spiller reinforce this point. Representative Spiller shared DOJ's mistaken view that Petteway “compelled” the Legislature to systematically eliminate coalition districts from the 2021 Map.²⁶⁴ Representative Spiller and Chairman Hunter identified districts that the bill would transform from coalition districts into single-race-majority districts.²⁶⁵ In doing so, they emphasized that changing the coalition districts in this way brought the map into “compliance” with Petteway.²⁶⁶

Finally, when a legislator from the opposing party directly asked Chairman Hunter whether he had “purposely altered” certain coalition districts to make them single-race-majority districts, Chairman Hunter did not deny that he had.²⁶⁷

All the evidence discussed so far overcomes the presumption of legislative good faith. Chairman Hunter and the other joint authors evidently strategized that a map that eliminated coalition districts and increased the number of majority-Hispanic and majority-Black districts would be more “sellable” than a nakedly partisan map.²⁶⁸ The legislators could point to the map's increased number of majority-minority districts to rebut accusations of racism.²⁶⁹ The Governor could promote the map to Hispanic voters who might be inclined to swing Republican.²⁷⁰ And legislators could deny they were redistricting for purely partisan reasons or to placate President Trump, and instead say that DOJ and Petteway had forced their hand.²⁷¹ It was, therefore, critical for the redistricting bill's authors to compile a legislative record replete with racial statistics and references to Petteway—which is exactly what they did.

Even though partisanship was undoubtedly a motivating factor in the 2025 redistricting process, “race was the criterion that, in the State's view, could not be compromised.”²⁷² It wasn't enough for the map to merely improve Republican performance; it also needed to convert as many coalition districts to single-race-majority districts as possible. That best explains the House bill's authors’ comments during the legislative process and the map's stark racial characteristics. The bill's main proponents purposefully manipulated the districts’ racial numbers to make the map more palatable. That's racial gerrymandering.²⁷³

We reach that conclusion even though Chairman Hunter stated repeatedly that the bill was primarily driven by non-racial partisan motivations.²⁷⁴ Chairman Hunter often referred to Rucho as another primary driver for the 2025 redistricting—sometimes in the same breath as Petteway,²⁷⁵ sometimes not.²⁷⁶ Chairman Hunter also stated on the House floor that he was “not guided” by the DOJ Letter in the redistricting process.²⁷⁷ He mentioned at various times that he had taken other race-neutral districting criteria like compactness into account.²⁷⁸ And he said on the floor that he “didn't go at” any coalition districts.²⁷⁹

But if Chairman Hunter's motives were exclusively partisan as the State Defendants contend, why mention Petteway at all? Why not just base the 2025 redistricting exclusively on Rucho?²⁸⁰ The answer must be that race and Petteway were essential ingredients of the map, without which the 2025 redistricting wouldn't have occurred.²⁸¹ The fact that Rucho was already the law when the Legislature redistricted in 2021 ²⁸² further cements the notion that Petteway was the primary driver behind the 2025 redistricting. Petteway was the only thing about the legal landscape that had changed since 2021.²⁸³

4. Contrary Direct Evidence of Legislative Intent

The State Defendants’ contrary direct evidence regarding the Legislature's intent primarily comes from:

(1) Senator Phil King, the Chairman of the Senate Redistricting Committee and the sponsor of the Senate counterpart to the House redistricting bill;284

(2) Senator Adam Hinojosa; and

(3) Representative Cody Vasut, who was the Chairman of the House Select Committee on Congressional Redistricting in 2025.285

These legislators each testified at the preliminary-injunction hearing that race played no role in the 2025 redistricting process. But their testimony is less probative than the Plaintiff Groups’ evidence.

a. Chairman King

At the preliminary-injunction hearing, as well as on the Senate floor, Chairman King insisted that the DOJ Letter did not motivate his votes and actions during the 2025 redistricting process.²⁸⁶ He claimed that he did not look at racial data at all,²⁸⁷ and that, to his knowledge, the 2025 Map was drawn blind to race.²⁸⁸

Chairman King maintained that his goals in the 2025 redistricting were to achieve three lawful, race-neutral objectives:

(1) to increase the likelihood that the districts would elect Republicans;

(2) to enact a map that complied with all applicable law; and

(3) to make several of the districts more compact.289

Chairman King said that he sponsored and voted for the enacted map because it achieved all three of those race-neutral objectives.²⁹⁰ Chairman King further testified that the motives of the Legislature as a whole were partisan rather than racial.²⁹¹ Chairman King's testimony thus supports the State Defendants’ position that race didn't play any role whatsoever, let alone predominate, in the 2025 redistricting process.²⁹²

For the following reasons, though, we find Chairman King's testimony and legislative statements less probative of the Legislature's intent than those of Speaker Burrows, Chairman Hunter, Representative Oliverson, Representative Toth, Representative Spiller, and Representative Pierson.

i. Chairman King's Minimal Role in the Redistricting Process

First, Chairman King played a much less significant role in the 2025 Map's development and passage than other legislators, even though he served as Chairman of the Senate Redistricting Committee. He testified that the House—not the Senate—took “the lead on redistricting.”²⁹³ He further admitted that he played “[no role] whatsoever” in drafting the map that the Legislature ultimately enacted.²⁹⁴ Chairman King merely took the same map that the House had introduced during the Legislature's first special session and introduced it, unchanged, in the Senate.²⁹⁵ He stated on the Senate floor that he didn't “really have any personal knowledge of the inner workings that went into who participated in drawing the maps.”²⁹⁶ And, by his own admission, Chairman King was “out of the loop” for key milestones in the 2025 redistricting process.²⁹⁷ Thus, as between Chairman King and Chairman Hunter—the latter of whom was far more intimately involved in the 2025 Map's development and passage—we find Chairman Hunter's statements regarding the purposes underlying the 2025 redistricting much more probative.

ii. Inconsistencies in Chairman King's Testimony

Second, a concerning portion of the hearing evidence was inconsistent with Chairman King's testimony and floor statements.

On direct- and cross-examination, the parties thoroughly explored conversations between Chairman King and Adam Kincaid during the legislative process. Mr. Kincaid was the outside mapmaker who drew nearly all of the 2025 Map.²⁹⁸ Significant aspects of Chairman King's testimony about those conversations were inconsistent with other evidence.

For instance, Chairman King spoke briefly with Mr. Kincaid at the American Legislative Exchange Council (“ALEC”) conference in mid-July 2025.²⁹⁹ As Chairman King tells it, he told Mr. Kincaid that he didn't want to talk about the redistricting maps, because he believed he'd likely be chairing the Senate Redistricting Committee, and he wanted all information about redistricting to come through public channels.³⁰⁰ By contrast, Mr. Kincaid testified that Chairman King openly questioned him about the redistricting efforts during their conversation at ALEC—without ever stating that he'd prefer not to talk about the maps due to his likely future position on the committee.³⁰¹ While Chairman King testified that he never asked how many seats Republicans would potentially gain under the 2025 Map,³⁰² Mr. Kincaid unequivocally testified that Chairman King specifically asked him how many seats Republicans could pick up under the new map, and Mr. Kincaid told him.³⁰³ When counsel confronted Chairman King with that discrepancy at the preliminary-injunction hearing, he conceded that either he was misremembering or Mr. Kincaid's testimony was incorrect.³⁰⁴ That leads us to question whether Chairman King, Mr. Kincaid, or neither one was accurately relaying the substance of their meeting at ALEC—and whether anything happened during that meeting that would betray an unlawful legislative motive.

Chairman King's testimony at the preliminary-injunction hearing was also inconsistent with statements he gave on the Senate floor. He testified that “sometime late in the first-called Special Session”³⁰⁵ —i.e., sometime shortly before August 15, 2025 ³⁰⁶ —he called Mr. Kincaid to ask whether he “use[d] racial data in drawing the map.”³⁰⁷ According to Chairman King, Mr. Kincaid answered that he hadn't used racial data.³⁰⁸

However, on August 22, 2025—shortly after that call allegedly occurred—Senator Roland Gutierrez directly asked Chairman King on the Senate floor if he knew whether the mapdrawer “looked at race in creating the[ ] map.”³⁰⁹ Despite having allegedly called Mr. Kincaid a little over a week earlier to ask him exactly that question, Chairman King told Senator Gutierrez that he didn't know whether the mapdrawer had looked at race.³¹⁰ In fact, Chairman King told Senator Gutierrez during that same exchange that he hadn't even “inquired as to who physically drew the maps.”³¹¹ Yet Chairman King clearly knew Mr. Kincaid had drawn the map, since he had allegedly called Mr. Kincaid just a week or two earlier to ask him whether he had based that map on race. Chairman King's testimony in court thus conflicts with his responses to Senator Gutierrez on the Senate floor—causing us to further question his credibility.³¹²

The record also contains discrepancies regarding:

(1) whether Chairman King's meeting with Mr. Kincaid at ALEC was unplanned or prearranged;313 and

(2) the substance and existence of other communications between Chairman King and Mr. Kincaid during the 2025 redistricting process.314

We might dismiss those inconsistencies as innocuous memory lapses if we considered either one of them in a vacuum. But the number of inconsistencies regarding potentially critical exchanges between the Chair of the Senate Redistricting Committee and the person who drew the 2025 Map makes us doubt the veracity of Chairman King's testimony.

For the foregoing reasons, we do not credit Chairman King's testimony about the Legislature's motives.

b. Senator Hinojosa

We next consider the testimony of Senator Adam Hinojosa. Senator Hinojosa delivered a speech on the Senate floor stating that he was voting for the 2025 Map for partisan rather than racial reasons.³¹⁵ While we have no reason to doubt the truthfulness or sincerity of that speech, we don't think Senator Hinojosa's testimony and contemporaneous legislative statements move the needle. Senator Hinojosa had little involvement in the redistricting process beyond voting for the bill and delivering a brief speech in support.³¹⁶ Thus, Senator Hinojosa's testimony tells us, at most, why one single legislator voted for the 2025 Map. Precedent cautions us not to “overemphasiz[e] statements from individual legislators,”³¹⁷ as “[w]hat motivates one legislator to make a speech about a statute is not necessarily what motivates scores of others to enact it.”³¹⁸ We find the contemporaneous statements of the 2025 Map's sponsors and primary champions more probative of the Legislature's intent.³¹⁹

c. Chairman Vasut

Finally, Chairman Vasut. In contemporaneous statements to the media, Chairman Vasut insisted that the 2025 Map was motivated by partisan rather than racial considerations,³²⁰ and that the DOJ Letter did not influence the Legislature in the redistricting process.³²¹ Chairman Vasut likewise stated in legislative hearings that he wasn't influenced by the Governor's media statements conveying a desire to eliminate coalition districts.³²²

We do not disregard Chairman Vasut's testimony on credibility grounds like Chairman King's. And unlike Senator Hinojosa, Chairman Vasut held a key position in the redistricting process as Chair of the House Select Committee on Congressional Redistricting ³²³ and as one of the House bill's joint authors.³²⁴ Accordingly, we do not dismiss Chairman Vasut's statements as the views of a rank-and-file legislator who wasn't intimately involved in the redistricting process.

On balance, however, the direct evidence of a predominant racial motive outweighs the direct evidence on the other side. The fact that one witness provided testimony that challenges the Plaintiff Groups’ claims doesn't prevent them from meeting their burden at this stage.

We conclude that the contemporaneous statements of legislators involved in the 2025 redistricting are more indicative of racial motives than partisan ones. When we consider that direct evidence with the circumstantial evidence of racial gerrymandering, the totality of the record persuades us that the Plaintiff Groups have shown the requisite likelihood of success on the merits of most of their racial-gerrymandering claims.

d. Adam Kincaid

As previewed above, the person who drew all but a small portion ³²⁵ of the 2025 Map was Mr. Adam Kincaid.³²⁶ Mr. Kincaid wasn't a member of the Legislature; instead, the Republican National Committee hired Mr. Kincaid as an outside mapmaker to draw the State's congressional plan.³²⁷

i. Mr. Kincaid's Testimony

At the preliminary-injunction hearing, Mr. Kincaid testified extensively about his thought process when drafting the 2025 Map.³²⁸ He stated that although he has the ability to display racial demographic data on his mapdrawing software,³²⁹ he did not look at any racial data when drafting the 2025 Map.³³⁰ Mr. Kincaid thus testified unequivocally that he drew the 2025 Map completely blind to race.

Mr. Kincaid testified that he instead based his districting choices entirely on partisan, political, and other race-neutral criteria:

(1) “[E]very Republican incumbent who lived in their seat” under the 2021 Map needed to “stay[ ] in their seat” under the 2025 Map.331

(2) “[E]very Republican incumbent who was in a district that President Trump had won with 60 percent of the vote or more in 2024” needed to “stay[ ] in a district that President Trump won ․ with 60 percent of the vote or more.”332

(3) For incumbent Republican members “who were in districts that President Trump had carried but by less than 60 percent of the vote,” Mr. Kincaid “either had to improve” the Republican performance of those districts “or keep their Partisan Voting Index” (“PVI”) “the same.”333

(4) The map needed to create five new Republican-leaning seats (“pickup opportunities”)334 in which:

(a) President Trump carried the district by at least 10% in the 2024 Presidential Election;335 and

(b) Senator Ted Cruz carried the district in the 2024 U.S. Senate Election.336

(5) It needed to appear likely, based on various predictors, that the map's Republican districts would remain Republican after the 2026 midterms.337

(6) Mr. Kincaid also sought to improve the map's compactness and respect for municipal and geographical boundaries.338

(7) To comply with the constitutionally mandated “one person, one vote” requirement,339 the districts needed to be as close to equipopulous as possible.340

(8) Finally, Mr. Kincaid needed to comply with certain district-specific instructions from the Republican congressional delegation, like keeping certain counties together 341 or keeping district offices within the district.342

Even when Mr. Kincaid opted not to follow certain traditional districting criteria, he did so in a partisan fashion. For example, while Mr. Kincaid prioritized protecting Republican incumbents,³⁴³ he gave no consideration to keeping Democrat incumbents in their districts.³⁴⁴ Mr. Kincaid likewise prioritized core retention in Republican districts but not Democrat districts.³⁴⁵

On the stand, Mr. Kincaid went district by district—sometimes line by line—explaining the logic behind each of the redistricting choices he made.³⁴⁶ Rather than relaying a blow-by-blow recitation of Mr. Kincaid's testimony, we'll simply acknowledge that Mr. Kincaid gave political or practical—i.e., non-racial—rationales for his decisions at every step of the mapdrawing process.³⁴⁷ In Mr. Kincaid's own words, he “drew the map using political data from start to finish.”³⁴⁸

ii. The Court Does Not Credit Mr. Kincaid's Testimony

While Mr. Kincaid's statewide tour of his map was compelling,³⁴⁹ we nonetheless discredit his testimony that he drew the 2025 Map blind to race. We find it extremely unlikely that Mr. Kincaid could have created so many districts that were just barely 50%+ CVAP by pure chance.

The Supreme Court's opinion in Cooper v. Harris illustrates the point.³⁵⁰ As here, lawmakers commissioned an outside (i.e., non-legislator) mapmaker “to assist them in redrawing district lines.”³⁵¹ Like Mr. Kincaid, the outside mapmaker in Cooper claimed that “he displayed only [political] data, and no racial data, on his computer screen while mapping the [challenged] district.”³⁵²

However, the mapmaker achieved an “on-the-nose attainment of a 50% BVAP” in the challenged district ³⁵³ —a feat that, in the district court's view, the mapdrawer would have been unlikely to achieve by blind adherence to partisan data alone.³⁵⁴ The district court deemed it far more likely that the mapdrawer used a 50% racial target to “deliberately redr[a]w [the challenged district] as a majority-minority district.”³⁵⁵ The district court “disbelieved [the mapmaker's] asserted indifference to the new district's racial composition,”³⁵⁶ and the Supreme Court ruled that the district court didn't clearly err by doing so.³⁵⁷

The facts here are even starker. Mr. Kincaid would have us believe that, with racial data on his mapping program turned off, and relying purely on race-neutral criteria like partisan performance, compactness, and incumbency protection (for Republicans), he just coincidentally happened to transform not one, but three, coalition districts into districts that are single-race-majority by half a percent or less:

(1) CD 9 (whose Hispanic CVAP increased from 25.6% to 50.3%);

(2) CD 18 (whose Black CVAP increased from 38.8% to 50.5%); and

(3) CD 30 (whose Black CVAP increased from 46.0% to 50.2%).358

While we acknowledge the possibility that Mr. Kincaid might have done that for one district by pure chance,³⁵⁹ it is very unlikely that he would have hit a barely 50% CVAP three times by pure chance. Mr. Kincaid's “on-the-nose attainment of a 50% [C]VAP” in three districts causes us to doubt his testimony that “he displayed only [partisan] data, and no racial data, on his computer screen while mapping” those districts.³⁶⁰ We find it far more likely that Mr. Kincaid “deliberately redrew [those districts as] majority-minority district[s].”³⁶¹

Mr. Kincaid would also have us believe that it's just a coincidence that the 2025 Map achieves three of the four explicit racial directives outlined in the DOJ Letter:

(1) eliminating CD 9's status as a coalition district;

(2) eliminating CD 18's status as a coalition district; and

(3) radically transforming the racial demographics of CD 29.362

Mr. Kincaid was well aware of the DOJ Letter. He saw a preliminary draft of it in the West Wing of the White House and discussed it with key White House and DOJ officials—and Governor Abbott—a week before DOJ released it.³⁶³

Finally, Mr. Kincaid would have us believe that it's just a coincidence that blindly following the political objectives that Governor Abbott expressly disclaimed happened to achieve the Governor's publicly stated racial goal of creating several new majority-Hispanic districts.³⁶⁴

But, as Chairman Hunter announced on the House floor, “Nothing's a coincidence.”³⁶⁵ It is far more plausible that Mr. Kincaid had both racial and partisan data turned on while drawing the 2025 Map and that he used the former to achieve the racial targets that DOJ and the Governor had explicitly announced as he simultaneously used the latter to achieve his partisan goals.³⁶⁶ Only that would explain how Mr. Kincaid could point to putatively race-neutral criteria to justify his districting decisions at each step of the process while still arriving at such precise racial numbers.

Apart from the 2025 Map's racial numbers, we also reiterate the significant inconsistencies between Mr. Kincaid's testimony and Chairman King's testimony and his contemporaneous statements on the Senate floor.³⁶⁷ Just as those contradictions caused us to question Chairman King's credibility, they lead us to question Mr. Kincaid's veracity as well.

iii. Mr. Kincaid's Professed Lack of Racial Motive Isn't Attributable to the Legislature

Even if Mr. Kincaid just happened to hit those precise racial bullseyes without enabling racial shading in his mapmaking software, Mr. Kincaid's professed lack of racial intent still would not defeat the Plaintiff Groups’ racial-gerrymandering claims. Mr. Kincaid is not a member of the Legislature. The record contains no indication that the Legislature ever told Mr. Kincaid to draw the 2025 Map race-blind; Mr. Kincaid's instructions for how to draw the map came from the White House ³⁶⁸ and the Republican congressional delegation ³⁶⁹ rather than the Legislature or the Governor.³⁷⁰ Just as we can't automatically impute DOJ's or the Governor's racial intent to the Legislature,³⁷¹ we can't automatically impute Mr. Kincaid's alleged lack of racial intent to the Legislature either.³⁷² What ultimately matters is why the Legislature—not Mr. Kincaid—did what it did.

The Fifth Circuit's decision in Prejean v. Foster is illustrative.³⁷³ There—as here—a non-legislator drew an electoral map that the legislature ultimately adopted.³⁷⁴ There, too, the non-legislator mapmaker swore that he drew the map for predominantly political, non-racial reasons.³⁷⁵ The map contained a majority-Black district, which the plaintiffs challenged as a racial gerrymander.³⁷⁶

The Fifth Circuit concluded that the outside mapdrawer's stated “intent in drawing the [map]” could not be “taken as conclusive proof of the legislature's intent.”³⁷⁷ Instead, the Fifth Circuit focused on why the legislature introduced and enacted the map that the mapmaker drew.³⁷⁸ The Court indicated that even if the mapdrawer had truly based the map primarily on political rather than racial considerations, the Legislature's decision to introduce and pass that map for predominantly racial reasons could support a finding of racial gerrymandering.³⁷⁹

The Fifth Circuit then explained that DOJ had been pressuring the state to create a majority-Black subdistrict.³⁸⁰ The outside mapdrawer's plan proposed to do exactly that.³⁸¹ The court reasoned that if the legislature had introduced and passed the mapmaker's plan because “the legislature was ready to adopt whatever proposal would satisfy its objective of creating a black subdistrict,” then that could support a finding of racial gerrymandering ³⁸² —irrespective of the mapmaker's insistence that he based the map predominantly on political and other race-neutral principles.³⁸³

“[C]ontemporaneous statements attributable to the State suggest[ed] that the major purpose of” the enacted plan in Prejean “was to create a majority-minority subdistrict” as DOJ had demanded—not to achieve the mapdrawer's subjective political goals.³⁸⁴ By all objective appearances, “the state was rushing headlong into the arms of DOJ regardless of legal consequences.”³⁸⁵ Perceiving a “disjunction ․ between [the mapmaker's professed] intent and the intent of the legislature,” the Fifth Circuit concluded that the mapmaker's declarations regarding his own thought process when drawing the map were “far from determinative” of “the legislature's intent.”³⁸⁶

While we readily acknowledge factual and procedural distinctions between this case and Prejean,³⁸⁷ Prejean stands for the principle that when an outside mapdrawer professes to have drawn a redistricting plan based on political rather than racial criteria, courts should not automatically impute the mapdrawer's lack of racial intent to the legislature.³⁸⁸ The court should instead inquire why the legislature introduced and passed the map that the mapmaker drew. If other evidence in the record indicates that the legislature adopted the mapmaker's purportedly race-blind map because it happened to achieve some racial objective—such as creating a new single-race-majority district at DOJ's behest—that can potentially support a finding that race was the legislature's predominant motivation.³⁸⁹

Even if we credited Mr. Kincaid's testimony that he drew the 2025 Map completely blind to race, the fact remains that the map he gave to the Legislature proposed to eliminate numerous coalition districts and replace them with single-race-majority districts. Mr. Kincaid gave the Legislature a map that achieved DOJ's and the Governor's objectives, while enabling the Legislature to portray the map as being more favorable to minority voters than its 2021 predecessor. If the reason why the Legislature introduced and enacted that map is because it just happened to achieve those objectives, then Mr. Kincaid's subjective lack of racial motivation is irrelevant.

“[C]ontemporaneous statements attributable to the State” and other direct and circumstantial evidence “suggest that the major purpose of” the 2025 Plan “was to create [more] majority-minority [districts].”³⁹⁰ Mr. Kincaid's professed lack of racial intent is therefore “far from determinative” of “the legislature's [own] intent.”³⁹¹ The “disjunction ․ between” Mr. Kincaid's stated intent and the apparent “intent of the legislature” leads us to conclude that Mr. Kincaid's testimony does not preclude the Plaintiff Groups from obtaining a preliminary injunction.³⁹²

5. Circumstantial Evidence of Legislative Intent

Having canvassed the available direct evidence, we now discuss the circumstantial evidence.

a. The 2025 Map Achieved DOJ's and the Governor's Goals

First, the fact that the Legislature fulfilled almost everything that DOJ and the Governor desired supports the notion “that a majority of the [Legislature's] members shared and purposefully adopted (i.e., ratified) the [Governor and DOJ's racial] motivations.”³⁹³ It further suggests that the Legislature “was rushing headlong into the arms of DOJ regardless of legal consequences.”³⁹⁴

b. The Sheer Number of Just-Barely-50%-CVAP Districts Suggests that the Legislature Set and Followed a Racial Target

The 2025 Map's “on-the-nose attainment of a 50% [C]VAP” for so many districts ³⁹⁵ further suggests that the Legislature was following a “50%-plus racial target” “to the letter,” such that the “racial target had a direct and significant impact on [those districts’] configuration[s].”³⁹⁶ This fact is as much circumstantial evidence as it is direct.

c. The Legislature Left a Majority-White Democrat District Largely Unchanged

If the Legislature's aims were exclusively partisan rather than predominantly racial, it is reasonable to assume the Legislature would have also reconfigured single-race-majority Democrat districts to make them Republican. In particular, we'd expect the Legislature to also make significant modifications to CD 37, a majority-White district ³⁹⁷ that generally elected Democrats.³⁹⁸ Yet CD 37 remains a Democrat district under the 2025 Map.³⁹⁹ It also remains majority-White.⁴⁰⁰

That stands in stark contrast to how the Legislature treated majority-non-White districts with the same partisan attributes as CD 37. To illustrate, here is the most telling example. Whereas 67.8% of the 2021 configuration of majority-White CD 37 remains intact in 2025 Map,⁴⁰¹ only 2.9% of majority-non-White CD 9 remains intact in the new map.⁴⁰² The fact that the Legislature completely gutted majority-non-White CD 9 and not majority-White CD 37—even though the two districts had the same political lean—constitutes additional circumstantial evidence that the Legislature's predominant consideration was race rather than partisanship.⁴⁰³

d. The Legislature Transformed a Republican Coalition District into a Republican Majority-White District

Coming at it from the opposite angle, if the Legislature's aims were partisan rather than racial, one would expect the Legislature not to make fundamental changes to the racial demographics of Republican districts, as doing so would net no gain in the number of Republican seats. Yet the 2025 Map takes an existing majority-non-White Republican district (CD 27) and decreases the Hispanic CVAP from 48.6% to 36.8%, while raising the White CVAP from 44.1% to 52.8% to make it majority-White.⁴⁰⁴

e. The Testimony of Dr. Moon Duchin

Finally, the expert report and testimony of Dr. Moon Duchin (Professor of Data Science, University of Chicago) supplies additional circumstantial evidence that race, not politics, best explains the 2025 Map's contours.

i. Dr. Duchin's Methodology

Dr. Duchin is one of the pioneers of a technique for assessing whether an electoral map is more consistent with race-based decision-making than with race-neutral criteria, such as partisanship and traditional districting considerations.⁴⁰⁵ Using a computer program, Dr. Duchin randomly generates hundreds of thousands of congressional maps that the Legislature might have hypothetically drawn.⁴⁰⁶ The program is coded to generate maps that a Republican-controlled Legislature might have realistically enacted. The maps favor Republicans by various metrics,⁴⁰⁷ and they obey (or at least favor) traditional districting criteria like contiguity, compactness, respect for municipal subdivisions, and core retention.⁴⁰⁸

After generating those hundreds of thousands of maps, the program “winnows” the maps down according to political criteria like Republican performance and incumbency protection.⁴⁰⁹ That winnowing process yields approximately 40,000 hypothetical maps that the Republican-controlled Legislature could have conceivably passed.⁴¹⁰

None of the programmed criteria for generating or filtering the maps is race-based; they are all race-neutral.⁴¹¹ The program thus generates an enormous number of maps that the Legislature might have drawn if—as the State Defendants assert here ⁴¹² —the Legislature had truly based its redistricting decisions exclusively on race-neutral considerations like partisanship and traditional districting criteria.

Dr. Duchin's program then compares the racial demographics of the enacted map to those of the hypothetical race-neutral maps.⁴¹³ The idea is that, if the Legislature had truly drawn the 2025 Map based solely on race-neutral criteria, then the enacted map's racial characteristics would likely fall somewhere within the expected range of the maps generated by the program.⁴¹⁴ By contrast, if the enacted map's racial characteristics fall outside the demographic ranges of the randomly generated maps, then the enacted map is a statistical outlier.⁴¹⁵ This finding would suggest that the Legislature was predominantly motivated by race rather than partisanship.⁴¹⁶ This technique provides a mathematical method to “disentangle partisanship and race”⁴¹⁷ —just as the Supreme Court has instructed courts and litigants to do in racial-gerrymandering cases.⁴¹⁸

To visually depict the distribution of the randomly generated maps’ racial characteristics, Dr. Duchin's expert report displays her results in the form of “box-and-whiskers” or “box” plots,⁴¹⁹ which look like this:

[Editor's Note: The preceding image contains the reference for footnote ⁴²⁰ ].

The y-axis represents the minority population of each district in each randomly generated map, with the dotted line showing the 50% mark.⁴²¹ The x-axis arranges the districts in each randomly generated map from lowest to highest by share of minority population.⁴²²

The orange figures—which are the ones we're most interested in for our purposes ⁴²³ —represent the range of minority populations for each district in each randomly generated map.⁴²⁴ The “whiskers” (the T-shaped appendages on each end) measure from the 1st percentile to the 99th percentile.⁴²⁵ Taking the orange figure on the far left as an example, in nearly all of Dr. Duchin's randomly generated maps, the district with the lowest minority population in the Dallas/Fort Worth area had a minority population percentage somewhere between 26% and 41%:⁴²⁶

The edges of the “boxes,” meanwhile, measure “from the 25th to the 75th percentile[,] [m]eaning that 50 percent of the plans fall in the box.”⁴²⁷ So, in about half of Dr. Duchin's randomly generated maps, the district with the lowest minority population in the Dallas/Fort Worth area had a minority voter percentage between roughly 34% and 37%:⁴²⁸

The line in the box marks “the median or 50th percentile”:⁴²⁹

The blue dots, meanwhile, represent the minority population of each district in the enacted map.⁴³⁰ For instance, the minority population of the lowest-minority-percentage district in the Dallas/Fort Worth area in the enacted map is around 30%:

The “box-and-whiskers” plot is a pictorial method for comparing the enacted map's racial demographics to those of race-neutral hypothetical maps. If any particular dot falls within the same range as the “box,” the enacted district's minority population is within the range we'd expect if the Legislature were relying exclusively on partisanship and other race-neutral districting criteria. If a dot falls outside the box but within the “whiskers,” the enacted district's minority population is on the outer edge of what we'd expect if the Legislature were relying exclusively on partisanship and other race-neutral considerations. If the dot falls outside the whiskers entirely, none of the race-neutral maps that Dr. Duchin generated has the racial characteristics approximating that of the enacted district—and, thus, the enacted map is statistically anomalous.⁴³¹ These results would in turn suggest that race—not partisanship—is the variable that best explains the enacted map's configuration.⁴³²

ii. Dr. Duchin's Findings and Conclusions

Dr. Duchin applied that technique to the Houston area,⁴³³ where three of the four districts mentioned in the DOJ Letter are located (CDs 9, 18, and 29).⁴³⁴ The results are jarring:

[Editor's Note: The preceding image contains the reference for footnote ⁴³⁵ ].

Five of the dots fall outside of the whiskers—some by a sizable amount—while only one dot falls within its respective box. Four of the ten districts in the Houston area “have outlying low levels of minority citizens” under the enacted map, “while one district far above 50% is elevated to an outlying degree.”⁴³⁶ These results suggest that a Legislature motivated exclusively by partisan and other race-neutral concerns would be unlikely to produce a configuration of the Houston-area districts with racial characteristics similar to the 2025 Map.⁴³⁷ This evidence supports the notion that the Legislature purposefully manipulated the racial statistics of Houston-area districts like CDs 9, 18, and 29 at DOJ's behest.

While the patterns in the Dallas/Fort Worth area (where CDs 30, 32, and 33 are located) are less visibly stark than those in the Houston area, and those in the Travis/Bexar County area (where CDs 27 and 35 are located) are even less so, they nonetheless reinforce the conclusion that the enacted map's racial composition is a statistical outlier:

[Editor's Note: The preceding image contains the reference for footnote ⁴³⁸ ].

[Editor's Note: The preceding image contains the reference for footnote ⁴³⁹ ].

In the Dallas/Fort Worth area, one of the dots falls outside the whiskers entirely, while two dots fall precisely on a whisker's edge.⁴⁴⁰ Though all of the districts in the Travis/Bexar County area fall within the whiskers, there are three dots that are a comfortable distance away from their respective boxes.⁴⁴¹

According to Dr. Duchin's analysis, it is highly unlikely that a Legislature drawing a map based purely on partisan and other race-neutral considerations would have drawn a map with the 2025 Map's racial characteristics.⁴⁴² In other words, the best possible explanation for the 2025 Map's racial makeup is that the Legislature based the 2025 Map on racial considerations, and those racial considerations predominated over partisan ones.⁴⁴³

Dr. Duchin's results are fully consistent with the direct evidence and other circumstantial evidence in the record. Even more notably, Dr. Duchin's testimony was effectively unchallenged; no defense expert submitted a report rebutting Dr. Duchin's findings.⁴⁴⁴ For all those reasons, we find Dr. Duchin's testimony and report highly credible and persuasive.

iii. The State Defendants’ Critiques

The State Defendants—though none of their experts—attack Dr. Duchin's methods and conclusions on several fronts. They first note that in a different case in which Dr. Duchin served as an expert, Alexander v. South Carolina State Conference of the NAACP, the Supreme Court determined that Dr. Duchin's analysis suffered from “serious problems,” and thus had “no probative force with respect to [the plaintiffs’] racial-gerrymandering claim.”⁴⁴⁵

Dr. Duchin's report here doesn't suffer from the same defects that led the Alexander Court to reject her findings. For example, the Supreme Court discredited Dr. Duchin's report in Alexander because “various parts of [her] report did not account for partisanship or core retention.”⁴⁴⁶ Here, Dr. Duchin's report explicitly took both of those variables into consideration.⁴⁴⁷ The Alexander Court also discredited Dr. Duchin because her conclusions were “based on an assessment of the map as a whole rather than [the challenged district] in particular.”⁴⁴⁸ Here, instead of examining the State of Texas as a whole, Dr. Duchin focused exclusively on three geographic clusters containing only the challenged districts and their adjacent neighbors.⁴⁴⁹ Therefore, the issues that caused the Supreme Court to discredit Dr. Duchin's conclusions in Alexander don't lead us to do the same here.

The State Defendants also attack the criteria that Dr. Duchin used to generate and winnow her numerous hypothetical maps. To ensure that Dr. Duchin's computer-generated maps resemble plans that the Legislature might realistically have enacted, the program's variables must resemble the race-neutral partisan and political parameters that the Legislature purported to follow when drawing and enacting the actual map.⁴⁵⁰ In other words, if you don't tell the computer to follow the same race-neutral criteria that the Legislature purported to follow, then the maps it generates won't tell you anything reliable about whether the enacted map is an outlier. The State Defendants argue that Dr. Duchin didn't program her computer to follow the same partisan and political criteria that the Legislature followed—and, consequently, that her maps aren't appropriate comparators.

For example, the State Defendants claim that Dr. Duchin set her partisanship thresholds too low.⁴⁵¹ As one of her winnowing conditions, Dr. Duchin culled the randomly generated maps to only include plans in which “at least as many districts ha[d] a plurality win for Donald Trump from the 2024 election as in” the enacted map.⁴⁵² As a robustness check, Dr. Duchin then “executed a run seeking to match the number of districts with Trump's 2024 major-party vote share over 55%,”⁴⁵³ and achieved results consistent with her prior findings.⁴⁵⁴ The State Defendants argue that Dr. Duchin needed to set those thresholds higher to emulate the Republican performance of the 2025 Map,⁴⁵⁵ since “President Trump carried many of the disputed districts with nearly 60% of the vote in 2024.”⁴⁵⁶

We're not convinced that Dr. Duchin's 55% Trump threshold caused her to generate maps that deviated materially from the enacted one. While the State Defendants are correct that some of the challenged districts in the enacted map have Trump numbers that equal or approach 60%,⁴⁵⁷ there are also districts that fall short of 60%,⁴⁵⁸ including multiple districts hovering right around Dr. Duchin's 55% threshold.⁴⁵⁹ Additionally, Dr. Duchin's 55% threshold was a floor rather than a ceiling—meaning that it would capture districts with Trump percentages closer to 60% like those in the enacted map.⁴⁶⁰ The State Defendants have therefore failed to persuade us that Dr. Duchin's 55% figure is disqualifying.

In any event, if raising the floor to a value closer to 60% would have undermined Dr. Duchin's conclusions, the State Defendants could have introduced expert rebuttal testimony to that effect. Again, though, the State Defendants let Dr. Duchin's testimony go unrebutted.⁴⁶¹ The State Defendants have therefore given us no concrete reason to think that Dr. Duchin's results would have looked significantly different had she selected different partisanship thresholds.

The same goes for the State Defendants’ arguments that Dr. Duchin:

(1) should have programmed her computer to favor only core retention and incumbency protection in Republican districts (like Mr. Kincaid did);462 and

(2) used an out-of-date list of incumbent addresses.463

Absent any rebuttal expert testimony that programming the computer to address those critiques would have significantly changed Dr. Duchin's results, we have no basis to dismiss her testimony as unreliable. And the record shows Dr. Duchin made a good-faith effort to update incumbent addresses for her preliminary-injunction report but was unable to do so for reasons outside of her control.⁴⁶⁴

In sum, Dr. Duchin generated tens of thousands of congressional maps that follow traditional districting criteria and favor Republicans by various metrics, and not one of them had racial demographics that looked anything like those in the 2025 Map.⁴⁶⁵ That is entirely consistent with the rest of the direct and circumstantial evidence. The 2025 Map's racial characteristics did not result from the blind pursuit of partisan gain, but from the intentional manipulation of the districts’ racial makeup.⁴⁶⁶

6. Contrary Circumstantial Evidence

A few brief notes about circumstantial evidence that points in the opposite direction:

a. CD 33 Remains a Coalition District

Although the DOJ Letter instructs Texas to eliminate CD 33's status as a coalition district, CD 33 remains a coalition district under the 2025 Map.⁴⁶⁷ At least for CD 33, neither the DOJ Letter nor racial considerations more generally were the primary factor motivating the Legislature's reconfiguration of the district. Therefore, the Plaintiff Groups have not demonstrated a likelihood of success on the merits of their racial-gerrymandering challenge to CD 33.

That finding does not undermine our conclusion that the Plaintiff Groups have demonstrated a likelihood of success on the merits of most of their other racial-gerrymandering claims. Because “[r]acial gerrymandering claims proceed district-by-district,”⁴⁶⁸ it's entirely possible for the Legislature to gerrymander one district without gerrymandering another. CD 33 is the lone exception to the Legislature's general pattern of converting as many coalition districts to single-race-majority districts as possible.

b. The 2025 Map Comports with Traditional Districting Principles

As stated above, a plaintiff asserting a racial-gerrymandering claim bears the burden to “prove that the State subordinated race-neutral districting criteria such as compactness, contiguity, and core preservation to racial considerations.”⁴⁶⁹ To make that showing, plaintiffs “often need to show that the State's chosen map conflicts with” those “traditional redistricting criteria.”⁴⁷⁰ “That is because it may otherwise be difficult for challengers to find other evidence sufficient to show that race was the overriding factor causing neutral considerations to be cast aside.”⁴⁷¹

By some measures, the 2025 Map is more consistent with traditional districting criteria than its predecessors. For instance, the 2025 Map scores better on certain compactness measurements ⁴⁷² and core-retention metrics ⁴⁷³ than the 2021 Map.

That hurdle is not dispositive here. Even though plaintiffs “often need to show that the State's chosen map conflicts with traditional redistricting criteria” to prevail on a racial-gerrymandering claim,⁴⁷⁴ “a conflict or inconsistency between the enacted plan and traditional redistricting criteria is not a threshold requirement or a mandatory precondition in order for a challenger to establish a claim of racial gerrymandering.”⁴⁷⁵ “Race may predominate”—“even when a reapportionment plan respects traditional [districting] principles”—if:

(1) “race was the criterion that, in the State's view, could not be compromised,” and

(2) “race-neutral considerations came into play only after the race-based decision had been made.”476

“[T]here may be cases where challengers will be able to establish racial predominance”—even “in the absence of an actual conflict” between the enacted map and traditional districting principles— “by presenting direct evidence of the legislative purpose and intent or other compelling circumstantial evidence.”⁴⁷⁷

The Plaintiff Groups have introduced direct and circumstantial evidence that race was the criterion that could not be compromised in the 2025 redistricting ⁴⁷⁸ and that racial considerations predominated over political ones.⁴⁷⁹ Therefore, the fact that the 2025 Map generally complies with traditional districting criteria isn't fatal.

7. The Plaintiff Groups’ Failure to Produce an Alexander Map

Finally, we address whether the Plaintiff Groups needed to present a so-called “Alexander map” to obtain a preliminary injunction. An “often highly persuasive way to disprove a State's contention that politics drove a district's lines” is for the plaintiff to introduce “an alternative map that achieves the legislature's political objectives while improving racial balance.”⁴⁸⁰ Such a map “show[s] that the legislature had the capacity to accomplish all its partisan goals without moving so many members of a minority group” between electoral districts.⁴⁸¹ The idea is that if the Legislature was “really sorting by political behavior instead of skin color,” it “would have done—or, at least, could just as well have done—this.”⁴⁸² “Such would-have, could-have, and (to round out the set) should-have arguments are a familiar means of undermining a claim that an action was based on a permissible, rather than a prohibited, ground.”⁴⁸³

In Alexander v. South Carolina State Conference of NAACP, the Supreme Court ruled that, “[w]ithout an alternative map” of the sort described above, “it is difficult for plaintiffs to defeat [the] starting presumption that the legislature acted in good faith.”⁴⁸⁴ The Alexander Court further remarked that such alternative maps are not “difficult to produce”; “[a]ny expert armed with a computer can easily churn out redistricting maps that control for any number of specified criteria, including prior voting patterns and political party registration.”⁴⁸⁵ The Court thus held that “[t]he evidentiary force of an alternative map, coupled with its easy availability, means that trial courts should draw an adverse inference from a plaintiff's failure to submit one.”⁴⁸⁶ The Supreme Court further opined that this “adverse inference may be dispositive in many, if not most, cases where the plaintiff lacks direct evidence or some extraordinarily powerful circumstantial evidence.”⁴⁸⁷

At this early phase of the proceedings, the Plaintiff Groups have not submitted an Alexander map.⁴⁸⁸ For the following reasons, that is not fatal.

For one thing, Alexander states that “[t]he adverse inference may be dispositive in many, if not most, cases where the plaintiff lacks direct evidence” of the legislature's intent.⁴⁸⁹ Unlike the challengers in Alexander, who “provided no direct evidence of a racial gerrymander,”⁴⁹⁰ the Plaintiff Groups here have produced substantial direct evidence indicating that race was the predominant driver in the 2025 redistricting process.⁴⁹¹ This case is not the sort of “circumstantial-evidence-only case” in which Alexander’s adverse inference is typically dispositive.⁴⁹²

Moreover, it's not even clear that Alexander requires us to draw an adverse inference against the Plaintiff Groups at this early phase of the case. The logic behind Alexander’s adverse inference is that, because an alternative map is relatively easy to generate as a technical matter,⁴⁹³ if a plaintiff fails to present such a map at trial, it must be because it's impossible to draw a map that achieves the legislature's partisan goals “while producing significantly greater racial balance.”⁴⁹⁴

But unlike Alexander, which reached the Supreme Court at the permanent injunction stage,⁴⁹⁵ after the district court had conducted a full-fledged trial,⁴⁹⁶ this case is still at the preliminary injunction phase. It's one thing to draw an adverse inference if a plaintiff fails to produce a suitable Alexander map after preparing for a trial for a year or more; it's quite another if a plaintiff fails to produce a suitable Alexander map at an accelerated, preliminary phase of the litigation. For that reason, at least one lower court has ruled that Alexander’s alternative map requirement does not apply at a redistricting case's preliminary phases.⁴⁹⁷ It would be improper here to infer that the reason the Plaintiff Groups didn't produce an Alexander map at the preliminary-injunction hearing is because it's impossible to create one. The most likely reason is that they simply didn't have time.⁴⁹⁸

If anything, the preliminary-injunction record suggests that the Plaintiff Groups will be able to present an acceptable Alexander map at trial. Although the Plaintiff Groups didn't offer any of Dr. Duchin's randomly generated maps as an Alexander map at the preliminary-injunction hearing,⁴⁹⁹ the fact that she generated tens of thousands of pro-Republican maps that obey traditional redistricting principles without producing the enacted map's exaggerated racial features makes us confident that the Plaintiff Groups will be able to produce a suitable Alexander map once the Court ultimately tries this case on the merits.⁵⁰⁰

Thus, while Alexander will be a hurdle that the Plaintiff Groups will need to surmount at trial, it does not bar the Plaintiff Groups from obtaining a preliminary injunction here.

8. Texas's Use of Race When Drawing the 2025 Map Wasn't Narrowly Tailored to Achieve a Compelling Interest

We've thus determined that, at trial, the Plaintiff Groups will likely satisfy their initial burden to show that race predominated over partisanship for many of the districts they challenge. Assuming they do so, the burden will then shift to the State Defendants ⁵⁰¹ “to prove that its race-based sorting of voters serves a ‘compelling interest’ and is ‘narrowly tailored’ ” to that end.”⁵⁰²

Because the State Defendants’ theory of the case is that the Legislature didn't base the 2025 Map on race at all,⁵⁰³ they make no serious effort to argue that the Legislature's use of race was narrowly tailored to achieve a compelling interest.⁵⁰⁴ For that reason alone, we could rule against the State Defendants on this issue at this stage of the proceedings.

It's nevertheless prudent to consider whether DOJ's claim—that Texas needed to systematically eliminate coalition districts to break from its supposed “racially based gerrymandering past”⁵⁰⁵ —constitutes a compelling interest to support race-based redistricting here. “There is a significant state interest in eradicating the effects of past racial discrimination.”⁵⁰⁶ “When a state governmental entity seeks to justify race-based remedies to cure the effects of past discrimination,” however, courts “do not accept the government's mere assertion that the remedial action is required.”⁵⁰⁷ Instead, courts “insist on a strong basis in evidence of the harm being remedied.”⁵⁰⁸

As discussed, the evidence in the preliminary-injunction record suggests that the 2021 Legislature didn't discriminate in favor of minority coalitions—whether to comply with Campos or for any other purpose.⁵⁰⁹ Again, as far as the preliminary-injunction record reveals, the 2021 Legislature drew the 2021 Map based strictly on race-neutral criteria like partisanship.⁵¹⁰ By all current appearances, there was no past discrimination in favor of minority coalitions for the State to remedy—and, therefore, no “strong basis in evidence” to support the State's purposeful and predominant consideration of race in the 2025 redistricting process.

Besides remedying past discrimination, the Supreme Court has also “long assumed that complying with the VRA is a compelling interest.”⁵¹¹ The DOJ Letter appears to take the position that, post-Petteway, coalition districts violate the VRA.⁵¹² Therefore, we consider whether we can excuse the State's race-based redistricting as a well-intentioned but misguided attempt to comply with the VRA.

We can't. “Although States enjoy leeway to take race-based actions reasonably judged necessary under a proper interpretation of the VRA,”⁵¹³ courts cannot “approve a racial gerrymander ․ whose raison d’être is a legal mistake.”⁵¹⁴ As this opinion makes clear, the DOJ's interpretation of Petteway—that VRA § 2 and the Constitution render coalition districts per se unlawful—is obviously wrong.⁵¹⁵ Thus, the State's systematic, purposeful elimination of coalition districts and creation of new single-race-majority districts “was not reasonably necessary under a constitutional reading and application of [the VRA].”⁵¹⁶

Nor, if the State were so inclined, could it avoid liability by arguing that it was just following orders from DOJ. “[T]he Justice Department's objection” to a state's map is not “itself ․ a compelling interest adequate to insulate racial districting from constitutional review.”⁵¹⁷

We therefore conclude that, once this case proceeds to trial, the State Defendants will be unlikely to carry their burden to show that the Legislature's use of race was narrowly tailored to achieve a compelling interest. The Plaintiff Groups have therefore shown that they're likely to succeed on their racial-gerrymandering challenges to CDs 9, 18, 27, 30, 32, and 35.

C. Irreparable Harm

Besides showing that they're likely to succeed on the merits, the Plaintiff Groups have also established that they are “likely to suffer irreparable harm in the absence of preliminary relief.”⁵¹⁸ “In general, a harm is irreparable where there is no adequate remedy at law, such as monetary damages.”⁵¹⁹ Here, the Plaintiff Groups’ alleged harm is the violation of their constitutional rights under the Fourteenth and Fifteenth Amendments.⁵²⁰ “[T]he loss of constitutional freedoms,” such as the right to equal protection of the law and to exercise the right to vote free from racial discrimination, “for even minimal periods of time unquestionably constitutes irreparable injury.”⁵²¹ The inability to vote for and to elect a congressional representative under a constitutional map is undoubtedly “an injury that cannot be compensated with damages, making it irreparable.”⁵²² No legal remedy, including monetary damages, can make up for losing a constitutional right.

The State Defendants do not dispute that a violation of a constitutional right is an irreparable harm.⁵²³ Rather, the State Defendants argue that since the Plaintiff Groups are unlikely to succeed on the merits of their claims, the Plaintiff Groups also cannot show that they are likely to suffer irreparable harm.⁵²⁴ Since the Court finds otherwise, the State Defendants’ arguments fail.

Accordingly, the Court finds the Plaintiff Groups will suffer irreparable harm if the 2025 Map remains Texas's operative congressional map.

D. Balance of Equities and the Public Interest

The Court next addresses the remaining two factors necessary for imposing a preliminary injunction: (1) the balance of equities must favor the movant and (2) an injunction would not disserve the public interest.⁵²⁵ The Plaintiff Groups have satisfied both factors.

The balance of equities addresses “the relative harm to both parties if the injunction is granted or denied.”⁵²⁶ “The public-interest factor looks to the public consequences of employing the extraordinary remedy of injunction.”⁵²⁷ Because these two factors “overlap considerably,” federal courts routinely consider them together.⁵²⁸ Indeed, “[t]hese factors merge when the Government is the opposing party.”⁵²⁹ This is because “[w]hen a statute is enjoined, the State necessarily suffers the irreparable harm of denying the public interest in the enforcement of its laws, and the State's interest and harm thus merge with that of the public.”⁵³⁰ Accordingly, the Court considers both factors together.

1. Purcell Does Not Require the Court to Deny a Preliminary Injunction in This Case

The State Defendants argue that these factors weigh strongly against an injunction based on Purcell v. Gonzalez, 549 U.S. 1, 4–5, 127 S.Ct. 5, 166 L.Ed.2d 1 (2006) (per curiam).⁵³¹ Purcell stands for the principles “(i) that federal district courts ordinarily should not enjoin state election laws in the period close to an election, and (ii) that federal appellate courts should stay injunctions when ․ lower federal courts contravene that principle.”⁵³² These principles “require[ ] courts to consider the effect of late-breaking judicial intervention on voter confusion and election participation.”⁵³³

“[T]he Supreme Court has never specified precisely what it means to be ‘on the eve of an election’ for Purcell purposes.”⁵³⁴ Instead, courts have applied Purcell as “a consideration, not a prohibition,” based on a variety of factors and pre-election and election deadlines.⁵³⁵ Applying the same analysis to this case, the Court finds that Purcell does not require us to deny a preliminary injunction.⁵³⁶

Two Supreme Court applications of Purcell are especially relevant here.⁵³⁷ First is the Robinson line of cases. The Court will not belabor here these cases’ complex development.⁵³⁸ For this opinion's purposes, what matters is that the three-judge panel in Callais enjoined Louisiana's newly drawn congressional plan 189 days (about six months) before the November 5, 2024, general election.⁵³⁹ On May 15, 2024, the Supreme Court stayed the injunction on Purcell grounds.⁵⁴⁰ The Supreme Court's stay order included only a naked citation to Purcell without any accompanying reasoning or analysis about why Purcell compelled the stay.⁵⁴¹

Then there is Merrill v. Milligan.⁵⁴² In that case, the three-judge panel issued its preliminary injunction on January 24, 2022.⁵⁴³ The panel declined to stay the injunction on Purcell grounds because “the primary election [would not] occur [until] May 24, 2022, approximately four months from” the panel's preliminary-injunction order.⁵⁴⁴ The Supreme Court disagreed and stayed the injunction.⁵⁴⁵ Here again, the Supreme Court provided no reasoning for the stay.⁵⁴⁶ In fact, the Supreme Court did not cite to a single case to support its stay—not even to Purcell.⁵⁴⁷ The only reasoning offered to support the stay was in Justice Kavanaugh's concurrence discussing Purcell, which Justice Alito joined.⁵⁴⁸

In his concurrence in Petteway v. Galveston County, Judge Oldham cited to the Supreme Court's stay order in Milligan to observe that “the Supreme Court ․ refused to bless judicial intervention in State elections ․ 120 days before the primary election date” in that case.⁵⁴⁹ In addition to noting the Supreme Court's stay in Milligan, Judge Oldham noted the Fifth Circuit's own calendar constraints. The Fifth Circuit had already taken the case en banc, and the court's next en banc sitting was not until January 23–25, 2024, less than two months before the primary election.⁵⁵⁰ But unlike in Petteway, allowing time for intermediate appellate review of this opinion is not a complicating factor.

The State Defendants argue that these cases preclude the Plaintiff Groups from obtaining injunctive relief here.⁵⁵¹ Texas's congressional primary election is March 3, 2026, about four months from now.⁵⁵² If the Court were to apply Robinson’s timeframe to the next scheduled election, then the window to issue a preliminary injunction in this case before the March 3 primary election closed on August 26, 2025—three days before Governor Abbott even signed the redistricting bill into law.⁵⁵³ Similarly, under Milligan, if 120 days from the primary election is the cutoff, then the panel would have had only until November 3, 2025, to draft this opinion. If the Court applied these timeframes even further under Purcell precedent and considered the next scheduled election to begin when absentee ballots are issued for the primary election, those cut-off deadlines would be even earlier: July 12, 2025, under Robinson and September 19, 2025, under Milligan.⁵⁵⁴

We disagree with the State Defendants. Robinson and Milligan are not dispositive. “Purcell is [not] just a tallying exercise”⁵⁵⁵ or a “magic wand that bars [c]ourts from issuing injunctions some amount of time out from an election.”⁵⁵⁶ That is for good reason. If it were, the Purcell principle would effectively be “absolute”—and it is not.⁵⁵⁷ It is not the case “that a district court may never enjoin a State's election laws in the period close to an election.”⁵⁵⁸ Purcell “simply heightens the showing necessary for a plaintiff to overcome the State's extraordinarily strong interest in avoiding late, judicially imposed changes to its election laws and procedures.”⁵⁵⁹ Rather than setting a hard cut-off, Purcell sets a flexible standard based on a fact-intensive analysis that considers the disruption an injunction would cause.⁵⁶⁰ It's not just about counting the number of days until the next election.

An injunction in this case would not cause significant disruption. The Legislature passed the 2025 Map in August 2025, more than a year before the general election in November 2026. As of this writing, we are still one year out from the general election and four months out from the primary election. Even “critical deadlines that arise before election day itself,” like overseas and absentee primary voting, are more than two months away.⁵⁶¹ And the candidate-filing period remains open for several weeks.

Based on the credible testimony of Christina Adkins, the director of elections for the Texas Secretary of State, some preliminary election preparations have begun. The State has begun educating county election officials, including holding trainings about the 2025 Map, and some counties have started drawing county election voter registration precincts based on this map.⁵⁶² Candidates have also started relying on the 2025 Map, including determining which district to run in, collecting signatures, and campaigning.⁵⁶³ The Court also recognizes there is a trickle-down effect among elections because a candidate's decision to run for Congress means that candidate cannot run for another elected position.⁵⁶⁴ Candidates may make different choices under different congressional maps.

Yet in several critical respects, the State is still operating under the 2021 Map. The State's counties used the precinct boundaries under the 2021 Map for the November 4, 2025, election, and the State used the 2021 Map's lines for the special election in CD 18 on November 4, in addition to having used the 2021 Map for all congressional districts in the 2022 and 2024 elections.⁵⁶⁵ The special election in CD 18 is now proceeding to a runoff election under the 2021 Map on January 31, 2026.⁵⁶⁶ This means the runoff election for CD 18 under the 2021 Map will occur almost two months after the candidate-filing deadline for the November 3, 2026, election, two weeks after the overseas and absentee 2026 primary ballots are mailed, and mere weeks before the 2026 primary election—all of which is set to take place under the 2025 Map.⁵⁶⁷ This runoff also means that Harris County, the State's largest, will retain both its voter precinct boundaries and its district boundaries under the 2021 Map until after CD 18's special election has formally concluded.⁵⁶⁸

So, it is not the case that the entire State has been operating under the 2025 Map for months. The map wasn't even law three months ago, and Texas voters will continue to vote under the 2021 Map after several key pre-election deadlines for the 2025 Map have already passed. Although the filing period for precinct chairs opened in September 2025, its December 8, 2025, closing date will accommodate any changes to precinct filings that result from an injunction.⁵⁶⁹ And the Court is issuing its ruling well before the candidate-filing deadline of December 8. Simply put, the 2026 congressional election is not underway.⁵⁷⁰

In any event, any disruption that would happen here is attributable to the Legislature, not the Court.⁵⁷¹ The Legislature—not the Court—set the timetable for this injunction. The Legislature—not the Court—redrew Texas's congressional map weeks before precinct-chair and candidate-filing periods opened. The State chose to “toy with its election laws close to” the 2026 congressional election, though that is certainly its prerogative.⁵⁷² But any argument that this Court is choosing “to swoop in and re-do a State's election laws in the period close to an election” is wholly misdirected.⁵⁷³ In this case, “[l]ate judicial tinkering” with Texas's congressional map is not what could “lead to disruption and to unanticipated and unfair consequences for candidates, political parties, and voters.”⁵⁷⁴ The Legislature—not the Court—opened that door.⁵⁷⁵ No one disputes the fact that “state and local election officials need substantial time to plan for elections.”⁵⁷⁶ But for Purcell purposes, that fact became moot when the Legislature enacted a new congressional map days before the precinct chair filing period opened and two months before the candidate filing period opened. As between the Plaintiff Groups, who have a constitutional right to vote under a lawful map, and the State, who invited this issue by enacting a new map within Purcell’s range, the equities favor the Plaintiff Groups.

This finding is bolstered by the fact that the parties’ swift action has mitigated to the greatest extent possible the risk of “significant logistical challenges” for Texas election officials and of voter confusion.⁵⁷⁷ Unlike in other cases where the district court's injunction “would require heroic efforts by [ ] state and local authorities,” in this case the Legislature's decision to enact a new congressional map has required “heroic efforts” certainly by the parties, and to a lesser extent by the Court.⁵⁷⁸ The parties had approximately one month to prepare for a preliminary-injunction hearing in the most significant mid-decade redistricting case in recent memory. The Court likewise worked diligently to schedule a preliminary-injunction hearing at the earliest possible date and to issue substantive rulings on motions filed in the interim.⁵⁷⁹ Not to mention the Court's considerable efforts to issue its preliminary-injunction ruling on a nearly impossibly short fuse. Issuing a thoroughly researched and well-reasoned preliminary-injunction opinion of over 150 pages in just 38 days—after awaiting expedited proposed fact findings, legal conclusions, and briefing from the parties, which followed a nine-day evidentiary hearing featuring 23 witnesses and thousands of exhibits on the entire congressional map for the second-largest state in the country—is a Herculean task. Nevertheless, the panel has done everything in its power to rule as quickly as possible.

This case is not one in which “local elections [are] ongoing,” poll workers have already been trained, the voter registration deadline is looming, state election officials have been fully operating under the new map for months, a signature deadline has passed, or the state is only days or weeks away from an election.⁵⁸⁰ This case is one in which, despite the time constraints imposed by the Legislature, sufficient time remains for an injunction to take effect. Therefore, Purcell does not apply.

2. If Purcell Applies, the Plaintiff Groups Satisfy Purcell’s Heightened Showing

Even if Purcell were to apply, the Plaintiff Groups have satisfied its requirements. This litigation—under Purcell—is the prototypical “extraordinary case where an injunction” is “proper.”⁵⁸¹ Under Purcell’s heightened showing, a plaintiff “might be [able to] overcome [the Purcell principle] even with respect to an injunction issued close to an election if a plaintiff establishes at least the following: (i) the underlying merits are entirely clearcut in favor of the plaintiff; (ii) the plaintiff would suffer irreparable harm absent the injunction; (iii) the plaintiff has not unduly delayed bringing the complaint to court; and (iv) the changes in question are at least feasible before the election without significant cost, confusion, or hardship.”⁵⁸² Although the full Supreme Court has not adopted this Purcell exception, the Fifth Circuit has done so, so we apply it accordingly.⁵⁸³

First, undue delay. Without question, the Plaintiff Groups satisfy their showing on this element. The Court has already discussed this point and will re-emphasize it here: the Plaintiff Groups (as well as the State Defendants and the Court for that matter) could not possibly have acted faster or more diligently. On August 18, 2025, the Plaintiff Groups moved “the Court to schedule an expedited September preliminary injunction hearing on Texas's soon-to-be-enacted congressional map.”⁵⁸⁴ Two days later, the Court scheduled a status conference for August 27.⁵⁸⁵ By then, or within a day thereafter, all of the Plaintiff Groups had filed their motions for preliminary injunction—before Governor Abbott even signed the bill into law.⁵⁸⁶ During the status conference, the Court heard extensive argument on timing.⁵⁸⁷ The Plaintiff Groups asked—actually “begged”—the Court to set the preliminary-injunction hearing as soon as possible, vowing that they were ready to begin the hearing any day the Court scheduled it.⁵⁸⁸

The Court scheduled the preliminary-injunction hearing for October 1 to give the parties time to prepare while still giving the Plaintiff Groups the earliest possible hearing date.⁵⁸⁹ Preparing for a nine-day preliminary-injunction hearing in just one month—including the preparation of briefing, arguments, examinations, expert reports, witnesses, and exhibits—is no small feat. The Plaintiff Groups and the State Defendants met that challenge and in doing so exceeded the Court's expectations for preparedness, thoroughness, and professionalism.⁵⁹⁰ There is no evidence that the Plaintiff Groups unduly delayed bringing their claims to the Court. In fact, everyone—the Plaintiff Groups, the State Defendants, and the Court—worked as quickly as possible at every stage of these preliminary-injunction proceedings.

“This is not a situation in which [the Plaintiff Groups] were sleeping on their rights.”⁵⁹¹ The Plaintiff Groups moved for a preliminary-injunction hearing, the Court held a status conference on that motion and scheduled the preliminary-injunction hearing, and the Plaintiff Groups filed their motions for preliminary injunction all before Governor Abbott signed the 2025 Map into law. Then all parties proceeded one month later with a nine-day preliminary-injunction hearing—including a full day of trial on a Saturday—that involved more witnesses and exhibits than most trials on the merits. If that's not maximum diligence, what is?

Second, feasibility of changes close to the election. Because of the Plaintiff Groups’ (and the State Defendants’) rapid response to the new map, the changes necessary to use a map other than the 2025 Map are feasible at this stage of the election without “undue collateral effects.”⁵⁹² The Court has already discussed in detail the ways in which enjoining the 2025 Map would not disrupt the election or cause voter confusion.⁵⁹³ The Court need not repeat them here. The Court adds that even Ms. Adkins testified that the Texas election officials and systems are more than capable of proceeding with the 2026 congressional election under any map that is the law.⁵⁹⁴ As a result, any burden the State would incur is not only minimal, but also far outweighed “by the overwhelming public interest in enjoining C2333 [the 2025 Map] and protecting Plaintiffs’ constitutional rights.”⁵⁹⁵

That leads to the third element: irreparable harm. For the same reasons previously discussed, the Plaintiff Groups would suffer irreparable harm absent the injunction. The obvious harm here is the likely violation of the Plaintiff Groups’ constitutional rights absent the injunction.⁵⁹⁶ The Plaintiff Groups will be forced to proceed under a congressional map that likely unconstitutionally sorts voters on the basis of race. Proceeding in this way deprives the Plaintiff Groups of their right to participate in a free and fair election. That deprivation is a per se irreparable harm.⁵⁹⁷ And this irreparable harm outweighs any marginal voter confusion not already present because of the Legislature's late-breaking passage of the 2025 Map.

Fourth, the underlying merits. Again, the Court will not rehash its painstaking analysis of the merits. As explained above in great detail, this Court's majority finds the underlying merits are clearcut in favor of the Plaintiff Groups.⁵⁹⁸ The Court recognizes the panel's non-unanimous decision weighs against this finding.⁵⁹⁹ But given the indubitable direct evidence in this case, the circumstantial evidence, and the Court's inability to assign the mapdrawer's intent to the Legislature,⁶⁰⁰ “[a]t this preliminary juncture, the underlying merits” do not “appear to be close.”⁶⁰¹ The Plaintiff Groups have clearly shown a likelihood of proving that at trial.

3. As Both a Legal and Practical Matter, Purcell Cannot Apply to This Case

These legal conclusions are further buttressed by the fact that applying Purcell to this case would lead to absurd results.⁶⁰²

If the Court were to consider Robinson and Milligan dispositive, as the State Defendants suggest, the Plaintiff Groups would have had a right to bring their constitutional claims without any real opportunity for their requested remedy of a preliminary injunction. As this Court explained above, Robinson’s 189-day line would have foreclosed the Plaintiff Groups from even filing a motion for preliminary injunction, and Milligan’s 120-day line would have rendered that motion futile.⁶⁰³ Applying Purcell under either timeframe would mean the Plaintiff Groups’ motions for preliminary injunction were dead on arrival. Purcell and its progeny, like Robinson and Milligan, would bar the Plaintiff Groups from seeking a remedy that they have a legal right to seek. Reading Purcell and its progeny to lead to this result is diametrically opposed to the fundamental right of access to the courts that the Constitution affords plaintiffs.⁶⁰⁴

Even without an injunction, the Plaintiff Groups would not have been left without any remedy. The Plaintiff Groups could proceed with their claims to a full trial on the merits. Indeed, “practical considerations sometimes require courts to allow elections to proceed despite pending legal challenges,” even if those legal challenges may prove meritorious.⁶⁰⁵

But this case is not one of those times. The practical considerations that courts refer to in cases like this one are the “imminence of the election” and “inadequate time to resolve the factual disputes.”⁶⁰⁶ Here, those practical considerations arise solely because of how close to the election the Legislature drew the 2025 Map. A final adjudication on the merits after one or more election cycles have passed would run roughshod over the purpose of a preliminary injunction to provide merited, immediate relief. That is especially the case when, as here, the Court is working within—not creating—the timeframe dictated by the Legislature and when the Court finds in favor of the Plaintiff Groups on the merits of their preliminary injunction. Denying the injunction based on such practical considerations would also eschew this Court's obligation to bestow the Plaintiff Groups’ merited, preliminary relief. Purcell cannot be read to gut the Plaintiff Groups’ right to seek a preliminary injunction and this Court's obligation to award one when merited.

Applying Purcell to this case would also incentivize legislatures to redistrict as close to elections as possible. The Governor first placed redistricting on the proclamation for the first called special session on July 9, but the session didn't start until July 21.⁶⁰⁷ That means the first day the Legislature could even take up redistricting was less than eight months before the congressional primary election, less than four months before the candidate filing period opened, and less than two months before the precinct chair filing period opened. About seven weeks later, the Legislature passed the new map, and five days after that Governor Abbott signed it into law. Solely because of the Legislature's and the Governor's timing, the Court had less than seven months before the primary election and less than three months before the candidate filing period to determine whether the new map was constitutional. By acting late, the State has not wholly surrendered the reasonable deference Purcell provides it to run elections as it pleases.⁶⁰⁸ But if under Purcell this Legislature-imposed timeframe mandates denying an injunction, then the State would be immune from any immediate, legitimate constitutional challenge to its redistricting efforts. To secure an unchallenged election under a new map, the Legislature would need only to pass the map close enough to an election to foreclose any judicial review. No court has applied Purcell to mean legislatures have a license to belatedly redistrict at the expense of voters’ constitutional rights for even one election, if not more.

Taking this logic one step further, applying Purcell based on the timeframe established by the Legislature and the Governor would allow the State's executive and legislative branches to hamstring the courts. The Plaintiff Groups had a viable legal claim against the 2025 Map as soon as the 2025 Map became law on August 29. As the Court has explained, some readings of Purcell could foreclose that claim as early as July or at a variety of dates from then through November 3. Applying Purcell in this way would mean the Plaintiff Groups had a viable legal claim against the 2025 Map only after the point at which the Court could reasonably adjudicate any claim against that map for preliminary-injunctive relief. That application of Purcell would amount to placing the starting line beyond the finish line.

This particular dynamic has serious implications for the interplay between legislatures and the courts in the election context. To allow legislatures to redistrict as close to elections as possible while limiting the courts’ ability to review the constitutionality of that action—even in extraordinary cases like this one—would unduly tip the balance of the separation of powers between the legislative and judicial branches and impair the effectiveness of the Constitution's protections of voting rights. If all parties and the Court act with maximum diligence, and the Court finds the map is likely unconstitutional, and yet that likely unconstitutional map can still be deployed, then a legal proceeding like this one is a waste of time and a perversion of the Constitution. If the rule were otherwise and Purcell precluded relief in this case, any legislature could pass a blatantly unconstitutional new congressional map the day before the election, and the courts would be impotent to do anything about it. Denying an injunction in this case on the basis of Purcell permits such a scenario—a scenario that would allow for more election chaos, thereby undermining Purcell’s raison d’être.

It is precisely because of cases like this that Purcell is not “absolute.”⁶⁰⁹ The Court does not presume here “to articulate Purcell’s precise boundaries.”⁶¹⁰ “Whatever Purcell’s outer bounds” may be, this case does not fall within them.⁶¹¹ If it did, the law would hollow out the Plaintiff Groups’ right to seek a preliminary injunction, foreclose this Court's obligation to award a meritorious remedy, license legislatures to flout plaintiffs’ constitutional rights, and undermine the delicate balance of power between the State's law-making branches and the judiciary's obligation to review the constitutionality of even hastily passed redistricting legislation. The law does not and cannot compel that result, and this Court won't either.

* * *

This Court has been attuned to Purcell from the moment the Plaintiff Groups moved this Court for a preliminary-injunction hearing. At the August 27, 2025, status conference, this Court questioned the parties about how Purcell could affect a possible injunction.⁶¹² The Supreme Court has made clear that “lower federal courts should ordinarily not alter the election rules on the eve of an election.”⁶¹³ Indeed, the Supreme Court has stayed a lower federal court's election-related injunctions at least six times in the last 11 years.⁶¹⁴ This Court is not naïve to that reality.⁶¹⁵ But this Court is also not naïve to the likely unconstitutional realities of the 2025 Map.

Without an injunction, the racial minorities the Plaintiff Groups represent will be forced to be represented in Congress based on likely unconstitutional racial classifications for at least two years.⁶¹⁶ In this case, the Plaintiff Groups’ constitutional right to participate in free and fair elections is not outweighed by minor inconveniences to the State's election administrators and to candidates nor by any residual voter confusion, which would be marginal at best given the short timeframe since the 2025 Map was passed.

Based on the foregoing, this Court finds that the balance of equities and the public interest favor the Plaintiff Groups.

IV. REMEDY

Having found all four preliminary-injunction elements weigh in favor of the Plaintiff Groups, the Court next considers the appropriate remedy. Reverting to the 2021 Map is the proper remedy here. Despite the Plaintiff Groups’ previous legal challenges to the 2021 Map, there are several reasons why reverting to that map is the most legally sound and reasonable solution. First, this remedy is the one the Plaintiff Groups request.⁶¹⁷ Second, the 2021 Map was drawn by the Legislature, and courts favor legislative-drawn maps over judicial ones.⁶¹⁸ Third, the State has already used the 2021 Map in two previous congressional elections and is still using it in one special election that is ongoing, as we have already discussed.⁶¹⁹ As a result, the State could “easily ․ make the change” back to the 2021 Map.⁶²⁰ No “complex or disruptive implementation” is involved.⁶²¹

Reverting to the 2021 Map is also more proper than giving the Legislature an opportunity to redraw the map before issuing an injunction, as the State Defendants ask the Court to do.⁶²² “Since 1966, the Supreme Court has repeatedly reminded lower federal courts that if legislative districts are found to be unconstitutional, the elected body must usually be afforded an adequate opportunity to enact revised districts before the federal court steps in to assume that authority.”⁶²³ Courts should usually afford legislatures this opportunity because “redistricting and reapportioning legislative bodies is a legislative task which the courts should make every effort not to preempt.”⁶²⁴

Here, the Court does not need to afford that opportunity for both practical and legal reasons. Giving the Legislature that opportunity is impracticable.⁶²⁵ “Since the [L]egislature is not scheduled to be in session this year” or even next year, giving the Legislature an opportunity to fix the map “would require that the Texas Governor call a special session.”⁶²⁶ It is highly unlikely that the Governor could call a special session and that the Legislature could draw and pass a new map in that special session before the candidate filing deadline of December 8. Additionally, the Court has identified a serious legal flaw in the 2025 Map,⁶²⁷ and the 2021 Map is already a viable congressional map that was drawn by the Legislature.⁶²⁸ By reverting to the 2021 Map, this Court will not preempt the Legislature's authority to draw its congressional districts. Rather, this Court will uphold the Legislature's authority while requiring the least amount of change and disruption to both Texas's election officials and voters.

V. CONCLUSION

For the foregoing reasons, the Court GRANTS the Plaintiff Groups’ motions for preliminary injunction as to their racial-gerrymandering claims:

(1) “Plaintiff Texas NAACP's Motion for a Preliminary Injunction” (ECF No. 1142);

(2) “Plaintiff-Intervenors’ Motion for Preliminary Injunction” (ECF No. 1143);

(3) The “Gonzales Plaintiffs’ Motion for Preliminary Injunction” (ECF No. 1149); and

(4) The “Brooks, LULAC, and MALC Plaintiffs’ Joint Motion for Preliminary Injunction” (ECF No. 1150).

The Court thereby ENJOINS the State of Texas from using the 2025 congressional map and ORDERS the State to use the 2021 Map, as it did in the 2022 and 2024 elections.

So ORDERED and SIGNED on Galveston Island this 18th day of November 2025.

DISSENT FROM THE MEMORANDUM OPINION AND ORDER GRANTING PRELIMINARY INJUNCTION

“Fasten your seatbelts. It's going to be a bumpy night!”¹

I dissent from the entirety of Judge Brown's opinion granting a preliminary injunction.

* * * * *

PRELIMINARY STATEMENT

I append this Preliminary Statement to dispel any suspicion that I'm responsible for any delay in issuing the preliminary injunction or that I am or saw slow-walking the ruling. I also need to highlight the pernicious judicial misbehavior of U.S. District Judge Jeffrey Vincent Brown.²

In my 37 years on the federal bench, this is the most outrageous conduct by a judge that I have ever encountered in a case in which I have been involved.

In summary, Judge Brown has issued a 160-page opinion without giving me any reasonable opportunity to respond. I will set forth the details. The readers can judge for themselves.

This three-judge district court held a nine-day evidentiary hearing/trial on the motion for preliminary injunction. That hearing was concluded Friday October 10. The judges immediately retired to confer. Judges Brown and Guaderrama voted to grant the preliminary injunction. I voted to deny. It was understood that the majority judges would begin putting together an opinion.

During the next 26 days, there was silence—nary a word from either judge.

On Wednesday November 5, Judge Brown sent me a 13-page outline of the expected majority opinion “so that you and your chambers might be able to begin preparing your dissenting opinion.”

Nothing else for a week.

On Wednesday November 12, Judge Brown sent a message stating, “We currently anticipate issuing our injunction on Saturday, November 15. We will endeavor to get you a draft before we issue it. Sadly, we do not believe we can wait for a dissenting opinion before we rule—the fuse is simply too short in light of Purcell. We will, however, note on the opinion that you are dissenting. We are not trying to cut you out, we just don't have the time. Ideally, of course, we'd have liked to have seen your dissent before we issue our opinion, but that will also be impossible.”

Yes, you heard it right. To summarize, in case the reader doesn't get the point: Judge Brown was announcing that he would issue an opinion three days later—an opinion that I hadn't even seen and might not be furnished before its issuance. That is unthinkable, but it occurred—and not accidentally.

A day later, at 10:31pm Thursday November 13, Judge Brown sent a message stating, “I've attached a complete draft of our memorandum opinion and order granting the injunction. We still have revisions to make, but we wanted to get this to you to assist in the preparation of your dissent.” The draft was 168 pages, 655 footnotes, and departed noticeably from the outline I had received. Again, this was the very first actual opinion draft that I had been allowed to see (five calendar days before the actual opinion was sprung).

I was out of town on Thursday and Friday, November 13 and 14, to attend the funeral of (coincidentally) a District Judge of the Western District of Texas, having driven all day Thursday. In my absence, my staff continued working. I drove back home Friday, arriving after midnight, so that my staff and I could spend all day Saturday and Sunday working on the dissent.

Early Sunday morning, November 16, Judge Brown sent a message stating, “I've attached a newly revised draft of our majority opinion. We're still making revisions, but this is pretty close to the final version. We are now intending/hoping to issue it on Tuesday, November 18.” That second draft was 161 pages and contained some substantial revisions from the first (November 13) draft.

I replied that I had been out of town; was writing the dissent all weekend; and would be on the road all of the next day (Monday) to attend graveside services for the deceased federal judge. I said Judge Brown had no business issuing an opinion as soon as Tuesday.

At 11:27am Tuesday November 18, Judge Brown wrote the following: “I've attached a final version. We still intend to issue it today. I'm sorry that we can't wait on your dissent. Purcell compels us to get the ruling out as soon as we possibly can. It turns out that's today.” That third version, 160 pages, was issued a few minutes later (with a small number of additional changes) and was signed “So ORDERED and SIGNED on Galveston Island this 18th day of November 2025.”

This outrage speaks for itself. Any pretense of judicial restraint, good faith, or trust by these two judges is gone. If these judges were so sure of their result, they would not have been so unfairly eager to issue the opinion sans my dissent, or they could have waited for the dissent in order to join issue with it. What indeed are they afraid of?

Judges on multi-judge courts understand how important is the deliberative process to fair and accurate judicial decisionmaking. As I say later in this dissent, judges get paid to disagree as well as to find common ground. Judges in the majority don't get to tell a dissenting judge or judges that they can't participate. If the two judges on this panel get away with what they have done, it sets a horrendous precedent that “might makes right” and the end justifies the means.

The majority might even say “We don't need to wait for your dissent and wouldn't read it if we did.” Here, that sort of happened: The entry on the district court docket brings up only Judge Brown's opinion; the reader has no access to this dissent without opening a separate, non-consecutive docket entry. So this majority has “won” in terms of diminishing the impact of the dissent and the public's access to it. In the interest of justice, one can hope it is only a Pyrrhic victory.

When I was a newer on the bench, a friend asked me, “Now that you've been a judge for a few years, do you have any particular advice?” I replied, “Always sit with your back to the wall.”

* * * * *

DISSENT

The main winners from Judge Brown's opinion are George Soros and Gavin Newsom. The obvious losers are the People of Texas and the Rule of Law.

I dissent.

* * * * *

In the interest of time, this dissent is, admittedly, disjointed. Usually, in dissenting from an opinion of this length, I would spend more days refining and reorganizing the dissent for purposes of impact and readability. But that approach is not reasonably possible here because these two judges have not allowed it.

The resulting dissent is far from a literary masterpiece. If, however, there were a Nobel Prize for Fiction, Judge Brown's opinion would be a prime candidate.

* * * * *

Judge Brown could have saved himself and the readers a lot of time and effort by merely stating the following:

I just don't like what the Legislature did here. It was unnecessary, and it seems unfair to disadvantaged voters. I need to step in to make sure wiser heads prevail over the nakedly partisan and racially questionable actions of these zealous lawmakers. Just as I did to the lawmakers in Galveston County in Petteway, I'm using my considerable clout as a federal district judge to put a stop to bad policy judgments. After all, I get paid to do what I think is right.

* * * * *

In 37 years as a federal judge, I've served on hundreds of three-judge panels. This is the most blatant exercise of judicial activism that I have ever witnessed.

There's the old joke: What's the difference between God and a federal district judge? Answer: God doesn't think he's a federal judge. Or a different version of that joke: An angel rushes to the head of the Heavenly Host and says, “We have a problem. God has delusions of grandeur.” The head angel calmly replies, “What makes you say that?” The first angel whispers, “He's wearing his robe and keeps imagining he's a federal judge.”

Only this time, it isn't funny.

I dissent.

* * * * *

Judge Brown is no stranger to a spirited attack on a legislative body's exercise of its duly-elected power to redistrict. Before being roundly reversed by the Fifth Circuit sitting en banc, Judge Brown, imagining himself to be a legislator, wrote the following:

The 2021 redistricting process ․ occurred within a climate of ongoing discrimination affecting Black and Latino voting participation.

․

․ Black and Latino residents of Galveston County bear the effects of discrimination ․

․

Anglo commissioners are evidently not actively engaged in specific outreach to Galveston County's minority residents.

․

Black residents in Galveston County are more likely to be arrested, and Black and Latino residents comprise a disproportionate percentage of jail and prison inmates ․

․

[T]he plaintiffs do not need to initially show that partisan affiliation does not cause divergent voting patterns.

․

․ Practices exist in Galveston County, including voter purges and racially disparate access to polling places.

․

․ [I]t is stunning how completely the county extinguished the Black and Latino communities' voice on its commissioners court during 2021's redistricting.”

․

This is not a typical redistricting case. What happened here was stark and jarring. The commissioners court transformed Precinct 3 from the precinct with the highest percentage of Black and Latino residents to that with the lowest percentage. The circumstances and effect of the enacted plan were “mean-spirited” and “egregious” given that “there was absolutely no reason to make major changes to Precinct 3.

Petteway v. Galveston Cnty., 698 F. Supp. 3d 952, passim (S.D. Tex. 2023) (Brown, J.), rev'd, 111 F.4th 596 (5th Cir. 2024) (en banc).

Concluding that the district court “was wrong,” the en banc court remanded “for the district court to consider the intentional discrimination and racial gerrymandering claims ․” 111 F.4th at 614. Today, as a legislator/activist jurist, Judge Brown finds a likelihood of success on the instant racial gerrymandering claims.

In regard to the Galveston County matter: Stay tuned for what Judge Brown will rule on remand. In regard to the preliminary injunction in the case at hand, read on.

* * * * *

The ultimate question is whether unrestrained ideological judicial zeal should prevail over legislative choice. This isn't my first rodeo. Fourteen years ago, dissenting from a flawed three-judge redistricting order in this very court, I wrote the following:

․ “[R]eapportionment is primarily a matter for legislative consideration and determination.” White v. Weiser, 412 U.S. 783, 794 ․ (1973). Accordingly, district courts are bound to “follow the policies and preferences of the State, ․ in the reapportionment plans proposed by the state legislature, whenever adherence to state policy does not retract from the requirements of the Federal Constitution.” Id. at 795 ․ (emphasis added). The aim of giving such due regard to plans proposed by the State is so the court will “not preempt the legislative task nor intrude upon state policy any more than necessary.” Id.

․

Justice Samuel Alito, in a recent debate discussing “activist judges,” explained that judges are not theorists or social reformers․ Because the conscientious and well-intentioned majority has ventured far beyond its proper role ․, I respectfully dissent ․, in the hope that on appeal, the Supreme Court will provide appropriate and immediate guidance.[ ]

Two weeks later, the High Court noted probable jurisdiction and set a special oral argument. Less than two weeks after argument, the Court unanimously vacated the order from which I had dissented.

Unfortunately, here we go again.

I dissent.

* * * * *

Speaking of fortune: Just a few weeks ago, the Fifth Circuit answered the main question at hand, holding that “[t]he most obvious reason for mid-cycle redistricting, of course, is partisan gain.”³ The question for this three-judge district panel is whether the Texas Legislature did its mid-decade congressional redistricting to gain political advantage or, instead, because the main goal of Texas's Republican legislators is to slash the voting rights of persons of color.

Once again, here we go again: Criticizing the behavior of DOJ lawyers in last decade's redistricting battle, I noted the following:

It was obvious, from the start, that the DoJ attorneys viewed state officials and the legislative majority and their staffs as a bunch of backwoods hayseed bigots who bemoan the abolition of the poll tax and pine for the days of literacy tests and lynchings. And the DoJ lawyers saw themselves as an expeditionary landing party arriving here, just in time, to rescue the state from oppression ․ The [DoJ] moreover views Texas redistricting litigation as the potential grand prize and lusts for the day when it can reimpose preclearance via Section 3(c).[4]

Although the United States is no longer participating in the instant case, the same attitudes about Texas Republican legislators have been reflected in the testimony of multiple experts and witnesses presented by these plaintiffs and, occasionally but not always, by their talented counsel and the statements of some parties.⁵

Because the “obvious reason” for the 2025 redistricting “of course, is partisan gain,” Judge Brown commits grave error in concluding that the Texas Legislature is more bigoted than political.

I dissent.

* * * * *

It's all politics, on both sides of the partisan aisle. George and Alex Soros have their hands all over this.

One of the plaintiffs' top experts is Matt Barreto. He is a paid Soros operative and does not attempt to hide it. His CV confirms it. He expects to receive $2.5 million ⁶ from George and Alexander Soros.⁷ Nor is this something new. Soros has been pumping money into Barreto's UCLA Voting Rights Project for years.⁸ And this steady supply of money won't stop until 2026, at the earliest. Unsurprisingly, Barreto has been on quite a road show for years, parading across the country opposing Republican redistricting.⁹

That is the tip of the iceberg. The lawyers are involved as well.¹⁰

To his credit, the lead counsel for plaintiffs does not try to hide it, either. Chad Dunn acknowledged so in open court—he works with Barreto at the same Voting Rights Project ¹¹ that receives Soros funding. Dunn is a respected attorney in Texas election law cases, most recently serving as counsel in the Jackson case,¹² in which the Fifth Circuit squarely declared the political nature of mid-decade redistricting. Mr. Dunn, along with his Voting Rights Project colleague Sonni Waknin, also represented the plaintiffs before Judge Brown in the Petteway case, which was overturned by the en banc Fifth Circuit.¹³

Mark Gaber also appeared in Petteway and Jackson. He is the Senior Redistricting Director at Campaign Legal Center, a Soros-funded group.¹⁴

It does not stop there. The Elias Law Group draws from the Soros coffers, too. Counsel for the instant Gonzales plaintiffs, David Fox, is a partner at Elias, which “has collected more than $104 million” from Democrat Party committees and donors, including Mr. Soros.¹⁵ Firm Chair Marc Elias formed entities, “tucked inside large existing nonprofits,” that “raised tens of millions of dollars from some of the richest donors on the left—including from foundations funded by Mr. Soros.”¹⁶

On a silver platter, Judge Brown hands Soros a victory at the expense of the People of Texas and the Rule of Law.¹⁷ Judge Brown won't tell you that. I just did.¹⁸

Relatedly, Gavin Newsom took a victory lap in Houston to celebrate the Democrat redistricting win with Proposition 50.¹⁹ Indeed, he did so “on rival Gov. Greg Abbott's home turf Saturday and called on other blue states to push back on a GOP effort to retain control of the U.S. House.”²⁰ And after the improperly premature issuance of Judge Brown's opinion, the Houston Chronicle pointed out that Governor Newsom quickly tweeted, “Donald Trump and Greg Abbott played with fire, got burned -- and democracy won ․ This ruling is a win for Texas, and for every American who fights for free and fair elections.”²¹

That tells you all that you need to know—this is about partisan politics, plain and simple.

I dissent.

* * * * *

Regardless of one's political slant, it's obvious what Texas is trying to do in 2025. The Republicans' national margin in the House of Representatives is so slim that squeezing out a majority might even depend, day-to-day, on whether some seats are vacant because of deaths or resignations.

In 2021, the Texas Legislature, with both houses controlled by Republicans, devised a strategy of creating safe seats for both Republicans and Democrats, but with a decided majority of the state's delegation still Republican. Whether (as a matter of political clout) that was the wisest strategy is disputed and indeed was fulsomely debated in 2021.

In mid-2025, the strategy changed: The new plan was to make more seats winnable for Republicans by moving some Democrats incumbents from their districts and rendering other districts unwinnable by Democrats. That sacrificed the wider margins in some of the old districts. The tradeoff is obvious.

There is some speculation that this new strategy will backfire on Republicans in 2026 because, if they do poorly in the mid-terms, the new Republican seats created in 2025 will be a Pyrrhic victory, because they will lose elections in the closer districts. That is purely a matter of political strategy that federal judges have no business touching.

The challenge faced by these plaintiffs and Judge Brown is to explain how it could be that the Republicans would sacrifice their stated goal of political gain for racial considerations. It makes no sense to advance the notion that the Republican Legislature would draw districts for the purpose of disadvantaging racial and ethnic minorities if, by doing so, they lessen the number of new Republican seats they might gain.

The plaintiffs' theory is both perverse and bizarre. They actually contend that if the Republicans are sincere about gaining more seats, they could have drawn not five, but six, seven, or eight additional seats and that the reason they did not is that the real reason is racial animus. The absurdity of that notion speaks for itself. Yet it's all that the plaintiffs and Judge Brown have to offer to defeat the State's claim that the 2025 lines were drawn for the sake of politics and not race.

That's the central dispute in this case. But “[t]he most obvious reason for mid-cycle redistricting, of course, is partisan gain.”

I dissent.

* * * * *

Judge Brown rushes to issue this injunction before the tension between Section 2 of the Voting Rights Act and racial-gerrymandering jurisprudence is resolved by the Supreme Court in the currently-pending Callais case.²² Given Judge Brown's creative read of the facts and novel approach to the law, he should have considered denying this injunction for that reason alone, recognizing that a fundamental shift in voting-rights jurisprudence is not unlikely. Because the power to stay proceedings “is incidental to the power inherent in every court to control the disposition of the causes on its docket with economy of time and effort for itself, for counsel, and for litigants,” it would have been well within the authority of this three-judge court.²³

The fact that Callais may fundamentally change the nature of this case also weighs in favor of a stay. It is reckless for this court to proceed with opining on the merits, which amounts to nothing more than a general guess as to whether existing voting-rights jurisprudence will survive Callais.

* * * * *

Judge Brown has a lingering habit. He correctly recites part of a legal principle, then veers off track along a spectrum—intentionally misleading at best to false at worst. The opinion is replete with selectively copying and pasting parts of legal rules or standards. Beyond that, things get dicey.

This holds especially for Judge Brown's discussion of the standard for preliminary injunctions.

Judge Brown admits that the first factor—likelihood of success on the merits—is the “most important” and that granting a preliminary injunction is “an extraordinary and drastic remedy which should not be granted unless the movant clearly carries the burden of persuasion.”²⁴

Then, the opinion entirely goes off the rails.

Judge Brown quibbles with the omission of the word “substantial” next to the phrase “likelihood of success on the merits” in the Fifth Circuit mid-decade redistricting opinion from just a few weeks ago,”²⁵ claiming that the omission suggests that “the plaintiff need only show ‘a likelihood of success on the merits.’ ”²⁶ This is intentionally misleading at best and disingenuously false at worst.

How does he get there?

Judge Brown justifies his wish-list formulation of the first factor by noting the factual similarities between Jackson and the instant case: Both involve Texas mid-decade redistricting at the preliminary-injunction stage. But surely he knows that the phrase “extraordinary and drastic remedy” never appears in Jackson. Judge Brown, relying on the factual and procedural analogies between the two cases, would lead the reader to think that that gives him carte blanche authority to excise the “extraordinary and drastic remedy” from his opinion, as well. Nevertheless, he keeps the phrase “extraordinary and drastic remedy” in the standard because he knows he cannot remove the phrase at will.

Judge Brown, no stranger to inconsistency, is wrong.

He should give less consideration to the omission and more consideration to the actual words on the page. Judge Brown accurately cuts and pastes the following: A preliminary injunction is “an extraordinary and drastic remedy which should not be granted unless the movant clearly carries the burden of persuasion,” and the likelihood of success on the merits is “the most important” factor of the framework.

But the cut-and-paste job is selective. Judge Brown left out the fact that, giving attention to the relevant cases cited in Jackson, “the most important” factor language in Jackson ²⁷ is a direct quote from Mock v. Garland.²⁸ And any cursory reading of Mock easily reveals that the word “substantial”²⁹ (the word Judge Brown tries to avoid) is part of the first factor in no uncertain terms: “a substantial likelihood of success on the merits.”^{30  31}

Judge Brown doesn't tell you that. I just did.

The opinion is caught in an illogical straitjacket from which it cannot escape.

Knowing that his argument is weak, Judge Brown declares that the omission of the word “substantial” does not matter anyway because of the Fifth Circuit's sliding-scale ³² approach to the first factor, which is likelihood of success.³³ With a magic wand, the quibble with the omission of “substantial” is no longer consequential and vanishes into the ether. This is part of the activist, result-oriented bag of tricks that tinkers with the allegedly “most important” first factor, such that the quibbles that he proclaimed mattered no longer do.

Judge Brown says “ ‘[w]here the other factors are strong,’ the movant need only show ‘some likelihood of success on the merits” to obtain a preliminary injunction.”³⁴ This is intentionally misleading at best, disingenuously false at worst.

There he goes again.

Judge Brown overlooks what immediately follows the passage on which he relies:

Where other factors are strong, a showing of some likelihood of success on the merits will justify temporary injunctive relief. But when a plaintiff applies for a mandatory preliminary injunction, such relief should not be granted except in rare instances in which the facts and law are clearly in favor of the moving party.

TitleMax, Inc. v. City of Dallas, 142 F.4th 322, 328 (5th Cir. 2025) (internal quotations omitted) (emphasis added).

Judge Brown is wrong on multiple levels. First, he claimed that the first factor alone suffices, indicating that the other factors do not matter. Second, the other factors, discussed below, are extraordinarily week in this case. Third, TitleMax differentiates between temporary injunctive relief and the narrower category of a mandatory preliminary injunction. Judge Brown must surely know that, which is likely why he cherry-picked the language he liked (“some likelihood of success on the merits”), omitted the language he didn't (“temporary injunctive relief”), and inserted what he wanted—a preliminary injunction. If this is not judicial activism, I am not sure what would be. Fourth, Judge Brown is issuing a mandatory preliminary injunction because he is enjoining the implementation of the 2025 Texas Congressional Map and requiring Texas to use the 2021 map. Fifth, the facts and law are not clearly in favor of the moving party.

If this were a law school exam, the opinion would deserve an “F.”

Remember that recent Fifth Circuit redistricting case, the one that Judge Brown said was procedurally and factually analogous to the instant one. Judge Brown conveniently omits the key sentence in that mid-decade redistricting case: The “most obvious reason for mid-cycle redistricting, of course, is partisan gain.”³⁵ Judge Brown doesn't even pretend to grapple with Justice Stevens's relevant quote. It is far from a mere coincidence that the opinion goes to the mats over the omission of one word, when it suits the results-driven outcome, but overlooks the most significant sentence about the most obvious reason for mid-decade redistricting, which is partisan gain.

The combined weight of the procedural and substantive law is against what these plaintiffs and Judge Brown are trying to do. Not only do plaintiffs have to show clearly that they are entitled to the drastic and extraordinary remedy of an injunction, but they must also do so when Supreme Court and Fifth Circuit precedent is stacked against them. Nothing in any bag of results-oriented tricks can save that wished-for result.

Judge Brown is an unskilled magician. The audience knows what is coming next.

Moving past the recitation of the preliminary-injunction factors: Judge Brown does not hesitate to make excuses for plaintiffs (and their “experts”) for failing to produce an Alexander map. He has no other choice on the merits. He claims that “they [the experts] didn't have time”³⁶ and that it would be too much to ask plaintiffs to produce an Alexander map at this stage in the litigation. This is not how the law works for a preliminary injunction.

Judge Brown overlooks that plaintiffs seeking a preliminary injunction bear the burden of proving that they are entitled to it. With nothing more than meager direct evidence in the instant case, Plaintiffs must produce an Alexander map, plain and simple. They either cannot or don't want to—because it's really all about politics. In any event, this court has no business coming to the rescue by giving students who didn't do their homework a homework pass. Nor should Judge Brown make excuses for them for failing to show their work.

The last time I checked, a preliminary injunction is an extraordinary and drastic remedy. This is serious business that we are about.³⁷

Judge Brown boasts that “Plaintiff groups have successfully shown a likelihood of success on their racial-gerrymandering challenges ․ [and] that alone suffices to preliminarily enjoin the 2025 Map.”³⁸ Yes, you read that right. Judge Brown is so determined to issue an injunction that he does not need any help from the other factors.³⁹

How could that be? Because Judge Brown said so.

With his creative formulation of the preliminary-injunction standard, Judge Brown is intentionally misleading at best and disingenuously false. He engages in several layers of sophistry to water down the potency of the most important, first factor and to grease the skids for an injunction. He doesn't even make it clear which articulation of the first factor he uses.

Consider this bizarre multiple-choice question from hell: Which formulation of the first factor is he using? Is it the “likelihood of success” factor that is the (i) watered-down formulation because of the omission of the word “substantial,” (ii) the watered-down formulation because of the sliding scale, (iii) the watered-down formulation because of both the sliding scale and omission of the word “substantial,” (iv) the “substantial” formulation with the sliding scale, (v) the “substantial” formulation without the sliding scale, (vi) whatever Judge Brown thinks the law should be, or (vii) something else?

Confused yet? You can thank Judge Brown for that.

If we were to take him at his word that the first factor is dispositive (it is not)⁴⁰ to grant a preliminary injunction, it is not apparent why Judge Brown feels the need to discuss the other factors. His mind is made up on the first factor alone. But I will move on from that to discuss them anyway.

Judge Brown claims that the “Plaintiff Groups have made a very strong showing on the irreparable-injury factor.”⁴¹ Not so fast. First, plaintiffs are unlikely to succeed on the merits of their racial gerrymandering claim, so they are unlikely to suffer harm. Second, plaintiffs, bearing the burden of clearly showing they are entitled to an extraordinary and drastic remedy, cannot use circular reasoning to bootstrap their alleged likelihood of success from factor one into showing irreparable harm with factor two. Indeed, “[w]hen a statute is enjoined, the State necessarily suffers the irreparable harm of denying the public interest in the enforcement of its laws.”⁴²

He caps off the section by returning to the sliding scale again (the same one he claimed was not necessary) to reiterate his preferred standard that plaintiffs “need to show more than just ‘some likelihood of success on the merits’ to obtain a preliminary injunction, but not much more.”⁴³ This is wrong, again.

Judge Brown gets creative with the final two factors, balance of equities and public interest, and stands the Purcell framework on its head. He wants a “federal court to swoop in and re-do a State's election laws in the period close to an election”⁴⁴ and issue a “late-breaking injunction”⁴⁵ with disastrous, unintended consequences for “candidates, political parties, [ ] voters,”⁴⁶ the State, counties, and local officials. This injunction will affect down-ballot races because those interesting in running for Congress must make plans not to run for State House and Senate seats. And others are sure to run for the newly-vacant state seats. This trickle-down effect is only the tip of the iceberg. Judge Brown's injunction is the epitome of judicial tinkering.

The 2025 map is the status quo. Counties have begun preparations with 2025 map and educating local officials about the current law. Although Judge Brown acknowledges that the State has the prerogative to “toy with its election laws,”⁴⁷ he quickly contradicts himself that the State “invited this issue by enacting a new map within Purcell's range.”⁴⁸ Contrary to what Judge Brown wants to hear, the State, which has the prerogative to redistrict mid-decade, is in a fundamentally different position from that of a federal court, which must exercise extraordinary caution before intermeddling with an intimately vital local prerogative such as redistricting.⁴⁹

Judge Brown parrots plaintiffs' argument that the State is using the 2021 map in some limited circumstances.⁵⁰ But Judge Brown doesn't attempt to grapple with what the Fifth Circuit has made clear: A duly enacted Texas congressional districting map is the “status quo.”⁵¹ There, the Fifth Circuit said in no uncertain terms that “the Texas Legislature's duly enacted law” creating a new congressional districting map “became the new ‘status quo’ ” under Texas law.

Instead, Judge Brown cherry-picks the “status quo” language ⁵² out of another Fifth Circuit case,⁵³ where the court made it clear that “the Supreme Court has instructed that we should carefully guard against judicially altering the status quo on the eve of an election.” Whether Judge Brown likes it, he needs to acknowledge two realities. First, the duly enacted 2025 Texas Congressional Map is the status quo. But true to form, Judge Brown prefers living in fantasyland. Second, Judge Brown's late-breaking, eleventh-hour injunction is the precisely the kind of “judicial tinkering”⁵⁴ and judicial altering ⁵⁵ that the Court has repeatedly warned us about. I guess Judge Brown needs another reminder.

Whether Judge Brown likes it, gravity exists. So does the weight of Purcell against his late-breaking, eleventh-hour injunction.

There's more.

Judge Brown fails to recognize that some of these plaintiffs are seeking an equitable remedy, namely a preliminary injunction, with unclean hands. Contrary to his inventive contention that the State is to blame for the delay, some plaintiffs broke quorum and delayed the passage of the 2025 map for weeks.⁵⁶ Judge Brown contradicts himself again, claiming that Purcell does not bar him from issuing and injunction and then turns around to wag his finger at the State for the cause of the delay. He is mistaken. Plaintiffs should not get the benefit of the delay that they caused by breaking quorum. But, Judge Brown has no problem giving plaintiffs an equitable remedy, even though they have unclean hands. The so-called Purcell exception, which Judge Brown is eager to invoke, does not apply: Plaintiffs caused undue delay, the merits are not remotely in their favor, and plaintiffs have not suffered an irreparable injury.

I dissent.

* * * * *

To show the fallacies in Judge Brown's opinion, the following discussion of the direct and indirect evidence includes a granular examination of Texas's U.S. congressional districts in the 2021 maps, plan C2193,⁵⁷ and the various editions of the 2025 maps.

The 2025 maps first were offered as plan C2308,⁵⁸ in the first special legislative session on July 30, 2025.⁵⁹ Then, after August 15, the Texas legislature updated them to plan C2331.⁶⁰ The final version, introduced on August 18, passed on August 23, and signed into law on August 29 as HB4, was plan C2333.⁶¹ Immediately below, I reproduce the 2021 maps, plan C2193, and the 2025 adopted map, plan C2333.⁶² Careful consideration of these maps, and attention to changes in certain districts such as C2193-CD35 to C2333-CD35, is fundamental to understanding this case and to distinguishing between a racial gerrymander and a cynical partisan gerrymander by disentangling race from politics where “race and partisan preference are highly correlated,” as is strictly required under Alexander.⁶³

Tabular or graphical material not displayable at this time.

The 2021 Maps, C2193.

Tabular or graphical material not displayable at this time.

The 2025 Maps, C2333.

Everybody agrees that a plaintiff asserting a racial-gerrymandering claim may “make the required showing through direct evidence of legislative intent,”⁶⁴ such as “a relevant state actor's express acknowledgement that race played a role in the drawing of district lines,”⁶⁵ “circumstantial evidence of a district's shape and demographics, or a mix of both.”⁶⁶ The legislative intent is the critical question, and the Supreme Court has instructed that “legislators who vote to adopt a bill are not the agents of the bill's sponsor or proponents,” as “legislators have a duty to exercise their [own independent] judgment.”⁶⁷

So, let's talk about the direct evidence first, and then the indirect and statistical evidence.

* * * * *

This panel decides both law and fact. The salient issue of fact is whether the Legislature drew the new lines on account of race. The answer is easy: It did not. And that question is not even close.

Did I forget to mention: “The most obvious reason for mid-cycle redistricting, of course, is partisan gain.”

In that regard, everyone can agree that the star witness was Adam Kincaid. For months, there was controversy as to who drew “the map.” Without dispute, it turns out to be Kincaid. He is a paid, experienced, dedicated Republican operative, through and through. His lengthy testimony was the highlight of the preliminary-injunction trial.

Kincaid courageously spoke the truth, despite being the target of what authorities termed a “credible death threat” made shortly before he was scheduled to testify. As one of the finders of fact, I conclude that Kincaid was credible in every respect.

Knowing that Kincade is credible, Judge Brown makes every effort to ignore or circumvent Kincaid's solid testimony. Judge Brown avoids the details of that testimony. Because he won't tell you that, I do so now.

Adam Kincaid's testimony is credible and irrefutable. Beginning in the Panhandle and moving clockwise, he went district-by-district and described his map-drawing process with painstaking detail (and without any notes for two days). His testimony is methodically detailed, and he is a solid witness, especially on the key question of intent and race.

I begin with a roadmap. The preliminary discussion provides a brief background on Kincaid and his general approach to redistricting, which prioritizes partisanship and disclaims any reliance on race. First, I detail Kincaid's traditional redistricting criteria. Second, I highlight judges' questions to Kincaid and Kincaid's responses. Third, I describe Kincaid's district-by-district testimony organized by the relevant Texas region. Fourth, I describe what Kincaid noted as at least three changes between C2308 and C2333.

Adam Kincaid drew all or most of the Texas 2025 enacted congressional map. Tr. 10/7/25 AM 33:25-34:2.⁶⁸ Specifically, he used software, “Esri for Redistricting.” 41:7-13. In no uncertain terms, Kincaid stated “I don't think it's constitutional to draw maps based off of race.” 46:13-14. He unequivocally said “I do not” use race as a proxy for partisanship when drawing a map. 56:7-9. Instead, he reiterated that he used partisan data at the block level. 47:20-52:19. He said, “I drew my map using politics from start to finish and provided that to the Legislature.” Tr. 10/7/25 PM 93:11-12. As if he could not be clearer, Kincaid repeated, “I drew a race-blind map using partisan results, and that's how I created the map.” Tr. 10/8/25 AM 69:6-7.

Kincaid used traditional redistricting criteria. His top priority was to protect incumbents and improve or maintain existing Republican districts. His “top criteria was to make sure that every Republican incumbent who lived in their seat stayed in their seat.” 64:23-25. “Another criteria was to make sure that every Republican incumbent who was in a district that President Trump had won with 60 percent of the vote or more in 2024 stayed in a district that President Trump won by — with 60 percent of the vote or more.” 65:1-5. In fact, Kincaid “was not allowed to take any incumbent Republican who was above 60 below 60.” 65:5-6. For Republican districts with incumbents that Trump carried by less than 10 points, Kincaid had to either “improve [these seats] or keep their Partisan Voting Index exactly the same.” 65:10-11.

Kincaid's criteria in the five pickup opportunities were Trump+10, a Ted Cruz victory, a strong Abbott performance, and a durability test.

First, “every single one of [the Republican pickup opportunities] had to be a district that President Trump carried by ten points or more at a minimum” in the 2024 Presidential Election. 67:25-68:1, 68:12-14. Second, “every one of those seats had to be carried by Ted Cruz in 2024,” by any margin. 68:2-5. Third, the districts were generally those in which Governor Abbott “carried by as decent a margin as possible” in 2018 and 2022 because the “first test of this map would be in a midterm election versus a presidential election.” 72:9-17. Fourth, Kincaid ran a “durability test” on these districts, looking “at every presidential, senate, and governor's race in Texas, U.S. Senate and governor's race in Texas, from 2012 through 2024.” 73:8-20.

Kincaid admitted that was not looking at the Cruz and Abbott numbers in Republican districts that were not pickup opportunities because “it is a fair assumption that if you are drawing a seat at 60 percent Trump, it probably went Republican down ballot as well.” 150:17-25.

For other criteria, Kincaid used the balancing of population as well as compactness and neutral geographic features.

Kincaid had to balance population perfectly among the 38 districts in the state. 54:1-16. He “wanted to take [ ] districts [in the 2021 map] and make them cleaner, more compact, more city-based, [and] more county-based.” 66:22-25. He considered neutral geographic units or boundaries when drawing districts. 75:17-23. He “tried to use neutral boundaries across the entire map where possible.” 100:10-11.

Judge Brown actively questioned Mr. Kincaid. He asked, “When you drew the 2025 map, did you know that CDs 9, 18, 29, and 33 under the 2021 map were considered minority opportunity districts, in that they provided minorities an opportunity to elect candidates of their choice?” Tr. 10/8/25 AM 133:14-17.

Kincaid said that he knew. Id. 133:18.

Kincaid said that he was generally aware that a comfortable majority of Hispanics in Texas vote in favor of Democrat candidates, notwithstanding President Trump's better performance among Hispanics. Id. 133:19-134:1. Kincaid added that he “know[s] that President Trump carried Hispanic voters by about 10 percent statewide by various reports in 2024.” Id. 134:1-3.

When asked why he changed CD 9 from Democrat to Republican but left CD 7 Democrat, Kincaid said that “[t]here were political constraints on the west side of Harris County,” although he “actually wanted to flip that one.” Id. 134:4-9. The structural orientation of Congressman Luttrell's seat (in CD 8), Congressman McCaul's seat (in CD 10), and Congressman Hunt's seat (in CD38) prevented Kincaid from “restructur[ing] the population in 7 enough to redraw that seat.” Id. 134:12-15.

Judge Brown asked whether Kincaid received any instructions to protect (or not alter) Democrat districts similar to those instructions Kincaid received during the 2021 map drawing process. Id. 134:1-23. Although Kincaid testified that he received some instructions while drawing the 2021 map to protect some Democrat districts, he did not receive similar instructions regarding the 2025 map. Id. 134:16-23.

District by district, Kincaid drew the map by starting at the northwest corner and generally working clockwise.

I recount Kincaid's testimony in the order that it appeared in his direct examination, which typically coincides (but not necessarily) with the order in which he drew the Texas 2025 congressional map.

The only district that did not change at all was Texas District 19. 77:13-15. Beyond that, Kincaid began his map-drawing in the Texas Panhandle.

Texas 13 was the first district drawn, which is in the northwesternmost part of the state. 76:9-77:4. Intuitively, starting with the northwestern part of the state (the top left of the map) makes perfect sense. Indeed, Texas 13 is in the Panhandle and stretches across North Texas south of the Red River. 77:5-18.

Kincaid changed the lines in Wise and Denton Counties first. 77:16-18. Specifically, he moved some Democrats from the southwestern side of Denton County out of District 26 into District 13. 78:7-12. Because he had added some people into the 13th District, Kincaid had to take people out—he “took the line for Texas 26 and moved it north into Wise County.” 79:12-16. He also kept the cities at the center of Wise County whole. 78:14-17.

Kincaid reiterated that he “worked in a clockwise direction through Metro DFW.” 80:1-2. He took heavily Democrat precincts in the southeast corner of Denton, previously drawn out of 26 during the last redistricting, back into District 26 in this new map. 80:10-25. Kincaid put Democrats into the 26th District to “move Republican strength across the state from district to district” and “make sure that the 26th District didn't become too Republican.” 81:7-13.

District 4: After the piece of Frisco in Texas 26 was taken out, District 4 took on all of Frisco, making Frisco whole in District 4. 81:23-82:5. Kincaid took the 2021 map's three-way Plano split within Collin County and made it a two-way split with a clean line dividing Plano. 82:18-83:7. To the north, the part of the city of Celina, which is in northwestern Collin County, is whole in the 4th District. 84:10-14. Kincaid fixed the population of District 4 in the east, noting that he made the county with Clarksville (presumptively Red River County) whole. 84:17-22. The military installation in Bowie County was also made whole in the 4th District. 84:23-85:4.

District 3: Kincaid also made Allen and McKinney whole in District 3. 83:25-84:2. Because the 3rd District picked up more Democrats in the Plano area that it had before, he included more Republican strength, from rural East Texas counties, into the district. 85:10-13.

For the Dallas/Fort Worth Metroplex (DFW), we begin with District 32. The border between Districts 3, 4, and 32 is the city boundary of Richardson. 85:15-16. Kincaid made Richardson whole in District 32. 86:3-5. Four years ago, Texas District 32 could have been redrawn, but Kincaid did not take the opportunity to do so. 87:2-4. He took 40% Republican areas in North Dallas County, which were more Republican than the rest of the county, and paired them with more Republican counties east of Dallas County to create a new Republican district that extended from North Dallas County to the east. 87:12-88:9.

District 5: Kincaid had to keep Kaufman, Van Zandt, and Henderson Counties whole in District 5. 89:20-22. They had to remain the core of the district, per the instruction from the Texas Republican congressional delegation. 89:23-90:2. On the eastern side of Dallas County, Kincaid made Seagoville and Mesquite whole. 90:23-91:6. Kincaid used the Garland and Dallas city line between Districts 5 and 33 to move District 5 to the northwest, including areas that are more Republican. 91:6-13. However, Kincaid added the Democrat precincts north of 33 and east of 24 to District 5, which lowered the Republican support in the district. 10/7/25 AM 91:22-92:12. To counteract this and keep the district at 60% or above, he added Anderson County, which had been there in the previous decade, back into the 5th District. 92-18-20. Finally, Kincaid included north of downtown Dallas to bring District 5 to population. 92:19-20.

District 24: Kincaid kept the Park Cities, University Park and Highland Park, whole in District 24. 93:5-11. He made Farmers Branch, which was previously split, whole as the “conduit from the Park Cities to the west.” 94:22-24. Kincaid went into the southeast, where there were precincts in the “40s for President Trump versus the ones further down that are much bluer,” to balance the population. 94:7-14. Because District 24 was held by a Republican under the 2021 map, Kincaid made sure to ensure that the district office for District 24 stayed in the district. 95:1-21. Therefore, Addison had to be split slightly to keep the district office in District 24. 95:11-21. Admittedly, Kincaid did not prioritize keeping district offices for Democrat incumbents in the same way. 95:22-25.

Because the 24th District gets most of its Republican strength from Northeast Tarrant County, Kincaid used Farmers Branch as a conduit to “connect the western side of the district with the eastern side of the district in one continuous seat” and make the city boundary whole. 96:17-97:6. Kincaid made the city of Coppell whole and made the split in Irving to the north to make sure that Congresswoman Beth Van Duyne continues to live in District 24. 97:13-18. In Northeast Tarrant County, he “made sure that the district boundary aligned with the cities of Euless, Hurst, and Richland Hills, as well as North Richland Hills and Watauga.” 98:6-9. Kincaid made a small split of Haltom City to balance the population and added a few precincts to “clean up” the line on the western side of 24 between Districts 12 and 24. 98:10-14. The interstate forms the northwestern boundary of District 24. 98:15-20.

District 12: Kincaid left the Parker County line the same because he needed to ensure that Congressman Williams, a Republican incumbent, continued to reside in his seat in the 25th District. 99:3-8. The border between Districts 12 and 25 was set at the Haltom City line; from there, Kincaid used rivers down to the major road. 99:10-17. He balanced the population in Southwestern Tarrant County. 10/7/25 AM 99:17-18. His goal was to keep the district above 60% Trump, protect the Republican incumbent, and “absorb more Democrats in the seat.” 99:19-21. He used neutral boundaries whenever possible, including Interstate 20 and the South Fork of the Trinity River. 100:23-101:9.

Districts 30 and 33: Kincaid drew one “megadistrict ․ of the most Democrat VTDs [he] could find in Dallas and Tarrant County.” 102:8-11. He did so to avoid having to redraw districts that he was otherwise satisfied with. 102:2-11. After doing so, he moved to District 6. 102:23-103:4.

To divide the one “megadistrict” into two districts, Kincaid used partisan shading to put together clustered precincts south of downtown where President Trump received 20% or less of the vote (“very Democratic precincts”) into one seat for District 30. 109:18-110:8. From there, Kincaid worked west, assigning Democrat precincts to District 30. 110:11-14. Kincaid took about 250,000 people from heavily Democrat precincts in southeastern Tarrant County into District 30, creating a portion that juts into Tarrant County. 110:15-111:8. Using neutral boundaries, Kincaid set the border between Districts 30 and 33 — he used Interstate 20, working north to the local metro line, and then again joined a highway. 111:20-112:3. There is a small triangle with a “little nub” south of the interstate where Kincaid balanced the population. 112:25-113:6.

Kincaid made clear that his objective was to “make [District] 30 the more heavily Democrat seat of the two” to make for a more compact seat. 113:12-114:1. He had no concern about incumbents in Democrat districts. Tr. 10/7/25 PM 67:14-16. District 33 was simply the district “left over from the creation of [District] 30 within the super district.” 114:8-12.

Although there may be territory to the northeast that is in District 33 that is more Democrat than the territory in Tarrant County, Kincaid did not go for the District 33 territory because he “was using the footprint of [District] 30 as it currently existed.” Tr. 10/7/25 PM 71:11-19. Kincaid also noted that he considered building a more Democrat district by having it take on central Dallas County but did not do so because it created “a wall of a whole bunch of Democrats on the eastern side,” which he would have needed to move west. 73:3-21. This decision is why Kincaid “took the 30th District down ․ and put in its current footprint.” Tr. 10/7/PM 73:16-18. He said that he was generally maintaining the borders of District 33 and only moved small blocks to balance the population along the edges. Id. 21:10-20.

Kincaid indicated that Congresswoman Jasmine Crockett is no longer in the 30th District. Id. 114:22-24. He agreed that Congressman Veasey was no longer in District 33. Id. 114:14-21.

District 6: The areas in Irving moved significantly to Republicans in 2024 compared to 2020. 103:9-13, 104:6-13. The new District 6 was bound by the city of Irving on the eastern side. 104:16-20. In so doing, Kincaid put more Republicans into District 6 and out of Districts 30 and 33, which made the future Districts 30 and 33 as Democratic as possible. 105:2-9. Kincaid used the city boundary of Arlington and Rendon as a boundary for District 6. 106:2-6. Since the district became more compact and lost several counties to the east, Kincaid made changes to the south for population reasons. 107:3-6. Ultimately, the district “picked up a lot of Arlington.” 107:7-10.

Noting one of the changes between C2308 and C2333, Kincaid made Navarro County whole in the 6th District, which allowed him to get “more Republican strength into [District] 17.” 172:20-173:2.

District 25: The “entertainment district” had to remain in the 25th District. 106:8-9. While drawing the district, Kincaid prioritized the incumbency of Republican Congressman Williams, whose district office is in Cleburne and is a location of a split. See 106:22-25, 109:5-8. The border with District 6 set District 25, meaning that “the border between 6 and 25 was set between the two seats, all the way up through using the Rendon border.” 108:24-109:8.

Regarding the Houston metropolitan area: Kincaid “had already drawn the rest of the state and got to the Harris County area last” because he “like[s] to start in the corners” when drawing maps. 121:23-122:7. Because the central Texas area was the “most complicated to draw,” it was the next-to-last portion of Texas that Kincaid drew. 122:18-21.

District 36: Kincaid “changed the line in Harris to come in and pick up some Democrat areas closer in toward downtown.” 123:1-9. The Jefferson County line stayed “roughly the same” between Districts 14 and 36. 123:9-10. Kincaid used Interstate 10 as the dividing line between Districts 14 and 36. 174:21-22. Kincaid said that he drew CD 36 with 61.8% Trump 2024 general support. Tr. 10/8/25 AM 34:18-21.

Because Kincaid added Liberty County to District 9, District 36 became “underpopulated by about 93,000 people” and noncontiguous. 174:14-16. As a result, Kincaid had to change the way that District 36 was drawn through Jefferson County. 174:18-20. He took District 36 into the northern part of Brazoria County to “add population in 36 that was not too heavily Democrat.” 175:4-176:12. Kincaid also put three VTDs, previously in District 9, into District 36 to balance the population. 185:2-11. He added these three VTDs because he wanted to make District 36 contiguous and not add more Republicans to District 9 after District 9 got sufficient Republicans from Liberty County. 186:11-18. Kincaid moved these particular VTDs because he did not want to split Baytown or the downtown area in half. 186:19-22.

When asked on cross-examination whether he could have created CD 9 at over 60% Trump by swapping precincts with CD 36, Kincaid acknowledged that he could have done so. Tr. 10/8/25 AM 35:9-12. Responding to a hypothetical that if he had swapped those precincts back and forth to make CD 9 60% Trump whether the Hispanic CVAP would have dropped below 50%, Kincaid said, “I don't know that. That's certainly possible. But I wasn't targeting the Hispanic CVAP numbers.” Id. 35:15-20.

District 14: Kincaid moved District 14 “down through Galveston County and changed the orientation of Brazoria.” 123:11-12. Because he added Liberty County to District 9, the 14th District “ended up growing into Fort Bend County.” 176:17-19.

Kincaid said that he drew CD 14 with 61.5% Trump 2024 general support. Tr. 10/8/25 AM 35:5-7.

District 18: The goal of the redistricting process was to pick up five seats. 123:19-21. Because there used to be four Democrat seats in the middle of Harris County, “one of those seats had to be flipped.” 123:18-21. Kincaid “shaded on the partisanship and looked for the most partisanly Democrat precincts in Harris County and then into Ford Bend and Brazoria Counties and put all of those together in the 18th District.” 124:1-8.

In the northeast portion of District 18, there is an epiglottis-shaped region that sticks down, which consists of “two or three very Democrat VTDs,” a feature that also exists on the 2021 map. 130:10-18.

The 18th District needed to grow in population because District 14 moved into the southern part of Fort Bend County and both Districts 14 and 36 moved into the northern part of Brazoria County. 180:16-21. Therefore, Kincaid brought up District 18 to the Sam Houston Parkway to add population. 181:15-19. The Sam Houston Parkway was the northern border set in District 18. 181:16-19. On the eastern and northern borders between Districts 18 and 29, the more Republican VTDs were drawn in Republican districts. 189:17-21.

District 22: First, Kincaid “changed the southwestern Harris County a little bit ․ and then changed some of the area where 7 came down into 22.” 140:18-22. Specifically, he put the Sugar Land areas that were performing better for Republican candidates into District 22 to make the district as Republican as he could. 141:12-142:6.

On the border between District 14 in Brazoria County and District 22, Kincaid took territory to the south of District 14 and put it into District 22 to “keep the district at a good Republican Trump number ․ or better than it had been before.” 142:14-143:2. The northern part of Brazoria County is Republican, but not as Republican as the area that Kincaid swapped out of District 14 into District 22. 142:18-22. The 22nd District picked up more of Brazoria County, and the area in southwestern Harris County changed. 176:25-177:2. Kincaid moved Republicans from District 22 into District 8, and vice versa, to balance populations. 177:23-178:2. He was able to make District 22 a district that President Trump carried with 60% or more. 178:3-6.

Kincaid considered the Fort Bend County line between Districts 18 and 22 to make sure that District 22, “stayed as Republican as it had been before or got better.” 125:4-21. Indeed, some of the precincts between Districts 18 and 22 are not as “deep blue” as those in District 18, “but they are still much more Democrat than the rest of 22.” 126:21-24.

District 9: Kincaid drew District 9 after he drew District 18. 130:23-24. In fact, Kincaid notes that the “9th kind of drew itself” after he drew Districts 18 and 36 — the eastern border of District 18 and northern border of District 18 were set, so he “took the 9th District up the eastern side of Harris County.” 131:2-7. However, the 9th District did not completely encompass the area north of Baytown because Republican Congressman Crenshaw lives in that area and Kincaid drew around his house to avoid putting him into the 9th District. 131:8-20. Kincaid was “trying to make the 9th district as Republican” as he could so District “36 ended up taking Baytown” and he “took the 9th north from there.” 132:4-13.

When asked where he started when redrawing the Harris County map for Plan 2333, Kincaid said that he “added Liberty County to the 9th District” to make it “redder.” 173:15-174:1. Indeed, Kincaid said that he drew CD 9 in Plan C2333 with about Trump '24 general support at 59.5%. Tr. 10/8/25 AM 34:8-11. Kincaid said that he did not make any change to District 9 based on racial data. 174:5-6. Adding Liberty County to District 9 “created a clockwise rotation around the Houston area.” 176:13-17.

Comparing C2193 to C2333, Kincaid acknowledged that District 29 has been distributed into five districts, the biggest chunk (43%) of which went into the new 9th District. Id. 24:13-20. Kincaid indicated that Congressman Green no longer lives in Congressional District 9—he lives in the new 18th District. Id. 114:25-115:16.

District 2: Indeed, Congressman Crenshaw lives in District 2, a district that President Trump carried with at least 60% of the vote. 131:24-25. If Kincaid drew him in District 9, Congressman Crenshaw would be in a district that President Trump did not carry with at least 60% of the vote. 131:25-132:3.

Kincaid drew District 2 after he drew District 9. 132:17-19. Because District 2 lost population in eastern Harris County based on the way District 9 was drawn, Kincaid added Humble, slightly above 40% Trump support and “redder than the other areas around it,” into District 2. 132:21-133:5.

Kincaid brought District 2 further north into the Conroe area in Montgomery County to add more Republicans because District 2 “had shed a whole bunch of Republicans in northeastern Harris” County. 133:7-20. To keep District 2 above 60% Trump support, Kincaid extended District 2 “along the northwestern side of 29,” where there “are a series of competitive but Democrat-leaning precincts.” 133:23-134:17. He also made sure that The Woodlands was “relatively whole” in District 2, as it had been before. 135:5-9.

Kincaid added the Kingwood area in northeastern Harris County back into District 2 to help make it a reliable 60%+ Trump seat. 179:19-22.

District 29: District 29, north of District 18, was a “pretty straightforward draw.” 125:2-3. Kincaid drew District 29 after he drew District 2. 135:11-12. He took the heavily Democratic precincts on the northern border of the district and eastern side of Humble and put them in District 29, working his way south to create the most Democratic seat in the area. 135:15-21, 136:20-22. Kincaid could not have put the finger-like portion of eastern Humble, a heavily Democratic VTD, in District 2 “because that would have endangered the 60 percent Trump target in 2.” 136:5-15.

From the west side of CD 29 where a “finger ․ carves down on the right side” bordering District 2 to the bottom part of the district bordering the 610 Loop, Kincaid captured heavily Democrat precincts. 136:23-137:13. Notably, he used the 610 Loop as the southern border of District 29 because it was a natural boundary. 137:11-18. Kincaid brought in a small area south of the 610 Loop to balance the population. 138:8-10. Between Districts 18 and 29, Kincaid used a railroad track, instead of the VTD line, to clean it up. 139:7-14. And between Districts 7, 18, and 29, Kincaid used roads, interstates, and railroad tracks as boundaries, as done in the Dallas area. 140:7-11.

When asked about the change in District 29 from C2193 to C2333, Kincaid acknowledged that District 29 was “definitely reworked.” Tr. 10/8/25 AM 23:6-9. Comparing C2193 to C2333, Kincaid acknowledged that District 29 has been distributed into five districts, the biggest chunk (43%) of which went into the new 9th District. Id. 24:13-20. About 37% of District 29 remained in District 29. Id. 24:21-23. The remainder of the district went into District 7 (2%), District 18 (8%), and District 36. Id. 24:24-25:9.

District 38: Kincaid was trying to give District 38 as Republican a character as he could, so he tweaked the line between Districts 29 and 38 to make sure he got as many Republicans as possible into District 38 and out of District 29. 138:13-139:2. He adjusted the line between Districts 8 and 38 to “get the 38th District back to where it had been in the previous draw.” 145:6-9. District 38, which had lost Republican territory to District 2, was the last piece to fall into place in its area. 145:15-22.

District 14: The Congressman in District 14 wanted all seven ports that he represented to remain in the 14th District, which is why 14 is shaped the way it is at the bottom. 143:3-13. A heavily Democrat precinct on the south side of District 18 in C2308 was added to District 14 to make District 14 contiguous with the area just below District 18. 177:15-18.

District 7: When asked why he changed CD 9 from Democrat to Republican but left CD 7 Democrat, Kincaid said that “[t]here were political constraints on the west side of Harris County,” although he “actually wanted to flip that one” as well. Tr. 10/8/25 AM 134:4-9. The structural orientation of Congressman Luttrell's seat (in CD 8), Congressman McCaul's seat (in CD 10), and Congressman Hunt's seat (in CD 38) prevented Kincaid from “restructur[ing] the population in 7 enough to redraw that seat.” Id. 134:12-15. On cross-examination with Mr. Bledsoe, Kincaid added the 22nd District (with Congressman Nehls) as one of the seats, in addition to those listed above (CD 8, CD 10, CD 38), that constrained him. Id. 141:9-13. Specifically, the 22nd District has “a hook down in the middle of Fort Bend County,” which is a “carve-out for Mr. Nehls' home and a lot of population ․ [t]hat has to go somewhere.” Id. 142:2-6.

Kincaid said that he could not change District 7 from Democrat to Republican because of “the other parameters that [he] had and the constraints with the incumbents.” Id. 140:18-20. He reiterated that “[i]t was just an impossible thing to do,” even though he tried to “create only two Democrat seats in Houston instead of three.” Id. 140:20-23. The structuring of the neighboring seats, incumbent needs, and partisanship thresholds made it impossible to flip District 7 from Democrat to Republican. Id. 142:7-9. Kincaid said that putting the heavily Democrat areas of Harris County in District 18 “into one district on purpose” prevented him, in part, from flipping District 7. Id. 141:18-142:2.

Kincaid tried to “put as many Democrats” as possible into District 7, particularly to the north of District 18. 142:7-9, 143:23-144:2. After working on District 22, he addressed Districts 7, 8, and 38, simultaneously. 143:18-21. Kincaid cleaned up the border between what had been the 9th District and the 7th District, running the border along the bayou that runs to the highway and down to the county line. 144:6-10.

The line between Districts 7 and 22 changed slightly. 177:12-13. Kincaid moved some population from District 18 into District 7 to balance the population. 181:4-9.

District 8: Kincaid put some Republican-leaning, less Democrat VTDs bordering Districts 7 and 8 into District 8. 145:2-5. District 10 comes in over the top of District 8 and picked up Republican precincts from District 8. 124:20-23. District 8 lost some population it had in southwestern Harris County to District 22. 177:3-11. Indeed, Kincaid moved Republicans from District 22 into District 8, and vice versa, to balance populations. 177:23-178:2. Kincaid was “able to put a little more Republican strength back into the 8th District so it didn't sink too far down.” 178:7-9.

District 10: District 10 comes in over the top of District 8 and picked up Republican precincts from District 8. 124:20-23.

Kincaid then addressed the Travis County area.

District 37: Every VTD in District 37, which encompasses the Austin area, was less than 30% Trump support in 2024. 146:22-147:4. Controlling for population equality, the line between Districts 27 and 37 was a “strictly partisan” draw that differentiated along the 30% Trump number. 147:19-148:2.

District 27: Every VTD in District 27 was “30 percent or more Trump in 2024.” 146:22-25. Kincaid wanted to keep District 27 above 60% Trump support. 148:10-11. He moved the 27th District to the north along the Gulf and made sure that Victoria County, where the incumbent lives, was in the 27th District. 148:23-149:1. From there, Kincaid fit the 27th District underneath the 10th District and brought part of Hays County into District 27 to help get above 60% Trump support. 149:1-3, 150:2-8. Although he tried to avoid a split in Refugio, Aransas, and San Patricio Counties, Kincaid made sure that the 27th District was contiguous by road because, otherwise, it would have been only contiguous by water. 149:21-25.

District 34: Kincaid “had to carve out some heavily Democrat precincts in Nueces County and Corpus Christi” to get the 34th District to be a Trump+10 district. 148:18-22. Kincaid said, “Working up from the border, I knew 34 and 28 were already Trump seats, and I knew I was going to make those redder.” 10/8/25 AM 131:20-22.

District 21: Kincaid pulled the 21st District out of Travis County. 155:12-13. He had to keep this district a “60 percent Trump seat” because it was an incumbent Republican seat. 163:12-13, 163:25-164:3.

Kincaid then testified as to what he did with the Central Texas, Bexar County, and Travis County areas.

Districts 31, 17, 11, and 10 are all stacked above the 37th District in the form of a “layer cake.” 151:23-25. The 10th, 17th, 27th, and 31st Districts were all “barely over 60 percent Trump seats,” so much of Kincaid's work was to balance the partisanship among those districts. 152:1-5, 154:21-25.

District 10: Kincaid had to fit District 27 underneath District 10 because the 10th District had been stretched from western Travis County to the east to pick up Brazos County for at least two reasons—first, to accommodate incumbent Republican Congressman McCaul, who lived there, and second to keep the district above 60% Trump support. 149:6-10, 153:1-7 (referencing the “McCaul hook”). In District 10, the “McCaul hook” is so slender because Kincaid had to avoid picking up Democrats closer to downtown and by the university in Brazos County. 154:4-16.

Kincaid was “trying to get as few Democrat areas as possible” in District 10. 171:18-21.

District 11: Kincaid pulled Lee County out of the 11th District and brought north Travis County into the 11th District. 152:6-10. This allowed the 11th District to pick up more Democrat areas in Pflugerville, which is whole in the 11th District. 152:9-10, 153:20-21.

But the southern line of district 11 stayed the same. 155:9. Kincaid did not move any counties into District 11 between Districts 11 and 21 or between Districts 11 and 23. 155:10-11. He kept the north boundary between Districts 11 and 23 unchanged. 160:6-7.

District 31: Kincaid also wanted to make District 31 more compact than it had been under the previous draw. 154:25-155:2.

District 35: District 27 “abuts” District 35, which is in Central Texas. 151:4-5. Kincaid indicated that the drawing of the districts by the border, namely Districts 15, 16, 28 and 34, influenced the way in which he drew District 35. See generally 156:1-161:25. Based on the movement in the border counties (see infra), Guadalupe, Wilson, and Karnes Counties were “free to be worked with” and indeed were combined with the area of Bexar County to make the 35th District. 161:23-162:4.

Like the other pickup opportunities, District 35 needed to be a “Trump plus 10 seat that Ted Cruz had also carried in 2024.” 162:9-11. Kincaid looked at Governor Abbott's strong performance there and performed a durability analysis. 162:15-19. The south side of Bexar County approaching District 35 (but below District 20) allowed Kincaid to make District 35 more Republican. 167:9-12.

Although Kincaid technically could have evened out the Trump performance between adjacent Districts 21 and 35 by giving more heavily Republican precincts to CD 35 (54.6% Trump support in C2333), he could not do so without running afoul of the criteria that 60%+ Republican incumbent districts needed to be at 60%+ Trump support: He could not drop the 21st District (which had 60.2% Trump support and a Republican incumbent) much more. Tr. 10/8/25 AM 38:15-39:16. Kincaid acknowledged that as many as eight precinct splits occurred in a heavily Hispanic area in CD 35. Id. 41: 11-17.

District 20: To allow District 35 to become a true Republican pickup opportunity, District 20 had to “absorb as many Democrats” as possible. 164:10-14. Kincaid wanted to make District 20 as Democrat as he could. 164:16-19. Kincaid put parts that had previously been in District 35 into District 20. 164:7-22. He made a straight line between Castle Hills and Olmos Park as the northern border of District 20. 166:10-19. He said that San Antonio “had to be split no matter what.” Tr. 10/8/25 AM 38:1-5.

Kincaid noted that he drew the Kirby area into District 20, not District 35, because there is a “steady line of heavily Democrat precincts that are contained within 20 and then a smattering of 20 percent [Trump] precincts – or heavily Democrat precincts with smaller ones clustered in [the] Kirby area.” 168:16-22. He did so because he wanted to “maximize the Trump and Cruz numbers,” not simply maximize Republican performance overall. Tr. 10/7/25 PM 74:9-17. Kincaid was not concerned about an incumbent in District 20 or 35. Tr. 10/7/25 PM 76:6-15.

Kincaid remarked that the draw In Bexar County (Districts 20, 21, 23) was very complicated. For one, the 21st District could not move more to the west. 165:3-4. In an ideal world, Kincaid would have put the precincts on the west side of District 20 into a more Republican seat. 165:5-7. However, Kincaid could not do so because moving those precincts to District 23 would make District 23 more Democrat, causing it to miss its political targets. 165:6-10.

District 21: Kincaid made three small cities whole in the 21st District. 166:6-10. Although Kincaid could have evened out the Trump performance between adjacent Districts 21 and 35 by giving more heavily Republican precincts to CD 35 (54.6% Trump support in C2333) as a technical matter, he could not do so without running afoul of the criteria that 60%+ Trump districts with Republican incumbents in the 2021 map needed to remain at 60%+ Trump support — he could not drop the 21st District (which had 60.2% Trump support and a Republican incumbent) much more. Tr. 10/8/25 AM 38:15-39:16.

Kincaid then addressed the border counties.

District 34: Kincaid drew the 34th District as a “series of whole counties all the way up the Gulf Coast” until he “ran out of population in Corpus Christi.” 156:2-5. This took the 34th District out of Hidalgo County, making it a more compact district in the north. 156:5-7.

District 15: This was a complicated draw for Kincaid because the district was an “R plus seven district” for incumbent Republican Congresswoman Monica De La Cruz, and Kincaid needed to keep the district at the same margin. 156:15-23. Kincaid “had to pick up the eastern Hidalgo County part” that he “had just drawn out of 34,” which made things complicated because this part of Hidalgo County consisted of 52% Trump VTDs. 156:24-157:5. As a result, Kincaid included counties that had previously been part of District 34 into District 15 this time. 157:6-9.

Kincaid had to make sure that the incumbent congresswoman continued to live in her seat. 157:14-18. He reiterated that he starts at the corners while map drawing. 158:3-5. Overall, District 15 moved to the east. 161:20-21.

District 23: Kincaid needed to ensure that the 23rd District stayed at “R plus seven or greater during the draw” because it had a Republican incumbent. 158:20-25. Kincaid made Horizon City whole in District 23. 159:18-19. He included VTDs north of “where it says El Paso” in District 23 because those VTDs were 50% Trump. 159:19-24. Generally, Kincaid included Republican areas of El Paso County in District 23. 160:2-4. Kincaid kept the north boundary between Districts 11 and 23 unchanged. 160:6-7.

District 16: Kincaid's border between Districts 16 and 23 did not make it into the final map, and Kincaid did not draw the change between Districts 16 and 23 between C2308 and C2333. 7-14. Kincaid made Socorro whole in District 16. 159:19.

District 28: Kincaid took the remainder of Hidalgo County and put it into District 28. 160:12-14. Then, he “used whole counties up to Atascosa and balanced the population of [District] 28 in Maverick County.” 160:14-16. District 28 was a Republican pickup opportunity drawn to be a “Trump plus 10 seat.” 160:22-23. Overall, District 28 moved south. 161:21. Kincaid said, “Working up from the border, I knew 34 and 28 were already Trump seats, and I knew I was going to make those redder.” 10/8/25 AM 131:20-22.

Kincaid noted at least three changes between C2308 and C2333.

First, he made Navarro County whole in the 6th District. 172:20-25.

Second, the Texas House changed a part of the map in El Paso—Kincaid did not draw this change. 173:3-4.

Third, there was a rotation of seats in the Houston metropolitan area. 173:6-7.

In conclusion, Kincaid's testimony is credible and irrefutable. His two-day testimony (without any notes) was detailed, methodical, and meticulous. When given the opportunity to do so, on both direct and cross, he had a perfectly legitimate and candidly partisan explanation for his every decision.

Despite testifying under a death threat, Kincaid was calm and straightforward. He is a solid witness on the key question of intent and race, and I easily credit his testimony as wholly convincing and unassailable.

Kincaid's testimony is fully consistent with the law: “The most obvious reason for mid-cycle redistricting, of course, is partisan gain.” As Kincaid cogently explained, he was put in charge of that partisan gain for Texas in 2025. And as his testimony shows, it was all about politics, not race.

I dissent.

* * * * *

After outlining Mr. Kincaid's compelling testimony on the map-drawing process, we need to consider his statements, along with those of Senators Phil King and Adam Hinojosa, and Chairman Cody Vasut, which Judge Brown considers to be defense-favorable direct evidence,⁶⁹ and weigh them against those of Chairman Todd Hunter,⁷⁰ Speaker of the House Dustin Burrows, Representatives David Spiller, Tom Oliverson, and Steve Toth, which Judge Brown considers to be damaging direct evidence.⁷¹ Of course, Judge Brown buries this question of legislative intent—the principal question in the case—after a lengthy recitation of ambiguous and contradictory direct evidence on the White House's pressure, outside media coverage, the DOJ's letter, the Texas AG's letter, and Governor Abbott's statements,⁷² none of which can easily be attributed to the Legislature, and all of which butts up against Alexander's presumption of good faith for legislatures.⁷³

So, how should you weigh the evidence in this case? Judge Brown admits, as he must, that legislative intent remains the fundamental question.⁷⁴ Yet legislative intent is notoriously challenging to discern.⁷⁵

These are the main competing bodies of evidence:

• first, the Texas legislators' statements, notably including Hunter, Burrows, Vasut, Hinojosa and King;

• second, the actual outcomes on the map drawn in Plan C2333;

• third, Adam Kincaid's testimony as the map-drawer;

• fourth, Governor Abbott and other Texas politicians' statements, generally to the media;

• fifth, the Department of Justice and Donald Trump's statements.

Each one is relevant and probative, but some are more relevant than others. In particular, the (1) legislators' statements, (2) actual map adopted by them, and (3) the map-drawer's explanation—as agent for the legislature—of every choice made during drawing the map look the most probative.

Meanwhile, statements of politicians in Texas's executive branch (including the governor and attorney general) or statewide delegation to the United States congress are less probative of the Texas legislature's intent.

Further, statements by non-Texas federal politicians in Washington D.C. are even less probative, though Judge Brown repeatedly hangs his hat on this nigh-irrelevant body of information, contrary to Alexander and the manifest weight of the evidence.⁷⁶ Opposite to the clearly-established law, they fail to draw competing inferences as they are required to.⁷⁷ I will point out each of these wrong turns, so we can make a U-turn and get back on track.

To unwind this narrative, we may have to bounce around, so bear with me.⁷⁸

* * * * *

Judge Brown singles out representatives Hunter, Oliverson, Burrows, and Toth.⁷⁹ Simultaneously, it buries Vasut, Hinojosa and King's contrary evidence with little basis.⁸⁰ It also relies upon statements from members of the opposing party—notably Representative Thompson to Chairman Hunter and Senator Gutierrez to Senator King.

Judge Brown centrally focuses on Chairman Hunter's exposition of the racial demographics of the new map on the floor of the Texas House, including his colloquies with Representatives Pierson and Spiller.⁸¹

The Supreme Court, however, has emphasized that legislators will “almost always be aware of racial demographics” when drawing districts, so it imposes a higher standard before subjecting districts drawn with awareness of racial data to strict scrutiny—otherwise, redistricting might be impossible.⁸²

Nothing Judge Brown says gets past ambiguity. He argues that Hunter's reciting demographics and mentioning Petteway jointly “suggests that the mapdrawers purposefully manipulated the districts' racial demographics to convert coalition districts into single-race-majority districts.⁸³ Suggestion, as against the Alexander presumption of good faith, is not enough.

So, how to best interpret Chairman Hunter's exposition of these facts and figures? Interpreting Hunter's invocation of both Rucho and Petteway, Judge Brown flouts Alexander's presumption of good faith to draw the forbidden rather than permitted inference.⁸⁴

Faithfully applying the presumption of good faith, the more plausible explanation is that Chairman Hunter was publicly attacked in the 2021 redrawing, again bound-up in the history of this case, and felt motivated to defend his reputation and that of the Texas house by expositing the racial statistics of the new map. That easily covers his presentation of the new maps on August 1, 2025, and why he “volunteered” Hispanic CVAP statistics. Hunter had previously been attacked and pilloried as a racist in the 2021 cycle—so, for him to present figures that he explained were increasing the number of majority-Hispanic districts easily fits the inference that he was aiming to defend the bill and bolster his credibility.

Further, drawing this positive inference is consistent with legislative awareness of race—which Judge Brown concedes, but then breezily walks by, contrary to Alexander. Hunter provided more than enough favorable commentary to support the positive inference—discussing the race-blind drawing process, apparently delighting in the partisan advantage of Rucho—so, for Judge Brown to insist that he harbored inward racial animus on this ambiguous fact pattern unfairly paints Hunter, a former democrat, as an unreformed, unrepentant racist maintaining a flagging veneer of partisan nastiness over Strom Thurmond-like segregationism. This upside-down fantasy entertained by Judge Brown is plain error and justifies reversal.

But Judge Brown compounds his error of drawing a negative rather than positive inference from individual legislators' mixed and conflicting statements. He interprets Speaker Burrows' mix of partisan and post-Petteway or anti-coalitional thinking in a post-passage press release from August 20 as showing racist intent.⁸⁵ Keep in mind, this was weeks after drawing the maps and after heated floor debates involving Rucho and Petteway.⁸⁶ In that direct, 1:1 tradeoff, Alexander commands this court to draw the positive inference.

Similarly, out of 88 House Republicans voting for the bill, he snipes at Representative Oliverson's and Toth's statements to the press. In Oliverson's NPR interview, he mentions Petteway, but in the next breath disclaims specific knowledge of the bill and invokes Rucho.⁸⁷ On this conflicted piece of evidence, Alexander requires the partisan inference. Toth's statement was similarly made during a sprawling TV interview, with the added context that Toth is running for the U.S. House of Representatives. There, he said, “Texas just went ahead when we drew these maps, as Joan Huffman said, I drew the maps blind to race. And that's what we did,”⁸⁸ while offering a wide range of conflicting purely-partisan and Petteway rationales. Again, Alexander demands the partisan inference.

Judge Brown also hand-waves past Chairman Vasut, Senator Hinojosa, and even Senator King's statements showing partisan intensity as the legislature's motive.⁸⁹

Judge Brown ignores Chairman (of the Redistricting Committee) Vasut's contemporaneous statements, made during the map-drawing process on August 2.⁹⁰ Judge Brown also downplays Senator Hinojosa's speech defending the bill on partisan grounds, despite that speech, delivered in the legislature, having equal or greater probative significance than errant remarks from Oliverson or Toth outside the legislature.⁹¹

Where Judge Brown attacks Senator King for his minimal involvement in the bill-drafting process, he does not apply the same lens to Burrows, Toth, or Oliverson.⁹² Almost all the house Republicans co-sponsored the bill: 78 in total. And worse for him, Chairman Hunter disclaimed any knowledge of the redistricting process earlier in the summer until he was asked to carry it on the floor. Given King's prior discussions with Kincaid at ALEC, how can Judge Brown claim that King was uninvolved, but everyone else knew and embodied the legislative intent?

Instead of weighing those against Chairman Hunter's statements in the aggregate and applying the presumption of legislative good faith to the entire collective body of the Texas legislature, Judge Brown seizes onto a tendentious interpretation of Hunter's statements and then imputes that to the whole legislature—House and Senate alike!

Worse for Judge Brown, there is no evidence that Hunter drew the maps, so any of his exposition of the racial statistics resulting from the outcome of that process is a posteriori, rather than probative of the legislature's invidious racial intent in drawing the maps.⁹³ Is it really credible to think that Hunter could have had his own self-contained invidious intent to enact a clean map? That stretches credulity.

Instead, Kincaid presented remarkably credible and ultimately unrebutted evidence proving his drawing of the maps on race-blind criteria including partisan affiliation, natural geographic boundaries, representatives' home and office addresses, and greater compactness in the 2025 than 2021 maps.

Another big problem for Judge Brown is that Kincaid started drawing the maps before the DOJ letter, and far before Chairman Hunter was asked to carry the bill on the floor.⁹⁴ Kincaid was told about upcoming redistricting in Texas in March while on a visit to the White House.⁹⁵ Kincaid also drew the maps last time around, and regularly explores “what is possible or what would have been possible․ across the entire country.”⁹⁶ Concretely, he states that he started drawing these maps as early as June ⁹⁷ —weeks before the DOJ letter—and apparently around the time he told Senator King that five pickups statewide were possible.⁹⁸

So, contrary to what Chairman Hunter told his political opponent Representative Thompson on the floor of the Texas House, the Legislature was redistricting during June.⁹⁹ The probative value of Chairman Hunter's statement to his rival is nada and zilch—where Judge Brown relies upon it, that exposes the weakness of his position.¹⁰⁰ Similarly, where Judge Brown invokes New York Times articles from June discussing the mixed impressions of U.S. representatives from Texas in Washington, D.C., that is minimally probative of the Texas state legislature's intent in Austin.¹⁰¹ They are different people in different places, months before the final enactment.

Looks like Judge Brown's so-called “direct evidence” doesn't amount to a hill of beans.

* * * * *

On legislative intent, to the extent Judge Brown attributes Hunter's intent to the whole legislature, he likely violates Prejean v. Foster.¹⁰² There, the Fifth Circuit rejected on summary judgment and while granting every inference to the nonmoving party—rather than on preliminary injunction and assessing likelihood of success on the merits—the argument that the intent of an external mapdrawer who averred zero racial motivation could be “taken as conclusive proof of the legislature's intent.”¹⁰³ Instead, the fact that the Legislature adopted the external mapdraw's districting plan at best “support[ed] an inference that racial considerations did not predominate.”¹⁰⁴

Here, under a different procedural posture, the question is whether the fact that Kincaid's map was adopted by the Legislature suggests that his intent can be attributed to the legislature. Evaluating this as a standard piece of evidence, rather than granting every reasonable inference to the opposite party, the answer must clearly be yes (in part). At a minimum, Kincaid's intent is probative of the Legislature's intent, given that he acted as their agent in drawing the maps and was given numerous instructions related to incumbency protection at the level of voting thresholds, home addresses, district office addresses, and communities of interest.¹⁰⁵

Judge Brown also rushes past the nuance that courts must be careful not to “overemphasiz[e] statements from individual legislators,”¹⁰⁶ as “[w]hat motivates one legislator to make a speech about a statute is not necessarily what motivates scores of others to enact it.”¹⁰⁷ But in dismissing Chairman Vasut's and Senator Hinojosa's statements disclaiming racist intent, Judge Brown reduces them and dozens of the other members of the Texas legislature to mere cat's paws, or dupes, mopes and muppets following the leader, which theory the Supreme Court criticized in Brnovich v. Democratic National Committee.¹⁰⁸ There, the Court wrote,

The ‘cat's paw’ theory has no application to legislative bodies. The theory rests on the agency relationship that exists between an employer and a supervisor, but the legislators who vote to adopt a bill are not the agents of the bill's sponsor or proponents. Under our form of government, legislators have a duty to exercise their judgment and to represent their constituents. It is insulting to suggest that they are mere dupes or tools.109

The rule is clear: Judge Brown cannot treat the statements of Hunter or Burrows as dispositive of the intent of the full legislative body, not only excluding over 80 other Republicans in the House, but scores more in the Senate.

In sum, Prejean's refusal to equate the intent of an external mapdrawer to the legislature itself cuts in both directions: the statements of an outside drawer are not conclusive in either direction, and need to be weighed for their probativity and credibility, like any piece of evidence. Here, Hunter's statements are minimally probative, while Kincaid's statements are highly probative, consistently delivered, and credible. It is plainly in error for Judge Brown to reach the opposite conclusion.¹¹⁰

* * * * *

Having considered the mixed legislative statements—which individually and aggregately fail to overcome the presumption of legislative good faith—we consider Judge Brown's discussion of the maps' outcomes.

Judge Brown's tour of the circumstantial evidence is lackluster, especially considering his overarching theory of the facts is that “the redistricting bill's sponsors made numerous statements suggesting that they had intentionally manipulated the districts' lines to create more majority-Hispanic and majority-Black districts ․ [which] suggest that they did so because such a map would be an easier sell than a purely partisan one.”¹¹¹ Judge Brown begins by arguing that the Legislature “fulfilled almost everything that DOJ and the Governor desired.”¹¹²

This is fanciful framing at best and intentionally deceptive at worst.

The DOJ letter erroneously singled out four districts as coalition districts. One of those, CD 29, was a majority Hispanic CVAP, meaning the DOJ was incorrect as flagging it as a coalition district in the first instance. However, Judge Brown appears to suggest that changing CD 29 fulfilled the DOJ's goals, even though the Hispanic CVAP dropped below 50% and created a district where no race or ethnic group is a majority of the citizen voting age population.¹¹³ If the Legislature intends to sell this map by emphasizing how many Hispanic majority CVAP districts there are and to claim they were required to eliminate coalition districts, why in the world would they get rid of an Hispanic majority CVAP district and create what at least has the outward appearance of a coalition district? Judge Brown has no answer to that question. With the map's not fulfilling the DOJ's vision of CD 29 and CD 33 remaining a coalition district, the tally stands at 2-2 for doing things that the DOJ letter suggested. Two districts looked like the DOJ wanted them to look and two didn't. Far from the record's making it obvious that Kincaid and the Legislature did the DOJ's bidding, it seems as though Kincaid drew his map blind to race and the bill sponsors, who had virtually no input on the lines in question, just sought to pay lip service to Petteway.

As for Governor Abbott, Judge Brown claims that Abbott wanted to “increase[e] the number of majority-Hispanic districts,” and the Legislature obliged. However, Judge Brown doesn't connect the dots correctly.

There is no evidence in the record, before the map was revealed at the end of July, that the Governor said anything about increasing the number of Hispanic majority CVAP districts. Rather, it's only after the map is revealed that the Governor says anything that can be construed as stating the lines were drawn to increase Hispanic majority districts.¹¹⁴

Far from the map's being drawn with an eye toward achieving the Governor's goal, it appears he adjusted his rhetoric to defend the map in a forward-facing capacity. If the Governor's concern throughout the redistricting process was increasing the number of Hispanic majority CVAP districts, then one imagines he would have said something about it before the legislature revealed a map which happens to have a higher number of Hispanic majority CVAP districts.

Judge Brown then talks about how the map's “ ‘on-the-nose attainment of a 50% [C]VAP' for so many districts suggests that the Legislature was following a 50%-plus racial target’ ‘to the letter,’ such that the ‘racial target had a direct and significant impact on those districts' configurations[s].’ ”¹¹⁵ While it may feel odd or uncomfortable to see four of the thirty-eight districts right at that 50% mark, Judge Brown provides no serious rebuttal to the reasons Kincaid gave for the lines that he drew in those districts.

The Kincaid testimony is thorough and largely based on testable claims about the areas in which he drew the lines. Even if we assume that the plaintiffs don't need to produce an Alexander map, when provided with thorough reasoning concerning the lines that exist and contrary evidence, as found throughout this dissent, that undermines the existence of a racial target, it seems concerning that the only conclusion Judge Brown can come to is that these numbers suggest legislature followed the DOJ's order to the letter, even though they only did half of what the DOJ suggested.

Judge Brown also suggests that the fact that the legislature left a majority white Democrat district largely unchanged is further evidence of racial motivations.¹¹⁶ This claim does not even fit Judge Brown's theory of the facts. Across his lengthy opinion, Judge Brown's theory is that the legislature conspired to make a map that's easier to sell by intentionally creating more minority districts while also still achieving partisan aims. However, here he appears to pivot into a suggestion that the legislature is outright bigoted and that a partisan legislature would try and make significant modifications to CD37 just like it did to the non-white district of CD 9, but failed to do so because CD37 was a white Democratic district.¹¹⁷

This is cherry-picking of the highest order. Of the 5 pick-up opportunities that were majority-minority, CD28 (53.6%) and CD34 (61.6%) kept a majority of their 2021 district intact.¹¹⁸ In comparison, CD32(41.2%) is a white majority CVAP district and kept the third least of its original territory out of the five pickup opportunities.¹¹⁹ It is hard to imagine how a rational actor comes to the conclusion that majority-white CVAP CD37 keeping 6% more of its territory than majority-minority pick-up district CD34 and 26% more than majority white CD32 is evidence of racial predomination. Judge Brown's argument here is just plain faulty, and his discrediting of Kincaid's testimony is more of a judicial handwave than a legitimate, reasoned explanation.¹²⁰

Judge Brown also claims that the fact that a Republican coalition district (CD27) became majority-white is circumstantial evidence of racial gerrymandering. Here, Judge Brown truly shows his biases and nakedly shows that he has no true desire to disaggregate race and politics. Judge Brown doesn't seem to realize that in a political gerrymander, the voting power for flipped districts must come from somewhere. So, one should not “expect the Legislature not to make fundamental changes to the racial demographics of Republican districts” because the only way one is going to pick up seats in a partisan gerrymander is by taking strength from heavily Republican districts and adding them to slightly Democrat districts (or some similar formulation).¹²¹

It's entirely plausible and even expected that the racial composition of some of the Republican districts might change as a result. After all, the people that got added to the district are not the same ones who got removed from the district. When looking back at the record, it's unsurprising to find that, sure enough, CD27 was a district where Republican strength was taken, and Kincaid had to work to keep the Trump numbers above 60%.¹²²

From the very outset, Kincaod admits that the 2025 maps “achieved all but one of the racial objectives demanded by DOJ.”¹²³ Specifically, CD-27 in Houston remains a ‘coalition’ district as previously authorized by Petteway. But Judge Brown's Petteway analysis gets it logically wrong by suggesting that the outcomes were driven by the DOJ letter. If Texas had been responding to DOJ's threat, why would they have left one coalition district on the table still subjecting them to liability? That doesn't make sense.

Instead, the correct inference on Petteway is that if you do not have to draw coalition districts, you may or may not draw them.¹²⁴ And that is exactly what the state did. Texas drew some (CD-27) and dismantled others (CD-9). So, the concept that the Petteway change drove or explains all of the variance is at odds with the facts that some coalition districts still exist, and others do not exist—rather than every coalition district having been eliminated. The right inference is that they were conducting the draw on some other criterion than eliminating all coalition districts.

Plaintiffs also seize upon alleged racial shifts in CD-22 and CD-27, Republican performing districts under both Plan C2193 (2021 map) and C2333 (the final 2025 map), per LULAC Second Supplemental Complaint at *42, 56. They allege that the shifts in composition among those districts are performed for racial reasons. Indeed, Judge Brown suggests that changes to CDs 22, 27, 30, 32, and 35 are racial gerrymandering.¹²⁵ Here, though, he again struggles to disentangle race from politics, given that, as in South Carolina, “race and partisan preference are highly correlated”¹²⁶ in Texas, and these districts are drawn for Republican performance constrained by the knock-on effects from drawing other districts.¹²⁷

Indeed, there are clear knock-on effects in C2333 from creating CD-35, which pulls in Guadalupe and Wilson Counties from the C2193-CD15, which then pulls in counties from the east such as Dewitt and Lavaca, and in turn pushes CD27 further east into Wharton and Matagorda Counties to politically balance out new population from Hays and southeastern Travis Counties, in turn pushing CD-22 into Brazoria County to claw back absolute population. To accuse CD-22 and CD-27 of hewing to new racial targets neglects the far more parsimonious explanation consistent with legislative good faith, which is that those districts were moved east to reflect a partisan gerrymander.

Both Judge Brown and plaintiffs devote relatively little attention to CDs 15, 28 and 34 under the new plan because they reflect anodyne partisan tweaks, as well as reflect the politically-inconvenient reality of Hispanic Texans in the Rio Grande Valley shifting for Donald Trump.

CD-18 in C2333 does track Black CVAP voting precincts, but plaintiffs fail to disentangle race from politics here. While race is a proxy for partisanship, the problem is that partisanship is also a proxy for race. And Black voters in Harris county favor the Democratic party at overwhelming rates, north of 90%, suggesting that a partisan packed map grouping together all the most-intensive Democrat precincts would likely track racial lines, given the parallel trend of residential racial segregation.¹²⁸ Indeed, the much celebrated “dangly bit,” or the eastern prong of CD-18 reaching into CD29 on map C2193 and reaching into CD9 on C2333 tracks just such a residential concentration performing at extremely high rates for the Democrat party. To disentangle the partisan correlation from the racial correlation, where that correlation is above 0.9, requires sensitive statistical analysis. Judge Brown relies completely on Dr. Duchin's analysis—which was, unfortunately, mis-calibrated.¹²⁹

CD-33 remains a coalition district, despite being named in the DOJ letter, which undermines the 1:1 DOJ application theory advanced by Judge Brown, since there is not a pattern of actually dismantling pre-Petteway coalition districts. How can Judge Brown say that only eliminating three such districts in CD-35, CD-9 and CD-18, while leaving in one, amounts to a clear pattern of action? The state only does it 75% of the time, in the one observed instance. If they were really conducting a full-Petteway reversal, and abiding by the DOJ's letter, why would they leave in one coalition district that would subject them to the terrors of Harmeet Dhillon's DOJ enforcement arm? While racial gerrymandering claims may proceed “district by district,” the state map drawing process indisputably took place on a map-wide draw, given Kincaid's unrebutted testimony.¹³⁰

As to Judge Brown's attack—relegated to a footnote—on CD-7, he argues that Kincaid's failure to eliminate CD-7 is probative of racial intent, because a White Democrat, Lizzie Fletcher, holds that seat.¹³¹ Yet Kincaid credibly testified that there were just not enough degrees of freedom, compared to the core retention constraint, given the nearby presence of CD-38 (Wesley Hunt's district), which was itself relatively compact, and the pressure on CD-22 to move northeast from consolidating CD-35 in Bexar county and the changes in the Rio Grande Valley districts.¹³²

* * * * *

In sum, Judge Brown does fine in his recitation of some of the law governing racial gerrymandering claims, but recitation and application are different things, and his application of law to facts is sorely wanting. To begin, it has been stated multiple times by the Supreme Court that federal courts must “ ‘exercise extraordinary caution in adjudicating claims that a State has drawn district lines on the basis of race.”¹³³ We act so cautiously because reviews of districting legislation “represents a serious intrusion on the most vital of local functions.”¹³⁴ Judge Brown's analysis is not careful, nor does it appreciate how serious an intrusion is being made here.

Judge Brown's direct evidence analysis is contradictory and legally wrongheaded. He cites Common Cause Florida v. Byrd for the proposition that the purported motivations of the DOJ and the Governor “do not become those of the [Legislature] as a whole unless it is shown that a majority of the [Legislature's] members shared and purposefully adopted (i.e., ratified) the [Governor and DOJ's] motivations.”¹³⁵ This case helps demonstrate the flaws in Judge Brown's analysis, and I thank him for pointing it out.

Assuming that Common Cause represents a proper reading of the law in this circuit, Judge Brown does not provide evidence that the majority of the Legislature shared and purposefully adopted the Governor's and DOJ's motivation. Instead, the Judge Brown collects statements from a handful of representatives and then fails to explicitly assert that the majority of the legislature specifically acted to ratify the underlying conduct. Instead Judge Brown spends much of his direct evidence section talking about the secretive plan that was hatched between Hunter and his co-authors but fails to make any credible connection to the intent of the majority of the Legislature as is necessary in Common Cause. Judge Brown's only attempted connection is that the Legislature “fulfilled almost everything that DOJ and the Governor desired.”¹³⁶ As will be demonstrated shortly, this claim is simply untrue. Under Judge Brown's own rubric, the DOJ Letter and the statements of Governor Abbott are not direct evidence that race was the “ ‘predominant factor motivating the legislature's decision to place a significant number of voters within or without a particular district.’ ”¹³⁷

Even if Judge Brown decided to use the standard for direct evidence that was given in Alexander, neither the DOJ letter nor Governor Abbott's statements are direct evidence. Direct evidence “comes in the form of a relevant state actor's express acknowledgment that race played a role in the drawing of district lines.”¹³⁸ The logical implication of this description of direct evidence is that direct evidence needs to come from a state actor who has control over the drawing of district lines. Here, Judge Brown provides no evidence, and the record provides minimal support for the prospect that Governor Abbott or the DOJ actually controlled the drawing of district lines in any way.

* * * * *

Turning now to the “indirect evidence,” mainly developed by the experts' statistical analysis, Judge Brown gets things woefully off-base.

First, and importantly, Judge Brown studiously avoids any reference to Dr. Barreto, despite the plaintiffs' heavy reliance on him in their post-trial brief, LULAC Post-Hearing Brief at 25, 33. Judge Brown also fails to make any reference to Dr. Murray, Dr. Ansolabehere, or Dr. Ely, apparently abandoning days of expert testimony developed in the hearing to grasp after straws.

Instead, Judge Brown depends exclusively on Dr. Duchin's analysis.¹³⁹ While Dr. Duchin may be a fine mathematician, she was demonstrably unaware of several of the redistricting criteria used by the State of Texas. Thus, she would likely be forced to admit that her analysis is statistically skewed. On a correct appraisal of her report for its substance—rather than merely being cowed into accepting her conclusions by her strong credentials ¹⁴⁰ —one will quickly realize that her report is so flawed as to be irrelevant at best and cunningly misleading at worst.

As to the role of an expert in a bench trial, normally “jurors are supposed to reach their conclusions on the basis of common sense, common understanding and fair beliefs, grounded one evidence consisting of direct statements by witnesses or proof of circumstances from which inferences can fairly be drawn.”¹⁴¹ Where a fact-finder needs to draw complex inferences, however, expert testimony is helpful.¹⁴² But expert testimony does not supplant the factfinding role; the Supreme Court has warned that even meritless expert testimony “can be both powerful and quite misleading because of the difficulty in evaluating it.”¹⁴³ While judges normally sit as gatekeepers of expert testimony, in a bench trial we are tasked with evaluating it.¹⁴⁴ Therefore, we should not hesitate in poking readily-observable holes in expert testimony—precisely as the Supreme Court did in Alexander—with this exact expert witness.

In Alexander, plaintiffs challenged redistricting around the city of Charleston, South Carolina, for racial vote dilution.¹⁴⁵ The Supreme Court faulted Dr. Duchin's vote dilution analysis for failing to account for partisanship or core retention metrics.¹⁴⁶ It also faulted, in the vote-dilution context, Dr. Duchin's report for conducting a statewide draw rather than attending to the particular district at issue to identify whether the map cracked or packed it: “Dr. Duchin's conclusion was based on an assessment of the map as a whole rather than District 1 in particular. A state-wide analysis cannot show that District 1 was drawn based on race.”¹⁴⁷

Although her analysis was primarily directed toward claims of racial vote dilution, Duchin had three steps in her analysis relevant to a claim of racial gerrymandering.¹⁴⁸ First, she conducted a compactness analysis of the 2025 maps compared to the 2021 maps, and the 2021 maps compared to the 2012 maps.¹⁴⁹ Second, she generated ensembles of hypothetical maps across metro-area “clusters,” which were defined as all the territory included in the C2333 districts that touched Travis/Bexar counties (San Antonio), Dallas/Tarrant counties (Dallas and Fort Worth), and Harris/Fort Bend (greater Houston).¹⁵⁰ These maps were the results of random walks and spanning trees mapping out possible permutations within the defined areas.¹⁵¹ Third, she conducted a “winnowing” process, or adjustment of the simulation results from the second step, by applying her choice of “filters” including (i) Republican performance across the cluster, (ii) Trump performance, (iii) urban/rural composition, and (iv) a cap on incumbent double-bunking.¹⁵² The results of these latter two steps are offered at Duchin Report 14-15. (“the histograms”).¹⁵³

While Duchin's analysis is an interesting simulation, it contains several internal ¹⁵⁴ and external threats to validity.¹⁵⁵ Her report also contains several weaknesses in presentation—such as inadequately labelled histograms that we nevertheless do our game best to interpret, but which arguably fail the burden of production on the plaintiffs' side.¹⁵⁶ She also offers several conclusory leaps toward assuming intent.¹⁵⁷ While expert witnesses are welcome to opine on ultimate issues, in a bench trial this cuts both ways, where judges are then responsible for scrutinizing the conclusions advanced by an expert.¹⁵⁸

First, on compactness, the 2025 maps scored better on every measure in Duchin's own analysis, supporting a soft inference that traditional redistricting criteria were used.¹⁵⁹ One might say that Texas has a little less ‘mander’ to its ‘gerry.’

Further, Duchin's analysis of precinct splits was completely rebutted by Adam Kincaid's testimony.¹⁶⁰ Duchin announced the conclusion that “the state has not disclosed the use of any partisan data below the precinct level, while race data is available at the block level [so that] the high number of precinct splits․ is more indicative of a focus on race than on partisanship.”¹⁶¹ However, in Kincaid's testimony, he reveals the State's use of commercially available and State-provided partisan data available below the precinct level, directly undermining Duchin's conclusions on compactness and precinct splitting.¹⁶²

Second, I recognize that Dr. Duchin attempted to improve her analysis from Alexander by including partisanship and core retention weights in her map-drawing algorithm. However, several problems emerge. One is that she does not include the same partisanship constraints as those used by the map-drawer. Unlike Duchin's blanket 55-Republican metric, in real life Kincaid had included constraints to reflect that (i) any Republican in a greater than 60-R district could not be reduced below 60, (ii) any Republican in a below-60-R district had to be kept constant or improved, and (iii) any newly drawn districts were to be as Trump-favorable as possible while also winning Ted Cruz the senate seat, beginning at the 10% margin.¹⁶³ ( Judge Brown hand-waves past this concern, stating “The State Defendants have․ failed to persuade us that Dr. Duchin's 55% figure is off the mark,” while failing to recognize that this departure likely skews Duchin's outputs.¹⁶⁴ )

Duchin also conducted only metro-area or cluster-wide draws, rather than any state-wide draw, whereas we know that Kincaid conducted a statewide draw beginning in the northwestern corner of the state, rather than conducting metro-area or cluster-wide draws.¹⁶⁵ Therefore, the knock-on effects from one district affecting another may significantly affect the range of results included in the simulation outputs.¹⁶⁶ Each of her clusters includes a significant number of surrounding counties outside the metropolitan core of Dallas-Fort Worth, Houston, and San Antonio.¹⁶⁷ However, this underrepresents the degrees of freedom available to Kincaid—we know from his testimony that he drew eastern counties into CD-32 to make it perform for Republicans, but he likely had numerous other options available across rural, Republican-performing counties generally in the Dallas-Fort Worth area, such that constraining the map-drawing space to only the counties actually chosen underrepresents the available space and constrains the output of the ensemble.¹⁶⁸ Where Duchin then complains that the actual outcomes are outliers, that may be an artifact of her flawed map-drawing process.¹⁶⁹

A statewide map draw, rather than one localized to 7 or 8 congressional districts in the Harris and Dallas-Tarrant County metros, will necessarily have greater variance. But Dr. Duchin concedes that she limited her analysis only to the subsets of those metro areas, thereby hacking a lower variance figure that ultimately excludes the final outcomes.¹⁷⁰

While Duchin was criticized in Alexander for not analyzing a particular district for vote dilution purposes, Judge Brown uses her analysis to support a racial gerrymandering claim that depends on statewide statistics.¹⁷¹ Therefore, unfortunately, the opposite criticism carries water: that she failed to conduct a statewide draw that would fully capture the range of possible outcomes in the ensemble.

If Judge Brown had pursued a vote-dilution theory, the relevant interpretation of Duchin's analysis might differ. But here, her analysis is clearly flawed by constraining the space within which the spanning trees could generate sets of possible maps. She should have realized that the metro constraint foreseeably manipulates the variance in her derivative statistics in a way that favors her preferred outcome.

Third, Duchin's winnowing criteria did not accurately capture the possible distributions available to a state map-drawer, because she chose off-base and thereby skewing filters. When selecting a subset from a wider set, or even transforming a set entirely, using accurate winnowing criteria can affect the variance or the skew of your outcome. So, where she compares her adjusted sets (in orange) to the ultimate outcomes (in blue), the probative nature of her analysis is severely limited by the fact that she used off-base winnowing criteria.¹⁷² It is also worth pointing out that under Alexander, she needs to attend to particular districts—so it is legally insufficient for her to refer merely to some possible set of outcomes in black and orange without accounting for the actual outcomes in blue.¹⁷³ It is the plaintiffs' burden to establish which districts were cracked and/or packed, just as in Alexander it was the plaintiffs' burden to show that that specific Charleston district had been racially vote-diluted.

Dr. Duchin's generation and winnowing conditions explained in Appendix E indicate numerous loose ends.¹⁷⁴ For instance, leaving districts within 1% of population for Ensemble Generation does not exclude the possibility of splitting precincts at the census bloc level for the last mile.¹⁷⁵ Indeed, where districts total 766,987 leaves about 7,670 voters on the table for each district—allowing the variance in both directions actually doubles this to 15,340 potential swings, whereas the true maps were required to be strictly equi-populous. That distinction can warp the distribution in multiple ways—but the most logical inference is that the truly available sets were a more discrete or constrained set and therefore would look skewed relative to a set chosen on softer parameters. Considering that Duchin completely ignored independents, libertarians, and greens, when we are at the level of arguing about a few thousand voters, these ‘silent’ votes could be disruptive in years with stronger or weaker, e.g., libertarian performance or independent swings.

As to her implementation of core retention in the spanning trees with either a 0.1 or 0.2 surcharge for crossing counties, census-designated county subdivisions (natural communities of interest), or newly drawn districts, Duchin does nothing to suggest that this surcharge results in figures with equivalent core retention to the actual map, and therefore does nothing to suggest that her core retention weights resemble those actually used. If she used either lower or higher core retention rates, rather than deriving her core retention weights from real life, her departure could foreseeably skew her results. Indeed, we know that core retention was intentionally violated in CD-9, given Kincaid's testimony that he planned to pull one Republican-performing district out of Harris County. Her spanning-tree analysis completely fails to distinguish between core retention for Republican incumbents (which was favored) and core retention for Democrat districts (which was actively disfavored, as through targeting Greg Casar and Lloyd Doggett in Austin through substantially changing CD-35 and CD-37).

As for Duchin's partisan weightings, her partisan score lagged considerably, including elections from 2012,¹⁷⁶ whereas Kincaid's partisan shading principally incorporated President Trump's and Ted Cruz's recent performances.¹⁷⁷ Given the recent changes in Hispanic preferences for the Republican party in Texas, using a lagging indicator could foreseeably skew the distribution of ensemble maps away from recent changes and thereby falsely represent the actual maps as outliers.

Dr. Duchin's use of a 50.1% sharp cutoff for Republican wins on her simulated map was problematic ¹⁷⁸ and did not reflect the realities of the map-drawing process conducted by Kincaid, which aimed to provide far greater insulation.¹⁷⁹ Foreseeably, Dr. Duchin's ensemble likely included a bulk of sub-55 maps which drove statistical skew, at least in the original outputs, even if not in the adjusted outputs. On wider tails embracing 54%, or 53% Trump-performance benchmarks, the variance would predictably be wider because there are ‘more possible ways’ to draw permissible maps within that space. Fortifying in underperforming Republican incumbents such as Dan Crenshaw could also warp the map, she failed to account for wins above 51%, instead analyses win and loss at the 51% cutoff.

Keep in mind, the general goal in gerrymandering is win by a little, lose by a lot.

Further, Dr. Duchin leaves out at least three other constraints: current addresses of representatives' homes, keeping congressional offices within districts, and favoring natural geographic boundaries like highways and rivers. Duchin actually used outdated incumbent data ¹⁸⁰ —and while she claims that this had no effect, it can predictably have affected the skew and variance in generating thousands of maps, which discredits Judge Brown's “box-and-whiskers” histogram standard deviation interpretation.¹⁸¹ Dr. Duchin was aware that the Winnowing Condition incumbent addresses were out of date, and she has been requesting updated addresses for months from her own counsel. She knowingly conducted a flawed analysis, which would have skewed the maps in unpredictable ways—in particular by pushing the actual maps further toward outlier status. Also, the urban-rural winnowing condition forces the redrawn CDs 9 and 32 to face strong outlier conditions, given the constraint down to only 10% swing, and the substantial relocation of those districts across Harris and Dallas counties respectively.¹⁸²

Another more arcane statistical feature that likely reduces variance is the set transformation involved in her branching trees analysis, since that only moves the line between two districts at a time, excluding other permutations from its random walk.¹⁸³ The problem is that this takes the outer-boundary conditions as given and modulates down the variance, whereas the variance on unbounded line-drawing can be expected to be higher. This reflects the same principle developed supra as to the statewide draw, which is that a statistic on a statistic generally loses variance. For her to then fault the actual outcomes for being beyond her down-regulated variance tails may be in error.

Further, Dr. Duchin's conclusions derived from her ensemble analysis are misrepresented, even on her own terms. The correct interpretation, in the social science literature, of a distribution analysis such as Dr. Duchin's is that any outcome within a set number of standard deviations is not considered a statistically significant outlier. Her use of the 1st and 99th percentile cutoffs is slightly unusual, given that two standard deviations generally embrace the 95% of the central distributions, while three standard deviations embrace 99.7% of the distribution.¹⁸⁴ Even on merely two standard deviations (narrowing the tails at Duchin Report 14-15), however, her conclusions as to Travis/ Bexar Counties, suggesting that “patterns characteristic of packing and cracking․ are present in each of the three clusters,” is flatly untrue. The outcome needs to be an outlier to overcome the null hypothesis, which is that the map is normative and exhibits no evidence of cracking and packing. Therefore, as to the Travis/Bexar cluster, Dr. Duchin's analysis actually supports the opposite inference, which is that the maps were not racially gerrymandered, but instead were partisan draws.¹⁸⁵

This should have caused a dispassionate academic some pause. But Duchin plowed on. So next, the correct interpretation of Tarrant/Dallas suggests that one of the actually drawn districts (the sixth of eight in the series) was a statistically significant low outlier.¹⁸⁶ And for Harris/Fort Bend, four outlier districts were low, while one was high.¹⁸⁷

As for Dr. Duchin's conclusions elicited in testimony, she got more specific on the particular precincts. She asserts that you see residential splits by race across CD-18 and CD-7, but there is drift across in either direction of Whites and Hispanics.¹⁸⁸ In CD-29, there is a significant concentration of white voters, rebutting the claim that the census lines neatly follow paths of segregation.¹⁸⁹ Indeed, the CD-18-C2193 “Barbara Jordan” district in Harris County more clearly followed the Black population lines than the newly reconfigured CD-29, which has a lower Black CVAP but is overall more diverse.¹⁹⁰ For CD-9 vs CD-18, she also failed to contemplate the political protection of the “Barbara Jordan district.” As in Alexander, where the Supreme Court expressly approved South Carolina's protecting Jim Clyburn's seat with a sur-abundance of democrat voters, Duchin here again fails to disentangle race from politics by ignoring relevant political alternatives.¹⁹¹

On cross-examination, Duchin also expressly ruled out using commercially available datasets with partisan data at the house level to interpolate data below the voting precinct level,¹⁹² which Kincaid later discussed.¹⁹³ The State of Texas also likely had access to specific voter registration data, which it could have provided to the legislature.¹⁹⁴

* * * * *

All these holes having been poked, Judge Brown breathlessly wraps himself in Duchin's report.¹⁹⁵

Judge Brown fails to read the maps correctly, declaring that “the orange figures—which are the ones we're most interested in—represent the range of minority populations for each district in each randomly generated map.”¹⁹⁶

Not so simple.

What they actually represent is the adjusted or transformed set of maps after application of Duchin's winnowing criteria.¹⁹⁷ And it's not that these are a strict subset—they are really a transformation, given that they have different statistical features reflecting a different imputed underlying natural population.

Any statistic tries to capture, from a black box a priori condition or null hypothesis, the truth of an underlying population. The correct interpretation of the black boxes on the histogram is as representing the 50th percentile, 25th and 75th (box), and then 1st and 99th (whiskers) cutoffs on the ensembles of maps in terms of their expected all-minority “POC” CVAP composition. The orange boxes represent the ensembles after transformation.

So, when Judge Brown next says “the district with the lowest minority population in the Dallas/Fort Worth area had a minority percentage somewhere between 26% and 41%,” he is a bit off.¹⁹⁸ What that orange figure shows is that after applying the winnowing conditions, the 1st percentile of maps started around 26% POC CVAP, and the 99th percentile of maps started around 41% POC CVAP. On 40,000 maps, this means 400 were outside of the range in each direction, for 800 total.¹⁹⁹

Where Judge Brown concludes “if a dot falls outside the box but within the ‘whiskers,’ that suggests that the enacted district's minority population is on the outer edge of what we'd expect if the Legislature were relying exclusively on partisanship and other race-neutral considerations,” he gets it subtly wrong.²⁰⁰ A C2333 outcome showing up in the simulation ‘whiskers’ means there is no outlying behavior identified at all. It cannot be said to be “on the outer edge,” it just means nothing relative to the null hypothesis.

Additionally, where Judge Brown states “If the dot falls outside the whiskers entirely, that suggests that none of the race-neutral maps that Dr. Duchin generated have the racial characteristics approximating that of the enacted district,” he is again without foundation.²⁰¹ It does not mean that none of them had that characterization—it means that less than 1% did, rendering it an outlier relative to the simulation's imputed target.

Where Judge Brown then analyzes the Houston and Dallas–Fort Worth cluster maps, he fails to account for any of the statistical phenomena discussed above, which may affect a sensitive calibration.²⁰² If you are targeting the wrong natural imputed population because of off-base inputs, your outputs will be off-base.

Additionally, where Judge Brown suggests that “those [patterns] in the Travis/Bexar County area․ are even less [stark], they nonetheless reinforce the conclusion that the enacted map is a statistical outlier,” he reveals his statistical naivety.²⁰³ What the Travis/Bexar cluster actually reveals is that there is no statistically significant outlier behavior—so this evidence actually cuts in the opposite direction and supports an inference of a partisan gerrymander.²⁰⁴

Judge Brown praises Duchin's “enormous number of maps”²⁰⁵ and her “tens of thousands of congressional maps.”²⁰⁶ But this is similarly clueless. It would not matter whether there was 1 map, 1 million maps, or 1 billion maps drawn, provided that the criteria used for drawing those maps were off-base.²⁰⁷ As a matter of probability theory, the underlying imputed natural population being sampled is not the same population as that which was actually sampled. A statewide map draw, rather than one localized to the 8-or-so congressional districts in the Harris and Dallas-Tarrant metros, will necessarily have greater variance. But Dr. Duchin concedes that she limited her analysis only to the subsets of those metro areas, thereby hacking a lower variance figure that ultimately excludes the final outcomes.

Moreover, where Judge Brown praises Duchin's consideration of partisanship, he reproduces her phrase that she “executed a run seeking to match the number of districts with Trump's 2024 major-party vote share over 55%” and achieved results consistent with her prior findings.²⁰⁸ But she offers no report of those robustness tests, which would have different variability, and instead presents the 50.1% cutoff figures, which may impact the skew of her distribution.

Plaintiffs have the burden of production of showing Duchin's robustness—merely calling something “consistent” does not mean that it showed statistically significant outlier variance, where consistency is an ambiguous term. After all, Duchin misrepresents the Travis/Bexar cluster as affirmatively showing evidence of cracking and packing where that shows nothing as a statistical matter. So, Judge Brown can hand-wave over Duchin's nonconformity with Kincaid's constraints, but Judge Brown has no rational basis to reject the idea that “Dr. Duchin's 55% figure is off the mark.” He just does not know. And in the very next breath, he inverts the burden of proof, “if raising the floor to a value closer to 60% would have undermined Dr. Duchin's conclusions, the State Defendants could have introduced expert rebuttal testimony to that effect.”²⁰⁹

Exactly the same logical errors apply where Judge Brown hand-waves away the internal threat to the validity of Duchin's favoring core retention for both Democrats and Republicans (rather than Republicans only), and using an out-of-date list of incumbent addresses.²¹⁰ Judge Brown inverts the burden of proof and claims to know things he just cannot know from this record.

Duchin failed to prepare any state-wide Alexander map, which certainly would have included wider variance figures, that in turn may plausibly have included the truly-chosen districts within two standard deviations of the normative draw.²¹¹ But Judge Brown hand-waves away this issue as well.²¹² In particular, his drawing a favorable inference from the absence of a map is somewhat absurd.²¹³

It would have actually been quite easy for Duchin, Barreto, or any other expert to draw Alexander map(s) based on a statewide draw using the same software. Therefore, plaintiffs' failure to muster such a map supports a negative inference against them, where that negative inference would be that statewide draws include the actual maps within two-standard deviations of their statistical tails—and for that reason, plaintiffs studiously avoided producing any statewide maps or derivative statistical figures. So, I do not assert that it is impossible to draw an Alexander map—I just find it damaging that plaintiffs failed to muster one when mustering one would be so easy, and from which one may infer that mustering one would potentially have been more damaging than cherry-picking by metro area.

There is something wrong with this picture.

Moon Duchin contends that she could run “a million maps in a matter of seconds” on a digital watch and have her robot execute a hundred thousand simulations in about an hour.²¹⁴ Yet neither plaintiffs nor their experts produce a single Alexander map.

Let's think of this in the context of an on-off switch.

Suppose the switch is turned off, and plaintiffs cannot produce an Alexander map that achieves the same partisan mapmaking criteria and greater racial “balance.” It strains credulity to suggest that they should be given a pass because the experts “didn't have time”²¹⁵ when “[a]ny expert armed with a computer ‘can easily churn out redistricting maps that control for any number of specified criteria.’ ”²¹⁶ Dr. Duchin's digital watch (the same one she claims can run a million maps in seconds) is more than capable. Plaintiffs want the extraordinary and drastic remedy of enjoining the 2025 Texas Congressional Map, so it is their burden to clearly show that they are entitled to such drastic, equitable relief. Bearing this in mind, it is it is highly inappropriate, in light of the weight of the procedural and substantive case law, for Judge Brown to give plaintiffs a pass and suggest that timing is the issue.²¹⁷

The real issue is that Judge Brown is embarking on a results-oriented crusade against the Texas Legislature. On his misguided journey, Judge Brown does not bat an eye, improperly bestowing unearned deference to supposed “experts” such as Duchin, while conveniently omitting discussion of other “experts”²¹⁸ such as Matt Barreto, whose testimony is so problematic that it is unusable.²¹⁹ Yet Judge Brown has no problem tossing the longstanding presumption of legislative good faith straight into the trashcan, as if the presumption of legislative good faith were a relic of a bygone era. Judge Brown pretends to know better—and to prove it, he is willing to contort himself into an illogical straitjacket. He cannot escape.

Now, let's consider a more nefarious scenario.

Suppose the switch is turned on, and plaintiffs or their purported “experts” could have produced an Alexander map. The fact that they did not file an alternative map curing the alleged discriminatory infirmity (the one they purport to care about) tells you all that the instant case is about—partisan gain. Duchin makes no bones about this, either. She, made it clear that she would not hazard to draw an alternative map, despite her extensive experience drawing maps in other states, because partisan gerrymandering is not her “motivating influence.”²²⁰

But Duchin may have a motivating influence.

Her CV gives us a clue.²²¹ Duchin notes that her amicus brief ²²² was cited in the Rucho dissent.²²³ Partisan gerrymandering may be her main problem.²²⁴ She had her chance in Rucho. Her brief was not persuasive enough to convince the Court to rule the other way.²²⁵

Rucho is not the only case where Duchin wishes the Supreme Court ruled differently or found otherwise. The Court noted, in Allen v. Milligan, that “Duchin's maps were based on old census data—from 2010 and 2020—and ignored certain traditional criteria, such as keeping together communities of interest, political subdivisions, or municipalities.”^{226  227}

There's more.

A few years after Rucho, she retooled her conclusion in Alexander, to say “that it is ‘not plausible’ that the dilution was a mere ‘side effect of partisan concerns’ ”.²²⁸ The Supreme Court kept with tradition—it discredited Duchin for “good reason” because “various parts of Dr. Duchin's report did not account for partisanship or core retention.”²²⁹ The Court could have stopped there, but it didn't: “Moreover, Dr. Duchin's conclusion was based on an assessment of the map as a whole rather than District 1 in particular. A state-wide analysis cannot show that District 1 was drawn based on race.”²³⁰ The Court continued: “Given these serious problems, it is no wonder that the challengers cite Dr. Duchin's report only in support of their racial vote-dilution claim. It has no probative force with respect to their racial-gerrymandering claim.”²³¹

Notice the pattern.

To his credit, Judge Brown does point out how Duchin was discredited in Alexander. But he has no choice but to do so.²³² Her flawed methodology is so patently obvious in a case that was routinely cited in briefs and subject to great discussion at the preliminary injunction that even Judge Brown cannot escape this reality.

But merely acknowledging the truth would be an exercise unfamiliar to Judge Brown. Instead, he can't help himself. Judge Brown gives “extra credit” to Duchin for turning in the assignment the Supreme Court gave her in Alexander. I have news for Judge Brown. She turned it in late. But I am not surprised that Judge Brown is an easy grader—the lawyers in Petteway can tell you all about it. Judge Brown also uses the same exam every year, so it's easy to get an excellent grade in his class, especially if you've taken a class or two with him before.

Whether Judge Brown likes it, gravity exists. So does Alexander.

Article III courts have a solemn responsibility, especially in bench trials, to assess expert reports for what they are actually arguing and the substance of their statistical claims, rather than merely being impressed by credentials. Where an expert report fails to show anything, by virtue of its internal threats to validity and external threats to validity, it is judicial aggrandizement to leap across the bench to save an infirm expert report.

Put plainly in the light of day, given Judge Brown's lack of statistical foundation, Duchin's analysis is irrelevant. Plaintiffs fail as a matter of law to disentangle race and politics as is required under Alexander.

I dissent.

* * * * *

Relatedly, the plaintiffs' own supplemental PI briefing shows the importance of statewide changes in map drawing. While detailing the history of the editions from C2308 to C2331 and finally C2333—with which Judge Brown neglects to grapple—plaintiffs concede that the changes from C2331 to C2333 not only moved Liberty County (population approximately 115,000 compared to total districts of 766,987 persons), but then sliced off the top of C2331-CD9 and put that back into C2331-CD2, around Lake Houston and Huffman, which ultimately had knock on effects in the 36th, 14th, 10th, and 17th. So, the variance induced by these changes—where the only unchanged district statewide was in the 19th based around Lubbock—needs to be accounted for by both Judge Brown and Dr. Duchin.

As for Judge Brown's much-ballyhooed 49 and low-50 series numbers, Judge Brown makes zero effort to challenge or even discuss the prevalence of near-50 cutoffs in the opposite direction: indeed, would they suggest it is an intentional racial gerrymander when the legislature drew C2333 CD-8, a district west and north of Houston, at 49.3% Anglo?²³³ Conversely, plaintiffs also have little to say about CD-23, covering the Western reaches of the Rio Grande, but which is already held in Republican hands. Plainly, they only dispute near-50 cutoffs where those affect elected Democrats' chances in the next election—which gives away the goose that what offends plaintiffs is not racial injury, but partisan targeting permitted under Rucho. Nor do they identify any problems with CDs 6, 12, 14, or 25, even though those all enjoy top-line low-50s and high-49s in their relative Anglo and non-Anglo compositions. But that is because each of these districts is held by a Republican either equally advantaged or further fortified by the C2333 2025 maps, per Kincaid's undisputed map-drawing constraints.

Further, all of these top-line high 49 and low 50-51 figures reflect the statistical trend in Texas that the Black and especially Hispanic populations are younger than the White population, meaning that a district can have a 49.5 and 50.5 racial percentage split while enjoying the 6- or 7-point partisan percentage margin that a Republican-maximizing map-drawer is seeking to achieve, on account of to the differently shaped population age pyramids.²³⁴ This race-neutral explanation more plausibly explains the overall trend in the data, including as applied to CD-32 and CD-9, rather than the cherrypicked explanation preferred by plaintiffs, which fails to rationally account for and explain the overall trend in the data. This kind of statistical hacking, analogous to p-hacking in random control trials, should not escape Judge Brown's notice, apart from his motivated reasoning.

Tellingly, Judge Brown also avoids revisiting other expert testimony from Dr. Barreto and even Dr. Duchin on Ecological Inference, or deriving district-level racial voting preferences from statewide averages. That is because the ecological inference data suggested that, while Hispanic voters overall favor the Democratic party, there has been a breakdown of support in recent years as the Hispanic community becomes more diverse and more Trump-supporting.²³⁵ Tejanos in the Rio Grande Valley turned strongly for Trump, while Cuban and Venezuelan recent arrivals are more Republican-leaning than Mexican recent arrivals.

Relatedly, Judge Brown avoids discussing any of these inconvenient developments because he explains that 2 of the 5 newly-drawn Republican pickup districts are in the Rio Grande Valley and stand to elect Hispanic-supported Republicans to Congress in the next election. Indeed, the entire preliminary injunction hearing carefully danced around discussion of the 28th and 34th districts, even as those were material to the Republican gains disputed under HB4.²³⁶ That should strike the judges as a conspicuous omission and should support the negative inference that those areas' redistricting resists statistical sniping as racial gerrymandering. In fact, that is direct evidence cutting against racial gerrymandering that reinforces the strong positive inference of good-faith legislative intent under Alexander.

I dissent.

* * * * *

Beyond all of this analysis of the facts, the most egregious shortcoming of the opinion is its treatment of the presumption of legislative good faith. To be sure, Judge Brown pays ample lip service to the presumption, but the presumption is quite strong and can't easily be overcome. As a matter of fact, the presumption is so strong that it “directs district courts to draw the inference that cuts in the legislature's favor when confronted with evidence that could plausibly support multiple conclusions.”²³⁷ After running through all of his proposed evidence, Judge Brown concludes that the “Chairman Hunter and the other joint authors evidently strategized that a map that eliminated coalition districts and increased the number of majority-Hispanic and majority-Black districts would be more ‘sellable’ than a nakedly partisan map”²³⁸

Unfortunately for Judge Brown, overcoming the presumption of legislative good faith requires a stronger conclusion than race “evidently” guided the drawing of map lines, even at this preliminary stage. By implication, overcoming the presumption appears to require that the evidence be able to support no other conclusion.²³⁹ Here, the evidence can and does support alternate theories, including theories that make far more sense than Judge Brown's reading of the tea leaves. The most straightforward read of the facts is simple: The legislature had no real concern for Petteway and Representative Hunter and the handful of House members Judge Brown relies on were paying lip service to it in order to avoid talking about partisan gerrymandering. This conclusion is distinct from the far more involved and technical theory that Representative Hunter conspired with a man he never talked to,²⁴⁰ on a map that was being drawn before he was asked to carry the bill,²⁴¹ to create a map that doesn't even do everything the DOJ letter requests.

Judge Brown handwaves the fact that Kincaid's map doesn't do everything the DOJ letter requests because “it's entirely possible for the Legislature to gerrymander one district without gerrymandering another.”²⁴² This misses the mark. The problem with the map leaving a coalition district intact, as expressed earlier, is that it undermines Judge Brown's theory of the facts. Why would a legislature, conspiring to use the elimination of coalition districts as a cover for partisan gain, leave a coalition district called out by the DOJ letter in place? If race were the criterion “that, in the State's view, could not be compromised in the drawing of district lines” as part of a statewide scheme, why was it compromised in this district?

Judge Brown offers no plausible justification for this anomaly and fails to consider it when trying to discern if the presumption of legislative good faith is overcome. Such information supports the far more modest proposition that the few representatives that Judge Brown is able to point to were discussing Petteway pretextually in order to limit the focus on partisan gerrymandering, especially considering its unpopularity of the practice in the state and nationwide.²⁴³ It also supports the inference that the three districts reaching just over 50% could, in fact, be a coincidence or byproduct of the partisan line-drawing in areas where race and partisanship are highly correlated, especially since Judge Brown fails to provide competent evidence disentangling race from politics.²⁴⁴ In the face of such evidence, the plaintiffs have not produced evidence to overcome the presumption of legislative good faith and thus cannot show even some likelihood of success on the merits.

I dissent.

* * * * *

In his remedial section, Judge Brown similarly hand-waves over thorny problems of remedies and the current status of the 2021 and 2025 maps.²⁴⁵

Texas's House Bill 4 (“HB4”), the statute at issue, provides:

(a) This Act supersedes all previous enactments or orders adopting congressional districts for the State of Texas. All previous acts of the legislature adopting congressional districts for the State of Texas are repealed.

(b) Chapter 7 (S.B. 6), Acts of the 87th Legislature, 3rd Called Session, 2021, is repealed.246

On a straightforward reading, this repeal provision in HB4 means that the 2021 maps were voided by the 2025 maps. Therefore, if the 2025 maps are enjoined, there can be no elections because there are no maps in place—contrary to the majority's attempt to revive the 2021 maps.

A federal court cannot reinstate a statute that the legislature has explicitly repealed and voided.²⁴⁷ That move presents grave federalism concerns, commandeers the state legislature,²⁴⁸ departs from the standard remedial process in voting rights cases, and intrudes into the “sensitive area of state legislative redistricting.”²⁴⁹ The default remedy, as Judge Brown admits, is that “the elected body must usually be afforded an adequate opportunity to enact revised districts before the federal court steps in to assume that authority.”²⁵⁰ But Judge Brown ignores the law and denies the state any opportunity to hold a special session to exercise its own legislative power.²⁵¹

Judge Brown also fails to grapple with the fact that the prior maps have been voided.²⁵² Texas law is clear: the Texas Code's subchapter on “construction rules for civil statutes” provides that “The repeal of a repealing statute does not revive the statute originally repealed.” Tex. Gov. Code § 312.007.²⁵³ At the time of writing, given that the law was passed on August 20 and signed into law on August 29, HB4 has been on the books for more than 75 days.

Properly understood, Judge Brown's remedy is a novel and unlawful order imposing a new map on Texas, in an activist echo of the overturned § 5 pre-clearance regime.²⁵⁴

Judge Brown embraces a dinosaur-like understanding of equitable remedies.

The up-to-date view of injunctive relief is that injunctions represent a court-ordered policy of nonenforcement restraining an executive from enforcing a federal or state law. As the Supreme Court recently instructed in Trump v. CASA, Inc., “traditionally, courts issued injunctions prohibiting executive officials from enforcing a challenged law or policy only against the plaintiffs in the lawsuit.”²⁵⁵

This restrained view is deeply rooted in equitable jurisprudence: In Ex Parte Young, the Supreme Court interpreted injunctions as stripping a state actor from enforcing a statute that remains on the books:

In every case where an official claims to be acting under the authority of the state․ [and] the act to be enforced is alleged to be unconstitutional; and if it be so, the use of the name of the state to enforce an unconstitutional act to the injury of complainants is a proceeding without the authority of, and one which does not affect, the state in its sovereign or governmental capacity. It is simply an illegal act upon the part of a state official in attempting, by the use of the name of the state, to enforce a legislative enactment which is void because unconstitutional. If the act which the state attorney general seeks to enforce be a violation of the Federal Constitution, the officer, in proceeding under such enactment, comes into conflict with the superior authority of that Constitution, and he is in that case stripped of his official or representative character and is subjected in his person to the consequences of his individual conduct.”256

In Grupo Mexicano de Desarrollo S.A. v. Alliance Bond Fund, Inc., the Supreme Court definitively stated, “[equitable] jurisdiction․ is an authority to administer in equity suits the principles of the system of judicial remedies which had been devised and was being administered by the English Court of Chancery at the time of the separation of the two countries,”²⁵⁷ and “the equitable powers conferred by the Judiciary Act of 1789 did not include the power to create remedies previously unknown to equity jurisprudence,”²⁵⁸ such that any enlargement of district courts' equitable power was properly left to congress.²⁵⁹

Most recently in CASA, the Court struck down universal injunctions for departing from the non-enforcement model and exceeding the “confin[es] of the broad boundaries of traditional relief,”²⁶⁰ and cautioned that “[w]hen a court concludes that the Executive Branch has acted unlawfully, the answer is not for the court to exceed its power, too.”²⁶¹ Judge Brown's command of the state legislature not only violates the Tenth Amendment—it likely exceeds the bounds of equity, too.

Injunctions in Texas take the same, restrained form.²⁶² The Supreme Court of Texas has written, “When a court declares a law unconstitutional, the law remains in place unless and until the body that enacted it repeals it, even though the government may no longer constitutionally enforce it.”²⁶³ And Texas appellate courts have noted the “Ordering the repeal of an ordinance would present grave separation-of-powers problems.”²⁶⁴ This strict separation-of-powers view prevents Texas state courts from ordering the repeal of a statute—which power is reserved to the legislature—and finely delineates between calling a law unconstitutional and technically voiding it.²⁶⁵

The other view of injunctions, more consistent with the law-declaration model of judicial review, is that courts recognize that a given law was truly unconstitutional from the moment of its inception, thereby insinuating that the legislature was without power to create it in the first place.²⁶⁶

This null-and-void, or ‘discernment,’ approach to injunctions sometimes crops up in state courts too, like Texas's recent Dickson v. Afiya Center case.²⁶⁷ Nevertheless, the weight of Texas law easily indicates that the effects of an injunction follow the first model.

Dickson was itself reversed on other grounds by the Supreme Court of Texas.²⁶⁸ Further, the Texas Constitution provides for separation of powers between the Executive, Legislative, and Judicial Departments, “no person ․ being of one of these departments, shall exercise any power properly attached to either of the others.”²⁶⁹ It also vests the entire legislative power of the state of Texas in its legislature.²⁷⁰ Admittedly, “there is an overlap in the functioning of the three different branches of government.”²⁷¹ Still, the division between Texas's legislative power and judicial powers appears to mirror that of the federal constitution.²⁷²

This second discernment approach is easily the incorrect view of the effect of injunctions. Otherwise, how could a law spring back into effect after a higher court vacates a lower court's injunction?²⁷³ A fine case-in-point is Citizens United v. FEC.²⁷⁴ After that decision, while the Supreme Court's controversial ruling prevents enforcement of the federal campaign finance statutes, those laws actually remain on the books and are ready-to-go should First Amendment jurisprudence evolve.²⁷⁵ As mentioned supra, the discernment approach has been cut back by newer Supreme Court jurisprudence.²⁷⁶

Here, applying the first, nonenforcement approach, the issuance of a federal injunction cannot reinstate the 2021 maps because Texas's state legislature retains its separate power to issue or repeal statutes, leaving the 2025 maps on the books but unenforceable. Yet by the issuance of this injunction, Judge Brown's free-floating Hegelian interpretation of the law undermines the legislature's ability—and thereby the people's ability—to make laws governing themselves.²⁷⁷ As Judge Learned Hand said, this is “irksome” rule by “a bevy of Platonic Guardians.”²⁷⁸

A federal court trying to reinstate a statute that the legislature has repealed may represent a limit on the equity power. A couple of recent election law cases are relevant. In Democratic Nat'l Comm. v. Wisconsin State Legislature, Justice Gorsuch wrote in concurrence, “[t]he Constitution provides that state legislatures—not federal judges, not state judges, not state governors, not other state officials—bear primary responsibility for setting election rules.”²⁷⁹ Justice Kavanaugh, likewise, has invoked the “principle of deference to state legislatures.”²⁸⁰ In Andino v. Middleton, reversing a lower court ruling invalidating South Carolina's witness requirement for absentee ballots, Justice Kavanaugh wrote, “a State legislature's decision either to keep or to make changes to election rules to address COVID–19 ordinarily “should not be subject to second-guessing by an ‘unelected federal judiciary,’ which lacks the background, competence, and expertise to assess public health and is not accountable to the people.”²⁸¹ Therefore, in addressing the diagonal separation of powers between federal courts and state legislatures, strict separation of powers, deference, and comity apply.

The bottom line is this: first, Judge Brown must permit redrawing rather than imposing his own map,²⁸² and second, it may violate separation of powers and exceed the equitable power for a court to order the legislature to reinstate a voided statute, contrary to Texas's anti-repealer statute, and to order the State executive to administer that voided statute. Judge Brown's remedy is unlawful judicial aggrandizement.

I dissent.

* * * * *

Also, Judge Brown's chosen remedy engenders an interesting contradiction: The plaintiffs have insisted, for years, that the 2021 maps are themselves racist and unconstitutional. While Judge Brown's opinion exactly what they asked for, it is manifestly absurd for them to mandate an unconstitutional set of 2021 maps!²⁸³ The Judiciary Act of 1789 authorizes courts to hear suits in equity ²⁸⁴ —but it plainly exceeds that statutory authorization to issue an unconstitutional injunction.²⁸⁵

Is Judge Brown now saying, sotto voce, that the 2021 maps are affirmatively constitutional? He must be, given that it would be without the Article III power to order a racist injunction. This stance then credits Chairwoman Huffman's statements from the spring trial that the 2021 maps were drawn race-blind.

Again, if they were drawn in a racist manner, then Judge Brown's order would itself be unconstitutional, exceeding the Article III power and Judiciary Act of 1789 authorizing equitable relief. And Judge Brown cannot issue an unconstitutional order, as he knows well through his related reversal in Petteway.²⁸⁶

Yet this conclusion also unearths another contradiction in Judge Brown's reasoning: If Huffman was right last Spring that the 2021 maps were drawn race-blind, permitting them as a remedy in this case, that then enhances the likelihood that the 2025 maps, drawn by the same map drawer in Mr. Kincaid, were drawn with the same criteria. Judge Brown's attack on Kincaid's credibility should thereby implode, given that he credits the Texas legislature's use of partisan intensity in 2021.²⁸⁷ Judge Brown seems to acknowledge, at some level, that this preliminary injunction is merely the latest round in a multi-decade partisan struggle, rather than a one-time isolated episode beginning in July 2025 with the Governor's legislative call. Otherwise, how could Judge Brown approve less-partisan-gerrymandered maps from 2021, while necessarily affirming their constitutionality? Here, picking and choosing between partisan maps of different intensity nakedly defies Rucho's rule on the non-justiciability of partisan gerrymandering as a political question.

As mentioned supra, this court's intrusion into bare-fist partisan politics is particularly concerning where other states are redistricting in real time.²⁸⁸ Injunctions have a major trickle-down—indeed, the 2012 injunctions likely affected Lt. Gov. Dewhurst's and Sen. Cruz's electoral outcomes.²⁸⁹ Rucho is clear: federal courts do not pick partisan winners and losers—they uphold the constitution.

I dissent.

* * * * *

This injunction flies badly in the face of the Purcell principle, especially in light of the Supreme Court's stay of the injunction in Merrill v. Milligan.²⁹⁰ The Purcell principle reflects “a bedrock tenet of election law: When an election is close at hand, the rules of the road must be clear and settled.”²⁹¹ The principle also reflects judicial restraint so as not to interfere with the democratic process.²⁹² To reiterate, it represents a policy of judicial restraint, as distinguished from judicial activism and meddling: The legislature, with its democratic accountability, has greater authority to intervene and regulate the rules of elections as election deadlines approach.²⁹³

Judge Brown's approach to Purcell is judicial aggrandizement, plain and simple. Quite contrary to the presumption of legislative good faith that's supposed to undergird the judiciary's approach to these sensitive legislative questions, Judge Brown's opinion is shot through with a presumption of legislative bad faith.

The opinion raises the specter of the legislature's being incentivized to redistrict “as close to elections as possible.”²⁹⁴ The opinion assumes that legislatures are often out to break the law when they redistrict and that it is the noble and just court who must always have the opportunity to step in and remedy this wrong, no matter how close to the election that this change has been made by the legislature. Judge Brown seems to miss that legislatures' being able to intervene later in the election cycle than the judiciary is a feature, not a bug, of the Purcell principle and reflects the different roles played by the courts as distinguished from the legislature.²⁹⁵

Judge Brown's inventive reasoning effectively mutilates Purcell. He goes so far as to state that “Purcell cannot be read to gut the Plaintiff Groups' right to seek a preliminary injunction and this Court's obligation to award one when merited.”²⁹⁶ But what purpose does Purcell serve but to deny injunctive relief that might, hypothetically, be merited and to do so because of the proximity to an election? If injunctive relief were not merited, the court would deny such relief, or the injunction would be vacated, on appeal on non-Purcell grounds. Purcell exists for those situations where injunctive relief may, in fact, be otherwise warranted but inappropriate considering the timing of the election.²⁹⁷

Judge Brown's notion of Purcell is that it exists almost exclusively to prevent plaintiffs from bringing challenges on the eve of the election. Judge Brown faults the legislature for making a late-breaking change to election law and essentially claims that Purcell can't apply if the legislature causes an injunction to be on the eve of an election. This subordinates the legislature and exalts the judiciary and is counter to the principle of judicial restraint that undergirds Purcell.²⁹⁸

A comparison between both the facts and the timeline of Milligan will demonstrate how clear the Purcell issue is in this case. The primary election here closer than was the primary in Milligan. When the district court issued the preliminary injunction on January 24, the primary election process began via absentee voting sixty-six days later on March 30. In this case, Judge Brown took well over a month to issue his opinion, leaving the state of Texas with around 60 days until absentee voting begins by ballots' being sent overseas.

Judge Brown wishes to rest much of its confidence on the fact that the 2021 maps could be used in place of the 2025 map, but those maps are no longer Texas law. The 2025 bill repealed the 2021 maps for the 2026 election, and, importantly, Texas has an anti-repealer statute, meaning that even if the act were enjoined or otherwise repealed, the repealed 2021 maps cannot spring back into life.²⁹⁹ It is noteworthy that Judge Brown does not cite a single example in which a previously enacted map has been brought back from the dead by the court's enjoining a bill or by pure judicial fiat. Furthermore, both the Supreme Court and the Fifth Circuit have made it clear that the only two options for relief are judicially crafting a map or letting the legislature work:

[T]he Court has repeatedly held that redistricting and reapportioning legislative bodies is a legislative task which the courts should make every effort not to preempt. When a federal court declares an existing apportionment scheme unconstitutional, it is therefore, appropriate, whenever practicable, to afford a reasonable opportunity for the legislature to meet constitutional requirements by adopting a substitute measure rather than for the federal court to devise and order into effect its own plan.300

It's actually unclear whether Judge Brown mistakenly believes the 2021 maps are still in effect for the 2026 elections or if, instead, he wishes to foist an alternate, judicially created, 2021 map on Texas. Under either theory, a fatal Purcell problem obviously remains.

If Judge Brown believes that the 2021 maps are still on the books in Texas, he is sorely mistaken, as discussed in the repealer section of this dissent. Under this read of Judge Brown's opinion, that means the Texas Legislature must be reconvened in a special session in order to redraw the maps.³⁰¹ The court should afford the legislature at least an opportunity to do this regardless, as the Supreme Court has clearly stated,³⁰² but it would be necessary if the ruling of the court orders Texas to follow a repealed law.³⁰³

Judge Brown's contrary assertion—that such is not necessary on account of the 2021 map's being a “viable congressional map that was drawn by the legislature”—ignores the obvious fact that the legislature repealed the map.³⁰⁴ To place that map back in place, the court must be imposing it on the state. The fact that the legislature at one point preferred these lines does not change the fact that they no longer preferred those lines and that they are an imposition on the legislature's authority. If anything, this represents a more odious form of imposition because it involves a map that the legislature has consciously decided to reject.

It should go without saying that the state of affairs that Judge Brown creates on these grounds is more severe than the situation in Milligan, where the district court required an Alabama legislature that was already in the midst of its regular session to redraw its maps.³⁰⁵ The Governor will have to issue a new call, the legislature will have to reconvene, and any hearings will necessarily be truncated and minimal because the filing deadline for candidates, which is fixed statutorily, is on December 8.³⁰⁶ This court is rendering its decision closer to the primary than in Milligan, with a legislature that is out of session, with less than a month before the close of the filing deadline and only two months before the first primary ballots go out to service members as required by federal law. Forcing the state to adjust to a new map would be setting the stage for bedlam beyond even the facts of Milligan. Scarcely more should need to be said to indicate the depth of the Purcell problem on this version of the facts.

Even assuming that Judge Brown were able magically to bring the 2021 map back into being through judicial fiat, the Purcell problem remains. While it is true that the type of relief and the ease in which the state can make the change without undue collateral effects impact “how close to an election is too close,” reversion to the 2021 map by no means resolves the Purcell dilemma.³⁰⁷ An injunction and reversion to the 2021 map now threatens to create voter confusion, disadvantage cash poor candidates, and threaten the tight schedule of election deadlines in the state of Texas.

Christina Adkins, the Director of Elections for the Texas Secretary of State's office, provided ample testimony about the structure of Texas elections and how a reversion to the 2021 election map would sow confusion amongst the voters and harm the integrity of Texas's election process, which is a complex web of statutorily set deadlines and deadlines keyed to the date of the election.³⁰⁸

As previously mentioned, the filing period for candidates seeking public office runs from November 8 to December 8, 2025.³⁰⁹ Candidates can file to run for office only if they pay a filing fee or submit a petition in lieu of that fee.³¹⁰ The petition for congressional candidates requires 500 signatures from individuals who live in the congressional district.³¹¹ Many candidates choose to submit petitions in lieu of paying the filing fee both to avoid the “heftier” filing fee and to introduce themselves to voters.³¹² After the filing deadline, political party chairs enter candidate information into the candidate filing system, which takes several days.³¹³ After this, the counties must perform ballot draws and begin preparing ballots, which takes approximately three weeks.³¹⁴

All of this must be done before January 17, 2026, to comply with federal law.³¹⁵ Any waiver of that requirement at the federal level would require the state to create a “comprehensive plan to ensure that absent uniformed service voters and overseas voters” are able to both receive, submit their ballots in time to be counted in the election, and receive approval from the President, meaning that moving the federal deadline likely provides the state with little flexibility.³¹⁶

With the context of this complex web of interactions laid bare, Ms. Adkins testified that any change in election policy, including this injunction, would be “harder on candidates, harder on voters, [and] harder on election officials.”³¹⁷ Ms. Adkins emphasized that there's “not much time to play with,” and that delaying the opening of the filing period (and presumably extending the filing period) would threaten the ability of the counties to adequately prepare and test their ballots, thwarting the ability of the state to tabulate election results accurately.³¹⁸ Additionally, many of the counties have already began redrawing county election voter registration precincts, rendering all of that work useless.³¹⁹ Furthermore, candidates had already begun to campaign and collect signatures under the 2025 map when Ms. Adkins offered her original testimony, meeting voters and spreading their name amongst the new congressional district.³²⁰

Several weeks later and this has likely only gotten worse.

Many of these candidates will be shuffled between districts, and voters may not become aware of that fact until they enter the voting booth. In addition to the voter confusion, reverting the maps now means that some of those candidates will need to run in different districts, needing up to 500 new signatures if they need to get onto the ballot via petition. This seriously disadvantages outsider political candidate who are likely to have less money to dole out for filing fees. Furthermore, it has been reported that changing the map “could force candidates who have already filed or are considering entering the race to rethink their plans,” meaning court intervention will fundamentally alter the state of these ongoing races.³²¹

As a legal and practical matter, Judge Brown's injunction turns the Texas electoral and political landscape upside down. It creates mayhem, chaos, misinformation, and confusion. Certain statutory election deadlines for the 2026 cycle kicked in in September 2025. Candidates began filing for federal and state office beginning on the statutory launch date of November 8, 2025.

The prevailing expectation is that the 2025 congressional lines will be used to elect representatives in 2026. There is, of course, a trickle-down effect. Some incumbents have announced their retirements because of the new lines. Some have announced they will run in different districts. Officials holding other or “lower” or local offices have declared as candidates for Congress, meaning that other citizens have decided to run to replace them.

Lastly, Judge Brown claims that Ms. Adkins testified that “Texas election officials and systems are more than capable of proceeding with the 2026 congressional election under any map that is the law.”³²²

This is a blatant misstatement, to put it politely.

The passage that Judge Brown highlights actually says:

In all of our interactions with the counties, we have been reiterating that these [2025] maps are the maps that are in place for the primary. Unless there is something, a court order or something telling us otherwise, we have to proceed and move forward with the maps that are law, that will be law.323

Nowhere does Ms. Adkins indicate that the Texas is “more than capable” of proceeding under any map that's law, nor does she imply that. Rather, her statement represents the admirable but mundane proposition that Texas will do everything in its power to comply with the law under either map. Judge Brown's misrepresentation of this fact makes it clear, once again, that he is motivated by results, not a sound application of law to facts.

The concerns about timeline, voter confusion, and chaos for political candidates ring true here, just as they did with Milligan, even if it were possible to return to the 2021 map. As it was for Alabama in Milligan, the filling deadline is imminent and candidates who have already been campaigning will be shuffled between districts, meaning new petitions, a new voter base, and confusion about the options the voters have come March.³²⁴

Likewise, the minimal wiggle-room in Texas's statutorily mandated elections process means that Texas is faced with an impossible dilemma should this injunction go through, extend the filing deadline for candidates threatening the integrity of their ballot preparation process or keep the original deadline and disadvantage or outright bar cash-poor political candidates across the state from qualifying as congressional candidates. Truly, compliance with this injunction “would require heroic efforts by those state and local authorities in the next few weeks—and even heroic efforts likely would not be enough to avoid chaos and confusion.”³²⁵ The fact that one congressional district is retaining its boundaries for a special election for the current Congress does little to remedy many of these concerns. At best, this relieves a single congressional district of a small portion of the burden generated by redistricting.

It should go without saying that the Judge Brown's notion—that this case somehow fits into the narrow exception to Purcell outlined in Justice Kavanaugh's concurrence in Milligan—is absurd.³²⁶ Far from clearcut in favor of the plaintiffs, Judge Brown must strain credulity and distort the record to reach his desired result, as has been highlighted throughout this opinion. As to the feasibility of implementing the injunction without significant cost, confusion, or hardship, this entire section is a testament to how far the plaintiffs are from satisfying that requirement.

Unfairness is the word of the day, and this injunction is laden with unfair consequences. See id. (“Late judicial tinkering with election laws can lead to disruption and to unanticipated and unfair consequences ․”). It is unfair to the congressional candidates (not to mention some candidates for state office) who need to rework their entire campaigns after more than a month of campaigning. It is unfair to the election officials who will be put into an impossible bind. It is unfair to the political parties whose candidates will be chosen through a confused and muddied process as a result of judicial meddling. Most importantly, it is unfair to the Texas voters who are having a map implemented by their duly elected legislature overturned by a self-aggrandizing, results-oriented court.

I dissent.

* * * * *

Beyond the grave error in granting an injunction, Judge Brown adds insult to injury by failing to stay the order for, say, at least 72 hours to give the state a chance to appeal or move for a stay. It is obvious that there will be chaos and political posturing as soon as the injunction is announced. Any observance of judicial restraint would dictate providing an opportunity for provisional adjustments in anticipation of further judicial action. But ideological zeal sometimes overrides common sense.

District courts often stay their orders, either pending a full appeal or for a time certain, to allow for an orderly disposition on further review. A prominent recent example, in an election case, is Nairne v. Landry, 151 F.4th 666 (5th Cir. 2025), in which the district court wisely granted a stay pending appeal of its order enjoining certain elections.

The same should obtain here.

I dissent.

* * * * *

Judge Brown's analysis exposes either a naivete that is unbefitting of the judiciary or a willful blindness unbecoming of the judiciary. Collected below is a non-exhaustive list of misleading, deceptive, or false statements Judge Brown put forward. (The list would be considerably longer but for the press of time; there's no lack of fodder.)

• Judge Brown says “[w]hen the Trump Administration reframed its request as a demand to redistrict on exclusively racial grounds, however, Texas lawmakers immediately jumped on board.”327 Misleading at best.

• Judge Brown says “[b]y all appearances, however, Republican lawmakers didn't have much appetite to redistrict on purely partisan grounds—even at the President's behest.”328 Misleading at best.

• Judge Brown says “[a]nd as far as some influential members of the Legislature were aware, the prospect of redistricting in 2025 was just a rumor”329 Misleading at best.

• Judge Brown says “[w]here the other factors are strong,” the movant need only show ‘some likelihood of success on the merits’ to obtain a preliminary injunction.”330 Misleading at best.

• Judge Brown says “Supreme Court precedent establishes, however, that when: (1) a relevant political actor “purposefully establishe[s] a racial target” that voters of a single race “should make up no less than a majority” of the voting population; and (2) the Legislature “follow[s] those directions to the letter, such that the 50%-plus racial target ha[s] a direct and significant impact on [the districts'] configuration,” a factfinder may permissibly conclude “that race predominated in drawing” those districts.”331 Deeply misleading quote mining at best, intentionally deceptive at worst.

• Judge Brown says “[w]hy not just base the 2025 redistricting exclusively on Rucho? The answer must be that race and Petteway were essential ingredients of the map, without which the 2025 redistricting wouldn't have occurred.”332 False.

• Judge Brown says “[in Cooper v. Harris], the mapmaker had achieved an “on-the-nose attainment of a 50% BVAP” in the challenged district—a feat that, in the district court's view, the mapdrawer would have been unlikely to achieve by blind adherence to partisan data alone. The district court deemed it far more likely that the mapdrawer used a 50% racial target to “deliberately redr[a]w [the challenged district] as a majority-minority district.”333 Deeply misleading quote mining at best, intentionally deceptive at worst.

• Judge Brown says “[e]ven more notably, Dr. Duchin's testimony was effectively unchallenged; no defense expert submitted a report rebutting Dr. Duchin's findings.”334 Misleading.

• Judge Brown says “[i]n any event, if raising the floor to a value closer to 60% would have undermined Dr. Duchin's conclusions, the State Defendants could have introduced expert rebuttal testimony to that effect. Again, though, the State Defendants let Dr. Duchin's testimony go unrebutted”335 False.

• Judge Brown says “[i]n this case, ‘[l]ate judicial tinkering’ with Texas's congressional map is not what could ‘lead to disruption and to unanticipated and unfair consequences for candidates, political parties, and voters.’ ”336 False.

• Judge Brown says “[t]he Court adds that even Ms. Adkins testified that the Texas election officials and systems are more than capable of proceeding with the 2026 congressional election under any map that is the law.”337 False.

* * * * *

This order, replete with legal and factual error, and accompanied by naked procedural abuse, demands reversal.

* * * * *

Darkness descends on the Rule of Law. A bumpy night, indeed.

So SIGNED this 19th day of November 2025.

FOOTNOTES

FOOTNOTE.  

3.   Jackson v. Tarrant County, ––– F.4th ––––, ––––, 2025 WL 3019284, at *7 (5th Cir. Oct. 29, 2025) (citation modified).We adhere to our prior ruling that we must follow published Fifth Circuit opinions as binding precedent, see League of United Latin Am. Citizens v. Abbott, 767 F. Supp. 3d 393, 401 & n.18 (W.D. Tex. 2025), even though any appeals from this order will go directly to the Supreme Court instead of the Fifth Circuit, see 28 U.S.C. § 1253.

4.   See, e.g., League of United Latin Am. Citizens v. Abbott, No. 3:21-cv-00259, 2022 WL 4545757, at *1 n.9 (W.D. Tex. Sept. 28, 2022) [hereinafter Intervenors MTD Op.] (noting that “[c]ourts agree that racial gerrymandering can violate the Fourteenth and Fifteenth Amendments alike”).

5.   E.g., Bethune-Hill v. Va. State Bd. of Elections, 580 U.S. 178, 187, 137 S.Ct. 788, 197 L.Ed.2d 85 (2017) (citation modified).

6.   E.g., Alexander v. S.C. State Conf. of the NAACP, 602 U.S. 1, 7, 144 S.Ct. 1221, 218 L.Ed.2d 512 (2024) (citation modified).

7.   E.g., id. at 38, 144 S.Ct. 1221 (citation modified).

8.   E.g., id. (citation modified).

9.   We have no occasion or need to decide whether intentional vote dilution violates the Fourteenth Amendment to the Constitution, the Fifteenth Amendment, or both. See, e.g., Intervenors MTD Op., 2022 WL 4545757, at *1 n.7 (noting that “[t]he Supreme Court has not yet [answered that question] conclusively”).

10.   See, e.g., Garza v. County of Los Angeles, 918 F.2d 763, 766 (9th Cir. 1990) (“To the extent that a redistricting plan deliberately minimizes minority political power, it may violate both the Voting Rights Act and the Equal Protection Clause of the fourteenth amendment.”).

11.   E.g., Alexander, 602 U.S. at 39, 144 S.Ct. 1221 (citation modified).

12.   See, e.g., League of United Latin Am. Citizens v. Abbott, 601 F. Supp. 3d 147, 162 (W.D. Tex. 2022) [hereinafter 1st Prelim. Inj. Op.] (remarking that racial-gerrymandering and intentional vote-dilution claims “both require discriminatory intent”).

13.   See, e.g., League of United Latin Am. Citizens v. Abbott, 604 F. Supp. 3d 463, 493 (W.D. Tex. 2022) (“Before [the 1982 amendments to the VRA], intent was integral to any Section 2 claim ․ The 1982 amendments removed that requirement, allowing plaintiffs to show a violation by demonstrating discriminatory effect.” (citations omitted)).

14.   See generally Thornburg v. Gingles, 478 U.S. 30, 106 S.Ct. 2752, 92 L.Ed.2d 25 (1986); see also, e.g., Abbott v. Perez, 585 U.S. 579, 614, 138 S.Ct. 2305, 201 L.Ed.2d 714 (2018) (“To make out a § 2 ‘effects’ claim, a plaintiff must establish the three so-called ‘Gingles factors.’ ”).The plaintiff must also “show, under the totality of the circumstances, that the political process is not equally open to minority voters.” Allen v. Milligan, 599 U.S. 1, 18, 143 S.Ct. 1487, 216 L.Ed.2d 60 (2023) (citation modified). That additional requirement isn't pertinent here.

15.   See, e.g., Allen v. Milligan, 599 U.S. 1, 18, 143 S.Ct. 1487, 216 L.Ed.2d 60 (2023).

16.   See Campos v. City of Baytown, 840 F.2d 1240, 1244 (5th Cir. 1988) (“There is nothing in the law that prevents the plaintiffs from identifying the protected aggrieved minority to include both Blacks and Hispanics.”), overruled by Petteway v. Galveston County, 111 F.4th 596 (5th Cir. 2024) (en banc).

17.   See, e.g., Bartlett v. Strickland, 556 U.S. 1, 13, 129 S.Ct. 1231, 173 L.Ed.2d 173 (2009) (plurality opinion) (defining a “coalition district” as one “in which two minority groups form a coalition to elect the candidate of the coalition's choice”).

18.   See Petteway, 111 F.4th at 599 (“We OVERRULE Campos ․”).

19.   Id. at 603.

20.   See, e.g., id. at 610 (“When, as here, a minority group cannot constitute a majority in a single-member district without combining with members of another minority group, Section 2 does not provide protection.”).

21.   See, e.g., id. at 599 (“The issue in this en banc case is whether Section 2 of the Voting Rights Act authorizes coalitions of racial and language minorities to claim vote dilution in legislative redistricting.”); id. at 601 (“The primary issue here concerns the first Gingles precondition ․”).

22.   See id. at 600 (“Following a ten-day bench trial, the district court found that the enacted plan violated Section 2 ․ The district court declined to reach the intentional discrimination and racial gerrymandering claims brought by the Petteway Plaintiffs and NAACP Plaintiffs because the relief they requested with respect to those claims was no broader than the relief they were entitled to under Section 2.”).See also, e.g., id. at 599 n.1 (“[T]he issue of intentional discrimination was not part of the district court's Section 2 ruling. The court withheld ruling on that constitutional issue, which we remand for further consideration.”).

23.   See id. at 614.

24.   See generally id. at 599–614; see also infra Section II.D.

25.   See 588 U.S. 684, 718, 139 S.Ct. 2484, 204 L.Ed.2d 931 (2019) (“[P]artisan gerrymandering claims present political questions beyond the reach of the federal courts.”).

26.   See id. at 719–20, 139 S.Ct. 2484 (noting that “numerous other States” have “restrict[ed] partisan considerations in districting through legislation,” and that several States “have outright prohibited partisan favoritism in redistricting”); see also id. at 720–21, 139 S.Ct. 2484 (remarking that the U.S. Congress could theoretically pass legislation to restrict partisan gerrymandering).

27.   See, e.g., Alexander, 602 U.S. at 6, 144 S.Ct. 1221 (“[A]s far as the Federal Constitution is concerned, a legislature may pursue partisan ends when it engages in redistricting.”).

28.   Jackson, ––– F.4th at ––––, 2025 WL 3019284, at *6.

29.   See infra Section III.B.2.

30.   See, e.g., 1st Prelim. Inj. Op., 601 F. Supp. 3d at 155–56.See also, e.g., Jackson, ––– F.4th at ––––, 2025 WL 3019284, at *2 (“To comply with the federal ‘one person, one vote’ principle ․ states and their political subdivisions must generally redistrict upon release of the decennial census to account for any changes or shifts in population.” (citation modified)).

31.   See Brooks Prelim. Inj. Ex. 258, ECF No. 1326-5, at 1.Here and below, the numbers don't add up to 100% because the Court has omitted the percentages of voters belonging to racial groups that are not numerous in Texas, such as Native Hawaiians and American Indians. See, e.g., id. (noting that the 2021 version of CD 9 was 0.2% American Indian by CVAP). The Court of course does not imply any disrespect for those voters by doing so.Additionally, all CVAP figures in this opinion are subject to a margin of error. See, e.g., id.

32.   See id.

33.   See id.

34.   See id.

35.   See Petteway, 111 F.4th 596 (decided August 1, 2024); see also supra Section II.A.3.

36.   See Prelim. Inj. Hr'g Tr. Day 1 (Morning), ECF No. 1414, at 18.See also, e.g., id. at 17 (“[The 2021 Map] was drawn race blind. Any work we did on it was race blind.”); id. at 19 (“Based on [the Supreme Court's] warning against race-based redistricting, I drafted all the proposed maps totally blind to race.”).

37.   See id. at 17 (“[T]he maps were drawn blind to race. So adjustments were made for population. Sometimes for partisan shading and so forth. But those were the priorities that we used.”); id. (“All the race neutral objectives were used ․ in drawing the maps ․”).Mr. Adam Kincaid—who was the outside mapmaker who drew all of the 2021 Map except for the four districts highlighted above (CDs 9, 18, 29, and 33), see, e.g., Prelim. Inj. Hr'g Tr. Day 6 (Afternoon), ECF No. 1342, at 58–59—likewise testified that he didn't look at racial data when drawing the map. See, e.g., Prelim. Inj. Hr'g Tr. Day 7 (Morning), ECF No. 1420, at 19 (“I didn't look at the minority numbers in 2021 ․”); Prelim. Inj. Hr'g Tr. Day 6 (Afternoon), ECF No. 1342, at 87 (“[T]he entire 2021 map was drawn race-blind as far as I drew it.”).The four districts that Mr. Kincaid didn't draw resulted from amendments in the Texas House after the Senate passed Senator Huffman's bill. See, e.g., Prelim. Inj. Hr'g Tr. Day 6 (Afternoon), ECF No. 1342, at 58–59. Here too, the preliminary-injunction record contains no evidence that the Legislature made any of those changes to comply with Campos. The record instead suggests that the Legislature passed those amendments to eliminate incumbent pairing, respect communities of interest, and preserve economic engines within the districts. See, e.g., Prelim. Inj. Hr'g Tr. Day 2 (Morning), ECF No. 1415, at 79–86, 139; see also Chen v. City of Houston, 206 F.3d 502, 515 (5th Cir. 2000) (noting that “concern about communities of interest is a valid traditional districting tool that may serve to deflect an inference that race predominated in districting”).

38.   See, e.g., Prelim. Inj. Hr'g Tr. Day 1 (Morning), ECF No. 1414, at 19 (“Once I drafted the maps, I ensured that they underwent a legal compliance check to ensure that there were no inadvertent violations of the law, including the Voting Rights Act.”); id. at 17 (“All the race neutral objectives were used ․ in drawing the maps that were drawn blind to race and then submitted [to outside attorneys for a legal compliance check]. And then our attorneys gave us—we were advised that [the maps] did not violate the Voting Rights Act. They were legally compliant.”).

39.   Given the current procedural posture, we have no occasion to make binding, definitive findings about the 2021 Legislature's intent when devising and enacting the 2021 congressional map—or, for that matter, the Texas House and Senate maps that the Legislature also enacted in 2021. The latter were the subject of a bench trial we held several months ago, and the Court has yet to rule on them.See also, e.g., Univ. of Tex. v. Camenisch, 451 U.S. 390, 395, 101 S.Ct. 1830, 68 L.Ed.2d 175 (1981) (“[F]indings of fact and conclusions of law made by a court granting a preliminary injunction are not binding at trial on the merits.”).

40.   See, e.g., Prelim. Inj. Hr'g Tr. Day 9 (Afternoon), ECF No. 1345, at 123 (the State Defendants’ closing argument at the preliminary-injunction hearing, agreeing that none “of the districts in the 2021 map were drawn based on race”); Defs.’ Post-Hr'g Br., ECF No. 1284, at 30 (insisting that “districts in 2021 ․ were drawn race-blind”).

41.   See, e.g., Prelim. Inj. Hr'g Tr. Day 6 (Afternoon), ECF No. 1342, at 7–9, 17.

42.   See Defs.’ Prelim. Inj. Ex. 1415, ECF No. 1364-5, at 2.

43.   See id.

44.   See id.

45.   See id.

46.   See id.

47.   See Gonzales Prelim. Inj. Ex. 35, ECF No. 1388-19, at 1–2; see also Prelim. Inj. Hr'g Tr. Day 7 (Morning), ECF No. 1420, at 119–20; Prelim. Inj. Hr'g Tr. Day 8 (Afternoon), ECF No. 1344, at 19.

48.   See, e.g., Prelim. Inj. Hr'g Tr. Day 2 (Morning), ECF No. 1415, at 90–91 (“Q. “Now, it's been stated by others that redistricting was in the conversation prior to [the DOJ Letter discussed later in this opinion] ․ What do you say to that? | [REPRESENTATIVE THOMPSON:] I heard it all during the session, and I made inquiries about it. And I asked [Chairman Hunter] ․ if they were going to be redistricting․ [H]e said he didn't know. You know, I think he told me he was unaware of any redistricting. And he kind of brushed it off as though it just might have been just a rumor or something, you know.”).

49.   The Legislature amended the State's congressional map before our panel was able to rule on the 2021 Map's legality.

50.   Trial Tr. (June 10, 2025), ECF No. 1413, at 54.

51.   See generally Brooks Prelim. Inj. Ex. 253, ECF No. 1326.

52.   Abrupt line break in original. See id. at 2.

53.   Id. at 1–2.

54.   See Defs.’ Prelim. Inj. Ex. 1466, ECF No. 1380-25, at 4 (“My office stands ready to support President Trump, Governor Abbott, and the Texas Legislature in their redistricting goals and will defend any new maps passed from challenges by the radical Left.”).

55.   See Defs.’ Resp. Gonzales Pls.’ Prelim. Inj. Mot., ECF No. 1199, at 12.

56.   See id. at 20.

57.   See Defs.’ Resp. J. Prelim. Inj. Mot., ECF No. 1200, at 13, 30.

58.   See id. at 13.

59.   See Prelim. Inj. Hr'g Tr. Day 9 (Afternoon), ECF No. 1345, at 123.

60.   See supra Section II.A.3.

61.   See Brooks Prelim. Inj. Ex. 253, ECF No. 1326, at 1 (describing “coalition districts” as “unconstitutional”); id. at 2 (claiming that “ ‘coalition districts’ run afoul the [sic] Voting Rights Act and the Fourteenth Amendment”); id. (“It is the position of this Department that [CDs 9, 18, 29, and 33] constitute unconstitutional racial gerrymanders, under the logic and reasoning of Petteway.”).

62.   See id. at 1 (“Congressional Districts TX-09, TX-18, TX-29 and TX-33 currently constitute unconstitutional ‘coalition districts’ and we urge the State of Texas to rectify these race-based considerations from these specific districts.”); id. at 2 (“If the State of Texas fails to rectify the racial gerrymandering of TX-09, TX-18, TX-29 and TX 33, the Attorney General reserves the right to seek legal action against the State ․”).

63.   See generally Petteway, 111 F.4th at 599–614.

64.   See supra note 22.

65.   Easley v. Cromartie, 532 U.S. 234, 249, 121 S.Ct. 1452, 149 L.Ed.2d 430 (2001) [hereinafter Cromartie II] (emphasis omitted).

66.   Id.

67.   Cf. Voinovich v. Quilter, 507 U.S. 146, 156, 113 S.Ct. 1149, 122 L.Ed.2d 500 (1993) (“[T]he federal courts may not order the creation of majority-minority districts unless necessary to remedy a violation of federal law. But that does not mean that the State's powers are similarly limited. Quite the opposite is true ․” (citation omitted)); id. at 155, 113 S.Ct. 1149 (“Section 2 contains no per se prohibitions against particular types of districts ․ Instead, § 2 focuses exclusively on the consequences of apportionment. Only if the apportionment scheme has the effect of denying a protected class the equal opportunity to elect its candidate of choice does it violate § 2; where such an effect has not been demonstrated, § 2 simply does not speak to the matter.”).

68.   556 U.S. 1, 129 S.Ct. 1231.Under the “Marks rule,” “[w]hen a fragmented Court decides a case and no single rationale explaining the result enjoys the assent of five Justices, the holding of the Court may be viewed as that position taken by those Members who concurred in the judgments on the narrowest grounds.” Marks v. United States, 430 U.S. 188, 193, 97 S.Ct. 990, 51 L.Ed.2d 260 (1977). The plurality opinion in Bartlett decides the case on much narrower grounds than the concurrence. Contrast 556 U.S. at 6–26, 129 S.Ct. 1231 (plurality opinion) (concluding that a VRA § 2 plaintiff cannot satisfy the Gingles factors by proposing a crossover district), with id. at 26, 129 S.Ct. 1231 (Thomas, J., concurring) (concluding that VRA § 2 “does not authorize any vote dilution claim, regardless of the size of the minority population in a given district”). The plurality opinion is therefore the precedential one under Marks.

69.   See, e.g., 1st Prelim. Inj. Op., 601 F. Supp. 3d at 163 (our prior opinion interpreting Bartlett to mean that “it must be possible for a state to violate the Constitution by dismantling a district that does not meet all three Gingles requirements”).Given that Bartlett undermines DOJ's argument, it's puzzling that DOJ cited Bartlett in its letter. See Brooks Prelim. Inj. Ex. 253, ECF No. 1326, at 2.

70.   See 556 U.S. at 13–14, 129 S.Ct. 1231 (noting that Bartlett did “not address th[e] type of coalition district” that is at issue here).

71.   See id. at 13, 129 S.Ct. 1231.

72.   See id. at 23, 129 S.Ct. 1231 (“§ 2 does not require crossover districts ․”).

73.   See id.

74.   See id.

75.   Id. at 23–24, 129 S.Ct. 1231.

76.   Id. at 24, 129 S.Ct. 1231.

77.   See id. at 24, 129 S.Ct. 1231.Although the State Defendants dismiss this language as mere dicta, see Defs.’ Post-Hr'g Br., ECF No. 1284, at 22–23, Fifth Circuit precedent requires us to “take [dicta] from the Supreme Court seriously.” See, e.g., Croft v. Perry, 624 F.3d 157, 164 (5th Cir. 2010).

78.   See Petteway, 111 F.4th at 610 (“Each of the[ ] reasons articulated in Bartlett for rejecting crossover claims applies with equal force to coalition claims.”).

79.   See id. (“One need only transpose Bartlett’s language to indicate the problems [with coalition districts]: ‘What percentage of [black] voters supported [Hispanic]-preferred candidates in the past? How reliable would the [coalition] votes be in future elections? What types of candidates have [black] and [Hispanic] voters supported together in the past and will those trends continue? Were past [coalition] votes based on incumbency and did that depend on race? What are the historical turnout rates among [black] and [Hispanic] minority voters and will they stay the same?’ ” (quoting Bartlett, 556 U.S. at 17, 129 S.Ct. 1231)).

80.   Cf. Bartlett, 556 U.S. at 23, 129 S.Ct. 1231.

81.   Cf. id. at 24, 129 S.Ct. 1231.

82.   Cf. id.

83.   Contra Brooks Prelim. Inj. Ex. 253, ECF No. 1326, at 1–2.

84.   See Brooks Prelim. Inj. Ex. 253, ECF No. 1326, at 1 (“Congressional Districts TX-09, TX-18, TX-29 and TX-33 currently constitute unconstitutional ‘coalition districts’ and we urge the State of Texas to rectify these race-based considerations from these specific districts.”).

85.   See id. at 2 (describing CD 29 as “a majority Hispanic district” on the very next page).

86.   See Brooks Prelim. Inj. Ex. 258, ECF No. 1326-5, at 1 (indicating that CD 29's Hispanic CVAP was 63.5% under the 2021 Map); see also supra Section II.B.

87.   See generally Petteway, 111 F.4th at 599–614.

88.   See supra Section II.B.

89.   Contra Brooks Prelim. Inj. Ex. 253, ECF No. 1326, at 2.

90.   See id.

91.   See, e.g., Perry v. Perez, 565 U.S. 388, 391–92, 132 S.Ct. 934, 181 L.Ed.2d 900 (2012) (“As Texas’ 2012 primaries approached, it became increasingly likely that the State's newly enacted plans would not receive preclearance in time for the 2012 elections. And the State's old district lines could not be used, because population growth had rendered them inconsistent with the Constitution's one-person, one-vote requirement. It thus fell to the District Court in Texas to devise interim plans for the State's 2012 primaries and elections.”).See also, e.g., id. at 390–91, 132 S.Ct. 934 (explaining the VRA § 5 preclearance process).But see Perez v. Texas, 970 F. Supp. 2d 593, 598 (W.D. Tex. 2013) (explaining that, in Shelby County v. Holder, 570 U.S. 529, 133 S.Ct. 2612, 186 L.Ed.2d 651 (2013), the Supreme Court “str[uck] down the coverage formula in § 4(b) of the Voting Rights Act which, in turn, means that Texas is no longer automatically subject to § 5 preclearance requirements”).Texas legislatively adopted the court-drawn map as its own in 2013. E.g., Abbott v. Perez, 585 U.S. at 590, 138 S.Ct. 2305 (“The 2013 Legislature ․ enacted the Texas court's interim plans ․ The federal congressional plan was not altered at all ․”).

92.   See Perez v. Texas, 891 F. Supp. 2d 808, 830 (W.D. Tex. 2012) [hereinafter Perez v. Texas 2012] (acknowledging that “race was necessarily considered in drawing CD 33 to some degree”).

93.   See id. (“The contours of CD 33 are a result of addressing the ‘not insubstantial’ § 5 claims of cracking and packing and the application of neutral redistricting criteria․ [T]he use of race was appropriate to remedy the alleged race-based discrimination that occurred ․ The Court finds that [the court-drawn map] adequately resolves the ‘not insubstantial’ § 5 claims ․”).See also Perez v. Abbott, 274 F. Supp. 3d 624, 652 (W.D. Tex. 2017) (“To address the § 5 discrimination claims, [the court-drawn map] included new CD 33, spanning Dallas and Tarrant Counties. [The court-drawn map] withdrew many of the encroachments into minority communities from the Anglo districts surrounding DFW, and the population left behind in DFW from the removed encroachments was placed in new CD 33, while accommodating congressional incumbents and taking into account population growth.”), rev'd and remanded, Abbott v. Perez, 585 U.S. 579, 138 S.Ct. 2305, 201 L.Ed.2d 714 (2018).

94.   Contra Brooks Prelim. Inj. Ex. 253, ECF No. 1326, at 2.

95.   See Perez v. Texas 2012, 891 F. Supp. 2d at 830.

96.   See id.

97.   See Perez v. Abbott, 274 F. Supp. 3d at 653.

98.   See id. (“[CD 33] is majority-minority CVAP when Black and Hispanic CVAP are combined, and it has elected an African-American, Mark Veasey. It has thus performed as a minority coalition district under most [p]laintiffs’ view that such districts require minority cohesion only in the general elections.”).

99.   Contra Brooks Prelim. Inj. Ex. 253, ECF No. 1326, at 2.

100.   See id.

101.   See id. at 1–2.

102.   See id.

103.   See id.

104.   See supra Section II.C.

105.   See Brooks Prelim. Inj. Ex. 322-T, ECF No. 1327-22, at 3 (Harmeet Dhillon's statement that the DOJ Letter “is what triggered the Texas legislature and the Texas governor to call the legislature into session to put new maps together”).

106.   See Brooks Prelim. Inj. Ex. 253, ECF No. 1326, at 1.

107.   Brooks Prelim. Inj. Ex. 254, ECF No. 1326-1, at 3 (emphasis added).

108.   See Brooks Prelim. Inj. Ex. 253, ECF No. 1326, at 1–2.

109.   See Brooks Prelim. Inj. Ex. 254, ECF No. 1326-1, at 3.

110.   See supra Section II.A.4 (discussing Rucho).

111.   Prelim. Inj. Hr'g Tr. Day 1 (Morning), ECF No. 1414, at 12–14; see also Brooks Prelim. Inj. Ex. 335-T, ECF No. 1328-1, at 4–5.

112.   Compare Prelim. Inj. Hr'g Tr. Day 1 (Morning), ECF No. 1414, at 13 (“[Y]ou are doing this to give Trump and Republicans in the House of Representatives five additional seats, right?”), with id. at 14 (“[T]he reason why we are doing this is because of that court decision.”).

113.   See id.

114.   See id. (“It just coincides it's going to be Hispanic Republicans elected to those seats.” (emphases added)).

115.   See Brooks Prelim. Inj. Ex. 325-T, ECF No. 1327-25, at 3–4 (July 22, 2025, interview in which the Governor stated that “we want to make sure that we have maps that don't impose coalition districts”); see also Prelim. Inj. Hr'g Tr. Day 1 (Morning), ECF No. 1414, at 32.See also, e.g., Prelim. Inj. Hr'g Tr. Day 1 (Morning), ECF No. 1414, at 84 (“[The Fifth Circuit] decided that Texas is no longer required to have what are called coalition districts and, as a result, we[’]re able to take the people who were in those coalition districts and make sure they are going to be in districts that really represent the voting preference of those people who live here in Texas.”); see also Brooks Prelim. Inj. Ex. 332-T, ECF No. 1411-3, at 2.

116.   See, e.g., Brooks Prelim. Inj. Ex. 325-T, ECF No. 1327-25, at 4–5 (“STEVEN DIAL: ․ There's been criticism of you saying you're letting President Trump call the shots. | GOV. GREG ABBOTT: Listen, people are always going to lodge criticisms. I'm not worried about stuff like that. What I'm worried about is making sure that we are going to have congressional districts ․ that fit the structure of [Petteway] ․”).

117.   See Prelim. Inj. Hr'g Tr. Day 1 (Morning), ECF No. 1414, at 84 (“Four of the five districts that we are going to create are predominantly Hispanic districts that happen to be voting for Republicans as opposed to Democrats.” (emphasis added)); id. at 77 (“Four of the five districts we are drawing, they would be Hispanic districts. They happen to be Hispanic Republican districts.” (emphases added)).

118.   See supra Section II.E.

119.   See Defs.’ Prelim. Inj. Ex. 1466, ECF No. 1380-25, at 2.

120.   See supra Section II.D.

121.   See Defs.’ Prelim. Inj. Ex. 1466, ECF No. 1380-25, at 2–3 (“I am ․ keenly aware of the Fifth Circuit's decision in Petteway ․ We ․ agree that, had the Texas legislature felt compelled under pre-Petteway strictures to create coalition districts, the basis for such decisions—as you say—‘no longer exists.’ However, my office has just completed a four-week trial against various plaintiff groups concerning the constitutionality of Texas's congressional districts ․ The evidence at that trial was clear and unequivocal: the Texas legislature did not pass race-based electoral districts ․ Texas State Senator Joan Huffman, who chaired the Senate Redistricting Committee, testified under oath that she drew Texas districts blind to race, and sought to maximize Republican political advantage balanced against traditional redistricting criteria․ The Texas Legislature ․ has drawn its current maps in conformance with traditional, non-racial criteria to ensure Texas continues to adopt policies that will truly Make America Great Again. As permitted by federal law, the congressional maps in 2021 were drawn on a partisan basis.” (citations omitted)).

122.   See id. at 3–4 (“The Texas Legislature has led the Nation in rejecting race-based decision-making in its redistricting process—it has drawn its current maps in conformance with traditional, non-racial redistricting criteria to ensure Texas continues to adopt policies that will truly Make America Great Again․ For these reasons, I welcome continued dialogue about how Texas's electoral districts can best serve Texas voters without regard to outdated and unconstitutional racial considerations. My office stands ready to support President Trump, Governor Abbott, and the Texas Legislature in their redistricting goals ․”).

123.   See supra Section II.E.

124.   See Brooks Prelim. Inj. Ex. 265, ECF No. 1326-12, at 1.

125.   See Brooks Prelim. Inj. Ex. 267, ECF No. 1326-14, at 2 (indicating that 12.6% of new CD 9 consists of voters from old CD 2, 2.9% consists of voters from old CD 9, 43.7% consists of voters from old CD 29, and 40.7% consists of voters from old CD 36).

126.   See Brooks Prelim. Inj. Ex. 253, ECF No. 1326, at 1–2.

127.   See Brooks Prelim. Inj. Ex. 265, ECF No. 1326-12, at 1.

128.   See Brooks Prelim. Inj. Ex. 267, ECF No. 1326-14, at 3 (indicating that 64.5% of new CD 18's population came from old CD 9, and that a plurality of the population that the Legislature moved from old CD 9 (46.1%) was Black).

129.   See Brooks Prelim. Inj. Ex. 253, ECF No. 1326, at 1–2.

130.   Compare Brooks Prelim. Inj. Ex. 258, ECF No. 1326-5, at 1 (CD 29's CVAP statistics under the 2021 Map), with Brooks Prelim. Inj. Ex. 265, ECF No. 1326-12, at 1 (CD 29's CVAP statistics under the 2025 Map).See also Prelim. Inj. Hr'g Tr. Day 2 (Afternoon), ECF No. 1338, at 36–37.

131.   See Brooks Prelim. Inj. Ex. 267, ECF No. 1326-14, at 5 (indicating that only 37.2% of the voters who were in CD 29 under the 2021 Map remain in CD 29 under the 2025 Map).

132.   See Prelim. Inj. Hr'g Tr. Day 2 (Afternoon), ECF No. 1338, at 44–45 (stating that “Latino neighborhoods like Denver Harbor, Magnolia Park, Second Ward, Manchester, and Northside”—“historic centers of Latino political strength”—were “carved out” of CD 29).

133.   See Brooks Prelim. Inj. Ex. 253, ECF No. 1326, at 1–2; see also supra Section II.B.

134.   Compare Brooks Prelim. Inj. Ex. 258, ECF No. 1326-5, at 1 (indicating that, under the 2021 Map, CD 33 was 43.6% Hispanic, 25.2% Black, 23.4% White, and 5.7% Asian), with Brooks Prelim. Inj. Ex. 265, ECF No. 1326-12, at 1 (indicating that, under 2025 Map, CD 33 is 38.2% Hispanic, 19.6% Black, 35.5% White, and 4.4% Asian).

135.   See Brooks Prelim. Inj. Ex. 253, ECF No. 1326, at 1–2.

136.   See Brooks Prelim. Inj. Ex. 258, ECF No. 1326-5, at 1.See also, e.g., Prelim. Inj. Hr'g Tr. Day 3 (Morning), ECF No. 1416, at 40 (“Under [the 2021 Map], CD 22 was a plurality White district. That is, the majority of the population were [sic] of no particular racial group; but the largest group were [sic] White.”).

137.   See Brooks Prelim. Inj. Ex. 258, ECF No. 1326-5, at 1 (indicating that the 2021 version of CD 22 was 24.6% Hispanic, 12.7% Black, and 11.3% Asian).See also, e.g., Prelim. Inj. Hr'g Tr. Day 3 (Morning), ECF No. 1416, at 40 (“[T]he remainder would be non-Whites. So it was a majority non-White district.”).

138.   Coalition districts are also defined by whether the two aggregated minority groups can successfully “elect the candidate of the coalition's choice.” See, e.g., Bartlett, 556 U.S. at 13, 129 S.Ct. 1231. The preliminary-injunction record indicates that the 2021 version of CD 22 did not elect minorities’ candidate of choice. See Tex. NAACP Prelim. Inj. Ex. 208, ECF No. 1384-8, at 9.

139.   See Brooks Prelim. Inj. Ex. 265, ECF No. 1326-12, at 1; see also, e.g., Prelim. Inj. Hr'g Tr. Day 3 (Morning), ECF No. 1416, at 41 (“New CD 22 is majority White.”).

140.   See Brooks Prelim. Inj. Ex. 258, ECF No. 1326-5, at 1.See also, e.g., Prelim. Inj. Hr'g Tr. Day 3 (Morning), ECF No. 1416, at 41 (“[The 2021 version of] CD 27 was a Hispanic plurality district. 48.8 percent of the CVAP were [sic] Hispanic.”).

141.   The preliminary-injunction record indicates that the 2021 version of CD 27 did not elect a minority coalition's candidate of choice. See Tex. NAACP Prelim. Inj. Ex. 208, ECF No. 1384-8, at 9.

142.   See Brooks Prelim. Inj. Ex. 265, ECF No. 1326-12, at 1; Prelim. Inj. Hr'g Tr. Day 3 (Morning), ECF No. 1416, at 41 (“[The 2025 version of] CD 27 is ․ majority White.”).

143.   See Brooks Prelim. Inj. Ex. 258, ECF No. 1326-5, at 1 (indicating that, under the 2021 Map, CD 30 was 46.0% Black, 24.5% Hispanic, 24.0% White, and 3.2% Asian).

144.   See Brooks Prelim. Inj. Ex. 265, ECF No. 1326-12, at 1.

145.   See Brooks Prelim. Inj. Ex. 258, ECF No. 1326-5, at 1 (indicating that, under the 2021 Map, CD 32 was 43.9% White, 23.4% Black, 22.9% Hispanic, and 6.9% Asian).

146.   See Brooks Prelim. Inj. Ex. 265, ECF No. 1326-12, at 1.

147.   See Brooks Prelim. Inj. Ex. 258, ECF No. 1326-5, at 2 (indicating that, under the 2021 Map, CD 35 was 46.0% Hispanic, 35.7% White, 13.0% Black, and 2.7% Asian).

148.   See Brooks Prelim. Inj. Ex. 265, ECF No. 1326-12, at 2.

149.   See H.B. 4, 89th Leg., 2d Spec. Sess. (Tex. 2025) (signed on August 29, 2025).

150.   See generally Tex. NAACP Prelim. Inj. Mot., ECF No. 1142; Intervenors’ Prelim. Inj. Mot., ECF No. 1143; Gonzales Pls.’ Prelim. Inj. Mot., ECF No. 1149; Brooks, LULAC, & MALC Pls.’ Joint Prelim. Inj. Mot., ECF No. 1150.

151.   E.g., Brooks, LULAC, & MALC Pls.’ Post-Hr'g Br., ECF No. 1281, at 2 (emphasis omitted).

152.   See, e.g., id. at 4–38.The Plaintiff Groups also raise intentional vote-dilution challenges that we need not address in this opinion. See Chart of Claims, ECF No. 1208-1, at 2–4; see also infra text accompanying note 163.

153.   Each Plaintiff Group challenges a slightly different set of districts. See Chart of Claims, ECF No. 1208-1, at 2–4.

154.   See id.No Plaintiff Group challenges CD 29 under a racial-gerrymandering theory, as opposed to an intentional vote-dilution theory. See id. at 2. Although we discuss CD 29 at various points in this opinion to illuminate the Legislature's intent in drawing the map more broadly, we do not base our ruling on the State's alleged gerrymandering of CD 29. See Bethune-Hill, 580 U.S. at 191–92, 137 S.Ct. 788 (explaining that although plaintiffs “can present statewide evidence in order to prove racial gerrymandering in a particular district,” “[r]acial gerrymandering claims” must ultimately “proceed district-by-district” (citation modified)).

155.   See, e.g., Defs.’ Resp. Intervenors’ & Tex. NAACP's Prelim. Inj. Mot., ECF No. 1195, at 6 (“The Texas Legislature passed [the] 2025 congressional map on precisely partisan lines.”); Defs.’ Post-Hr'g Br., ECF No. 1284, at 11 (“Texas's 2025 map is, and always has been, about partisanship.”).

156.   See, e.g., Defs.’ Post-Hr'g Br., ECF No. 1284, at 23 (“Race was not used here.”).

157.   See, e.g., id. at 21 (“[T]he redistricting occurred because President Trump wanted a chance for Texas to elect up to five more Republicans to Congress in 2026.” (citation modified)); Defs.’ Resp. Gonzales Pls.’ Prelim. Inj. Mot., ECF No. 1199, at 10–11 (“Mindful of history showing that a president's political party tends to lose House seats in mid-term election years and concerned that a Democrat majority would disrupt his national agenda, President Trump ․ called on Texas lawmakers to find five additional congressional seats ․ It is this political arms-race that motivated Texas legislators to redistrict mid-decade, not race.”).

158.   See, e.g., Defs.’ Resp. Intervenors’ & Tex. NAACP's Prelim. Inj. Mot., ECF No. 1195, at 23–24 (“Given the danger to President Trump's legislative agenda posed by [the] 2026 elections and the historical trend of the presidential party doing poorly in non-presidential election years, there was a great deal of political pressure placed on the State of Texas to match the political gerrymandering of Democrat states. This pressure only intensified when other states, especially California, pledged to perform mid-decade redistricting to make their already one-sided congressional maps even more favorable to Democrats․ None of those factors indicate race was involved ․”).After we held the preliminary-injunction hearing in this case, California passed Proposition 50, which increases the number of Democrat-leaning congressional districts in California to counterbalance the 2025 Map's creation of additional Republican-leaning congressional seats in Texas.

159.   Some Fifth Circuit opinions state that a plaintiff seeking a preliminary injunction must show “a substantial likelihood that he will prevail on the merits,” see, e.g., TitleMax of Tex., Inc. v. City of Dallas, 142 F.4th 322, 328 (5th Cir. 2025) (emphasis added), whereas others state that the plaintiff need only show “a likelihood of success on the merits,” see, e.g., Jackson, ––– F.4th at ––––, 2025 WL 3019284, at *3 (emphasis added).We will go with the language in the Fifth Circuit's most recent redistricting opinion, since it's the preliminary-injunction opinion that's most factually and procedurally analogous to the instant case. See Jackson, ––– F.4th at ––––, 2025 WL 3019284, at *3.Either way, given the Fifth Circuit's sliding-scale approach to the likelihood-of-success inquiry, see infra note 167 and accompanying text, we perceive no substantive difference between the two formulations of the standard.

160.   Jackson, ––– F.4th at ––––, 2025 WL 3019284, at *3 (citing Winter v. Nat. Res. Def. Council, Inc., 555 U.S. 7, 20, 129 S.Ct. 365, 172 L.Ed.2d 249 (2008)).

161.   TitleMax, 142 F.4th at 328 (quoting Canal Auth. of Fla. v. Callaway, 489 F.2d 567, 573 (5th Cir. 1974)).See also, e.g., Mazurek v. Armstrong, 520 U.S. 968, 972, 117 S.Ct. 1865, 138 L.Ed.2d 162 (1997) (case cited by the State Defendants for a similar proposition); Barber v. Bryant, 860 F.3d 345, 352 (5th Cir. 2017) (same).

162.   The Plaintiff Groups have not shown that they're likely to succeed on their racial-gerrymandering challenge to CD 33. See infra Section III.B.6.a. Nor have the Plaintiff Groups shown that they're likely to succeed on their racial-gerrymandering challenge to CD 22. See infra note 358. Thus, we do not base our grant of a preliminary injunction on those claims.

163.   Nor do we base our preliminary injunction ruling in any way on the Gonzales Plaintiffs’ malapportionment claim. See Gonzales Pls.’ Post-Hr'g Br., ECF No. 1278, at 3 n.2 (stating that “[t]he Gonzales Plaintiffs continue to seek preliminary relief as to this claim”). We dismissed the count on which that claim was based on September 30, 2025. See generally Mem. Op. & Order, ECF No. 1226.The Court's ruling on the Gonzales Plaintiffs’ motion to enter an appealable partial final judgment on that claim is forthcoming. See generally Gonzales Pls.’ Mot. Rule 54(b) Entry of Final J., ECF No. 1265.

164.   E.g., Jackson, ––– F.4th at ––––, 2025 WL 3019284, at *3.

165.   See, e.g., Janvey v. Alguire, 647 F.3d 585, 596 (5th Cir. 2011) (noting that a plaintiff “need not show that he is certain to win” to obtain a preliminary injunction (quoting Charles Alan Wright, Arthur R. Miller & Mary Kay Kane, 11A Federal Practice & Procedure § 2948.3 (2d ed. 1995))).

166.   See, e.g., Jefferson Cmty. Health Care Ctrs., Inc. v. Jefferson Par. Gov't, 849 F.3d 615, 626 (5th Cir. 2017) (“[T]here is no particular degree of likelihood of success that is required in every case ․”); TitleMax, 142 F.4th at 328 (“The importance and nature of the likely success on the merits requirement can vary significantly ․” (citation modified)); Fla. Med. Ass'n v. U.S. Dep't of Health, Educ. & Welfare, 601 F.2d 199, 203 n.2 (5th Cir. 1979) (“[F]inding a substantial likelihood that [the] movant will ultimately prevail on the merits does not contemplate a finding of fixed quantitative value.” (citation modified)).

167.   See, e.g., TitleMax, 142 F.4th at 328 (“This court has applied a sliding-scale analysis to the four preliminary injunction requirements. The importance and nature of the likely success on the merits requirement can vary significantly, depending upon the magnitude of the injury which would be suffered by the movant in the absence of interlocutory relief and the relative balance of the threatened hardship faced by each of the parties.” (citation modified)).

168.   E.g., id.

169.   See infra Section III.C.

170.   See infra Section III.D.

171.   Cf., e.g., TitleMax, 142 F.4th at 328.

172.   See id. at 329 (citation modified).

173.   E.g., Cooper v. Harris, 581 U.S. 285, 291, 137 S.Ct. 1455, 197 L.Ed.2d 837 (2017) (citation modified).

174.   E.g., Alexander, 602 U.S. at 8, 144 S.Ct. 1221.

175.   E.g., Cooper, 581 U.S. at 291, 137 S.Ct. 1455 (citation modified).

176.   E.g., id. at 308, 137 S.Ct. 1455 (citation modified).

177.   See, e.g., Jackson, ––– F.4th at ––––, 2025 WL 3019284, at *7 (“[B]ecause the gravamen of a [racial-gerrymandering] claim is the sorting of persons with an intent to divide by reason of race, and this holds true regardless of the motivations of those doing the sorting, plaintiffs raising such a claim need not show that the legislature either intended or succeeded in diluting any particular racial group's voting strength. Rather, the racial classification itself is the relevant harm in that context.” (citation modified)).

178.   See, e.g., Perez v. Abbott, 253 F. Supp. 3d at 945.

179.   E.g., Alexander, 602 U.S. at 9, 144 S.Ct. 1221.

180.   E.g., id. at 6, 144 S.Ct. 1221; see also supra Section II.A.4.

181.   E.g., Alexander, 602 U.S. at 9, 144 S.Ct. 1221 (emphasis omitted).

182.   Miller v. Johnson, 515 U.S. 900, 916, 115 S.Ct. 2475, 132 L.Ed.2d 762 (1995).See also, e.g., Hunt v. Cromartie, 526 U.S. 541, 551, 119 S.Ct. 1545, 143 L.Ed.2d 731 (1999) (“[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.”).

183.   E.g., Miller, 515 U.S. at 916, 115 S.Ct. 2475.

184.   E.g., Alexander, 602 U.S. at 7, 144 S.Ct. 1221 (citation modified).

185.   E.g., id. at 6, 144 S.Ct. 1221 (citation modified).See also U.S. Const. art. I, § 4, cl. 1 (“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 ch[oo]sing Senators.”).

186.   E.g., Alexander, 602 U.S. at 7, 144 S.Ct. 1221 (citation modified).

187.   E.g., Miller, 515 U.S. at 915, 115 S.Ct. 2475.

188.   E.g., id. at 915–16, 115 S.Ct. 2475.

189.   E.g., Alexander, 602 U.S. at 6, 144 S.Ct. 1221.

190.   E.g., id. at 10, 144 S.Ct. 1221.

191.   E.g., id. at 11, 144 S.Ct. 1221.

192.   E.g., Cooper, 581 U.S. at 292, 137 S.Ct. 1455.

193.   See infra Section III.B.8.

194.   See infra Section III.B.5.

195.   E.g., Miller, 515 U.S. at 911, 115 S.Ct. 2475.

196.   See supra Section II.D.

197.   See Brooks Prelim. Inj. Ex. 253, ECF No. 1326, at 1–2.

198.   See id.

199.   See id.

200.   For that reason, we reject the State Defendants’ argument that the DOJ Letter was not “a demand for race-based redistricting,” but was instead a demand to conduct race-neutral redistricting. Contra Defs.’ Post-Hr'g Br., ECF No. 1284, at 30–31 (emphasis omitted).Even if the State Defendants’ interpretation of the DOJ Letter was correct, that's not how the Legislature interpreted it.

201.   See Brooks Prelim. Inj. Ex. 265, ECF No. 1326-12, at 1–2.Cf. Cooper, 581 U.S. at 313, 137 S.Ct. 1455 (concluding that “the redistricters’ on-the-nose attainment of a 50% BVAP” supported the district court's finding that the legislature “deliberately redrew [the challenged district] as a majority-minority district”); see also infra Section III.B.5.b.

202.   See Cooper, 581 U.S. at 299–301, 137 S.Ct. 1455 (citation modified).

203.   Notably, the Legislature did not pass redistricting legislation during the first called special session due to a quorum break. See Prelim. Inj. Hr'g Tr. Day 8 (Afternoon), ECF No. 1344, at 109–110. The Governor then called a second special session, see Defs.’ Prelim. Inj. Ex. 1055, ECF No. 1373-16, at 2–3, during which the Legislature passed the 2025 Map. The Governor's August 15, 2025, proclamation placing redistricting on the agenda for the second special session omits any reference to DOJ's “constitutional concerns.” Contrast Defs.’ Prelim. Inj. Ex. 1054, ECF No. 1373-15, at 2–3 (directing the Legislature “to consider and act upon ․ [l]egislation that provides a revised congressional redistricting plan in light of constitutional concerns raised by the U.S. Department of Justice” (emphasis added)), with Defs.’ Prelim. Inj. Ex. 1055, ECF No. 1373-16, at 2–3 (merely directing the Legislature “[t]o consider and act upon ․ [l]egislation that provides a congressional redistricting plan” (emphasis added)). See also Prelim. Inj. Hr'g Tr. Day 8 (Afternoon), ECF No. 1344, at 49–50.We don't interpret that omission as evidence that the Governor abandoned the racial goals he had espoused in the media just four days earlier. See Brooks Prelim. Inj. Ex. 335-T, ECF No. 1328-1, at 1, 4–5 (Governor Abbott's August 11, 2025, interview proclaiming that he “wanted to remove ․ coalition districts and draw them in ways that ․ provide more seats for Hispanics”).Nor do we agree with the State Defendants’ suggestion that removing the reference to DOJ's constitutional concerns from the second proclamation somehow cleansed the first proclamation's racial taint. Contra Prelim. Inj. Hr'g Tr. Day 8 (Afternoon), ECF No. 1344, at 50 (“Q. ․ [E]ven if, as Plaintiffs allege, that the Governor's stated reasoning for adding the subject of redistricting to the call had some significance to the Legislature during the first legislative session, could the Legislature be legally permitted to consider that language during the Second Special Session?” | A. No.”). The map that the Legislature passed during the second session was largely identical to the first, indicating that racial considerations have already infected the map by the time the Governor issued the second proclamation. See Brooks Prelim. Inj. Ex. 264, ECF No. 1326-11, at 1–3 (showing the significant overlap between the map introduced in the first session and the map introduced in the second); Brooks Prelim. Inj. Ex. 266, ECF No. 1326-13, at 1–4 (showing the significant overlap between the map introduced in the second session and the enacted map).In any event, the Legislature acted under the DOJ Letter's directive even after the second proclamation. When the House passed the bill in the second session, the Speaker's press release explicitly stated that the House had just “delivered legislation to ․ address concerns raised by the Department of Justice.” See Brooks Prelim. Inj. Ex. 282, ECF No. 1326-28, at 1; see also infra Section III.B.3.d.i.

204.   See supra Section II.D.

205.   See supra Section II.C.

206.   See supra Sections II.D–E.President Trump's July 15, 2025, press statement that he “want[ed] the Republicans to draw ․ five seats” is not particularly probative of the motivation underlying the 2025 redistricting. Contra Defs.’ Prelim. Inj. Ex. 1352, ECF No. 1360-2, at 7–8; see also Prelim. Inj. Hr'g Tr. Day 2 (Morning), ECF No. 1415, at 127–29 (introducing that statement to support the State Defendants’ argument that Texas redistricted for political rather than racial reasons). By the time President Trump made that statement, DOJ had already asked Texas to redistrict for exclusively racial reasons on July 7, 2025, and the Governor had already asked the Legislature to redistrict based on DOJ's letter on July 9, 2025. See supra Sections II.D–E.

207.   See, e.g., Brooks Prelim. Inj. Ex. 335-T, ECF No. 1328-1, at 4–5 (“[W]e wanted to remove those coalition districts and draw them in ways that in fact turned out to provide more seats for Hispanics.”); see also supra Section II.E.

208.   See Brooks Prelim. Inj. Ex. 335-T, ECF No. 1328-1, at 5.See also Brooks Prelim. Inj. Ex. 332-T, ECF No. 1411-3, at 2 (“[W]e saw in the aftermath of the Trump election[ ] that an overwhelming number of Hispanics and [B]lacks as well as others[ ] chose to vote for Trump․ Democrats think they have an ownership right to voters who are Hispanic or Black. They're now learning the hard way. Those voters are supporting Republicans.”).

209.   See Brooks Prelim. Inj. Ex. 332-T, ECF No. 1411-3, at 2.

210.   See, e.g., Miller, 515 U.S. at 911, 115 S.Ct. 2475.

211.   The Governor's assertions regarding Hispanic voting preferences are factually inaccurate. The preliminary-injunction record indicates that Hispanic voters in the relevant areas of Texas still favor Democrats over Republicans by a comfortable margin. See Prelim. Inj. Hr'g Tr. Day 4 (Morning), ECF No. 1417, at 60–63. The record further indicates that the shift in Hispanic support towards President Trump in the 2024 general election did not carry over to other Republican candidates on the ballot. See id. at 61–63.

212.   See Bush v. Vera, 517 U.S. 952, 968, 116 S.Ct. 1941, 135 L.Ed.2d 248 (1996) (plurality opinion).See also Tenn. State Conf. of NAACP v. Lee, 746 F. Supp. 3d 473, 488 (M.D. Tenn. 2024) (“Just as a State should not use race to identify the schools that children may attend, so too it should not use race to determine the districts in which citizens should vote.”).

213.   See Jackson, ––– F.4th at ––––, 2025 WL 3019284, at *10.

214.   Miller, 515 U.S. at 911–12, 115 S.Ct. 2475 (citation modified).See also Bush, 517 U.S. at 968, 116 S.Ct. 1941 (1996) (“[T]o the extent that race is used as a proxy for political characteristics, a racial stereotype requiring strict scrutiny is in operation.”).

215.   See supra Section II.E.

216.   See, e.g., Alexander, 602 U.S. at 7, 144 S.Ct. 1221 (citation modified).

217.   See, e.g., id. (citation modified).

218.   Brnovich v. Democratic Nat'l Comm., 594 U.S. 647, 689–90, 141 S.Ct. 2321, 210 L.Ed.2d 753 (2021).

219.   See Common Cause Fla. v. Byrd, 726 F. Supp. 3d 1322, 1364 (N.D. Fla. 2024) (“A public and collective decision-making body, like the ․ Legislature, is answerable only for its own unconstitutional actions and motivations.” (emphasis omitted)).

220.   Id. at 1364–65.

221.   See, e.g., id. at 1343–44.

222.   See, e.g., id.

223.   Id. at 1361.

224.   Id. (citation modified).

225.   Id. at 1363.

226.   Id. at 1366 (emphases omitted).

227.   Brooks Prelim. Inj. Ex. 282, ECF No. 1326-28, at 1 (emphasis added); see also Prelim. Inj. Hr'g Tr. Day 8 (Afternoon), ECF No. 1344, at 132–33.

228.   See Common Cause Fla., 726 F. Supp. 3d at 1363 (stating that “[r]atification of another's discriminatory motives ․ may be demonstrated with evidence that the decision-makers knowingly chose a particular course of action for the purpose of giving effect to the discriminatory motives”).

229.   See infra notes 277, 286, 321 and accompanying text.

230.   Brooks Prelim. Inj. Ex. 282, ECF No. 1326-28, at 1 (emphasis added).

231.   See id.

232.   Brooks Prelim. Inj. Ex. 327-T, ECF No. 1327-27, at 2–3; see also Prelim. Inj. Hr'g Tr. Day 1 (Afternoon), ECF No. 1337, at 68–69.

233.   Brooks Prelim. Inj. Ex. 339-T, ECF No. 1411-5, at 3; see also Prelim. Inj. Hr'g Tr. Day 9 (Afternoon), ECF No. 1345, at 67 (admitting that interview into the record).

234.   See Abrams v. Johnson, 521 U.S. 74, 87–88, 117 S.Ct. 1925, 138 L.Ed.2d 285 (1997).

235.   We refer to Representative Hunter as “Chairman” because he was the Chair of the Calendars Committee during the 89th Legislature. We emphasize that Representative Vasut—not Representative Hunter—was the Chair of both the House Redistricting Committee and the House Select Committee on Congressional Redistricting in 2025. See, e.g., Prelim. Inj. Hr'g Tr. Day 8 (Afternoon), ECF No. 1344, at 60; see also infra note 285 and accompanying text.

236.   See, e.g., Prelim. Inj. Hr'g Tr. Day 8 (Afternoon), ECF No. 1344, at 106.

237.   To be clear, we do not treat Chairman Hunter's floor statements as dispositive of the intent of the Legislature as a whole. “[S]tatements of individual legislators”—“even the sponsors of legislation”— “should not be given controlling effect.” N. & S. Rivers Watershed Ass'n v. Town of Scituate, 949 F.2d 552, 555 n.6 (1st Cir. 1991), overruled on other grounds by Blackstone Headwaters Coal., Inc. v. Gallo Builders, Inc., 32 F.4th 99 (1st Cir. 2022); see also, e.g., Fusilier v. Landry, 963 F.3d 447, 466 (5th Cir. 2020) (cautioning “against overemphasizing statements from individual legislators”).All we're saying is that (1) Chairman Hunter's statements about his reasons for introducing and passing the redistricting bill are relevant when assessing the intent of the Legislature as a whole, and (2) Chairman Hunter's role as the redistricting bill's sponsor makes his statements more probative than those of rank-and-file legislators who had minimal personal involvement with the bill. See, e.g., Brock v. Pierce County, 476 U.S. 253, 263, 106 S.Ct. 1834, 90 L.Ed.2d 248 (1986) (stating that although statements by a bill's sponsor “should not be given controlling effect,” they nonetheless “provide evidence of [the legislature's] intent” if “they are consistent with the statutory language and other legislative history”); Campbell v. McCarthy, 952 F.3d 193, 204 (4th Cir. 2020) (“In determining legislative intent, the statements of a bill's sponsor made during debate are entitled to weight.” (citation modified)).Our panel reached the same conclusion in our previous preliminary-injunction opinion in this case. See 1st Prelim. Inj. Op., 601 F. Supp. 3d at 175 n.13 (“[S]tatements of discriminatory intent by a committee chair made during floor debate would doubtless be of some weight in judging the intentions of the body as a whole, particularly at this preliminary stage.”).

238.   See Brooks Prelim. Inj. Ex. 309-T, ECF No. 1327-9, at 46; see also H.B. 4, 89th Leg., 1st Spec. Sess. (Tex. 2025).

239.   See, e.g., Brooks Prelim. Inj. Ex. 315-T, ECF No. 1327-15, at 4–6 (identifying changes the mapmaker made between the first special session and the second); Brooks Prelim. Inj. Ex. 316-T, ECF No. 1327-16, at 32–33 (Chairman Hunter's statement that “[h]e and [his] lawyers” made changes between the version introduced in the first legislative session and the enacted version to “increase[ ] Republican political performance”); see also H.B. 4, 89th Leg., 2d Spec. Sess. (Tex. 2025).

240.   See Brooks Prelim. Inj. Ex. 309-T, ECF No. 1327-9, at 1, 45–46.A “layout” is when a bill's sponsor first presents the bill to the body in a public hearing. Prelim. Inj. Hr'g Tr. Day 1 (Afternoon), ECF No. 1337, at 43. The layout was Chairman Hunter's “first opportunity to talk about the map as it was introduced.” Id.

241.   See Prelim. Inj. Hr'g Tr. Day 1 (Afternoon), ECF No. 1337, at 44 (“Q. Did anybody ask Chairman Hunter at this stage of the proceedings, ‘Tell us what the racial makeup of these five new districts are that you're drawing?’ | [SPEAKER MOODY:] No. This is his layout of the bill, so this is him explaining the bill to the members and to the public for the first time.”).See also Jackson, ––– F.4th at ––––, 2025 WL 3019284, at *10 (indicating that statements related to race are more probative of intent when unprompted, as opposed to a response to a question phrased in racial terms).

242.   See Brooks Prelim. Inj. Ex. 309-T, ECF No. 1327-9, at 54.

243.   See Prelim. Inj. Hr'g Tr. Day 1 (Afternoon), ECF No. 1337, at 48 (“Q. So these comments that Chairman Hunter is giving, are they in response to a question? | [SPEAKER MOODY:] No, I don't believe so. I think this is all still part of his layout. | Q. In other words, this is something he came in with his own notion to say? | A. I mean, that's typically how a layout works.”).

244.   See Brooks Prelim. Inj. Ex. 309-T, ECF No. 1327-9, at 58 (“[CHAIRMAN HUNTER:] In the 2021 plan, there were 7 Hispanic citizen voting age districts; and under this plan, there are 8.”).

245.   See id. (“[CHAIRMAN HUNTER:] There were no majority Black CVAP ․ districts under the 2021 plan. In the proposed plan today, there are 2 ․”).

246.   See id. at 57–58 (“[CHAIRMAN HUNTER:] Congressional District 9, the new district, has a 50.5-percent Hispanic CVAP. CD 28 ․ has an 86.70-percent Hispanic CVAP․ CD 34, 71.9 percent, is now a Hispanic CVAP. And CD 35, which is in San Antonio, is now a 51.6-percent Hispanic CVAP.”).

247.   See id. at 58 (“[CHAIRMAN HUNTER:] CD 18 is now 50.8 percent Black CVAP; in 2021 it was 38.8. CD 30 is now 50.2 percent Black CVAP; in 2021 it was 46 percent.”).

248.   See also Prelim. Inj. Hr'g Tr. Day 1 (Afternoon), ECF No. 1337, at 44 (Speaker Moody's testimony that he saw the notes that Chairman Hunter had prepared to deliver the layout, which contained “Black CVAP, HCVAP[,] [t]he shifts between this map and that map,” etc.); id. at 45 (“[SPEAKER MOODY:] [T]hey were like bulleted out ․ it looked like talking points․ like you're presenting a bill, you've got that broken down.”).

249.   See supra notes 182–183 and accompanying text.

250.   See, e.g., Defs.’ Post-Hr'g Br., ECF No. 1284, at 27 (“Rep. Hunter was criticized for not providing the racial makeup in 2021 ․ Democrat legislators wanted racial data during the [2025] layout․ In the [Texas S]enate, Sen. Menendez criticized Sen. King for not providing racial data like Rep. Hunter.”); Prelim. Inj. Hr'g Tr. Day 8 (Afternoon), ECF No. 1344, at 129 (“[CHAIRMAN VASUT:] [T]he last time we went through this in 2021 ․ [Chairman Hunter] was asked questions about CVAP by everybody, and every amendment that came up, it was constantly a question asked, particularly by members of the Democratic Party.”).

251.   See, e.g., Alexander, 602 U.S. at 10, 144 S.Ct. 1221 (“Th[e] presumption of legislative good faith directs district courts to draw the inference that cuts in the legislature's favor when confronted with evidence that could plausibly support multiple conclusions.”); see also supra notes 189–190 and accompanying text.

252.   See Brooks Prelim. Inj. Ex. 309-T, ECF No. 1327-9, at 59.

253.   See, e.g., id. at 95 (“REP. PIERSON: ․ Chairman Hunter, I just want to say: thank you for bringing the bill. I'm proud to be a joint author.”).

254.   See id. at 99–101 (“REP. PIERSON: ․ The stakeholders who testified during the field hearings [that the Legislature conducted before Chairman Hunter introduced the redistricting bill] testified that the population of Black voters in the state did not have proportionate representation․ Well, this current map that you have submitted actually shows where there's not just one but two majority Black CVAP districts drawn on this map; is that true? | REP. HUNTER: That is correct. And let me give everybody details. CD 18 is now 50.8 percent Black CVAP; in 2021 it was only 38.3 percent. CD 30 is now 50.2 percent Black CVAP; in 2021 it was 46 percent. | REP. PIERSON: So that's two Black CVAP districts. How many Black districts are there on the [2021 Map]? | REP. HUNTER: I don't have all the counts on that. | REP. PIERSON: The answer is zero. So overall, Black voters in the state of Texas go from zero to two majority Black CVAP seats out of the 38 seats in Texas; is that accurate? | REP. HUNTER: It's accurate ․ | REP. PIERSON: ․ So would it be fair to say that your proposed map directly resolves many of the concerns that were expressed during those field hearings in your proposed map and would, in fact, strengthen minority representation in our state. Would you agree? | REP. HUNTER: The answer is, ‘Yes.’ ”).See also Brooks Prelim. Inj. Ex. 316-T, ECF No. 1327-16, at 370, 373 (“REP. PIERSON: ․ They say we're diluting the minority districts. They call us racist, but the facts don't match your rhetoric. Texas currently has zero Black CVAP districts. And under the new map, there are two. Now, I haven't been to third grade in a really long time, but when you go from zero to two, that's an increase; or perhaps you're using liberal logic․ Increasing minority representation is the right thing to do ․”).

255.   See, e.g., Brooks Prelim. Inj. Ex. 309-T, ECF No. 1327-9, at 59 (“REP. SPILLER: ․ [T]hank you for allowing me the opportunity to joint author [the redistricting bill].”)

256.   See Brooks Prelim. Inj. Ex. 316-T, ECF No. 1327-16, at 82 (“REP. SPILLER: ․ And this claim, that a lot of this stuff is racially motivated and race negative—let me ask you, and you've touched on it before, but we went under the [2021 Map] from zero majority Black CVAP districts in the State of Texas. And now, under your map, we added two to the list [that were] there. There there [sic] are two majority Black CVAP districts, correct? | REP. HUNTER: Correct. 18 and ․ 30. | REP. SPILLER: And on the current map we have seven majority Hispanic CVAP districts, and that is increased ․ under your [b]ill to 8. So, we're adding one more majority Hispanic CVAP district, correct? | REP. HUNTER: Yeah.”).

257.   Brooks Prelim. Inj. Ex. 309-T, ECF No. 1327-9, at 54 (emphasis added).See also Brooks Prelim. Inj. Ex. 316-T, ECF No. 1327-16, at 29 (“It's important to note—please note members—four of the five new districts are majority/minority Hispanic ․”).

258.   See Brooks Prelim. Inj. Ex. 316-T, ECF No. 1327-16, at 122 (“REP. HUNTER: ․ [W]e created four out of five new seats of [sic] Hispanic majority. I would say that's great. That doesn't ensure that a political party wins them, but the Hispanic—four out of five Hispanic majority out of those new districts—that's a pretty strong message, and it's good.”).

259.   See, e.g., Brooks Prelim. Inj. Ex. 309-T, ECF No. 1327-9, at 77 (Chairman Hunter's statement that “the percentage for Black CVAP [was] better” under his proposed map); Brooks Prelim. Inj. Ex. 316-T, ECF No. 1327-16, at 220 (“REP. HUNTER: ․ CD 18 now becomes a 50.8 percent Black CVAP. [The 2021 version of CD 18 was] 38.8 percent [Black] CVAP. I think my map is much more improving [sic].”).

260.   See, e.g., Brooks Prelim. Inj. Ex. 316-T, ECF No. 1327-16, at 357–58 (“REP. DUTTON: So, what else happened between the last redistricting and this [b]ill that causes you comfort to make these changes? | REP. HUNTER: Well, number one, in 2024 the Petteway case ․ was decided․ And there they said, Section 2 of the Voting Rights Act does not authorize separately protected minority groups to aggregate their populations for purposes of a vote dilution claim, and it does not require political subdivisions to draw precinct lines for these particular groups. So, this changed a lot of the law that happened in 2021.”).See also Brooks Prelim. Inj. Ex. 309-T, ECF No. 1327-9, at 53 (“[CHAIRMAN HUNTER:] Under the Fifth Circuit—and this is a recent decision; they changed the law ․ [c]oalition districts were held by the Court that Section 2 no longer requires the drawing of coalition districts.”).See also Brooks Prelim. Inj. Ex. 315-T, ECF No. 1327-15, at 6, 11, 29 (similarly referencing Petteway); Brooks Prelim. Inj. Ex. 316-T, ECF No. 1327-16, at 34, 49, 93, 121, 215, 326, 328, 329, 343–44, 357–58 (same).

261.   Brooks Prelim. Inj. Ex. 316-T, ECF No. 1327-16, at 28–29 (“[T]he [Fifth] Circuit, in Petteway v. Galveston indicates that the law has changed. The court held that Section 2 does not require us to draw coalition districts. So, giving partisan political performance as an acceptable reason and clarification from these courts, we have redrawn the congressional map with that emphasis.”).See also Brooks Prelim. Inj. Ex. 309-T, ECF No. 1327-9, at 53 (similarly stating that the bill authors had “redrawn the congressional map” based in part on the “clarification from the Fifth Circuit on coalition districts”).

262.   See Brooks Prelim. Inj. Ex. 316-T, ECF No. 1327-16, at 110 (“You have had a discussion about a U.S. Supreme Court [case] and a [Fifth] Circuit [case] that has new impact on the law, which gives us justification further to look at redistricting. And we looked at redistricting, and we created five new congressional seats, four are Hispanic majority.”).

263.   See id. at 122 (“[Petteway] says there's not a requirement that you have to use coalition [districts]․ So, today, this map is taking th[at] in factor [sic].”).

264.   See id. at 76–77 (“REP. SPILLER: ․ [U]nder the [2021 Map], there are coalition districts that were created as such in ’21 because of the law as it existed in Texas under the 5th Circuit at that time. Is that fair to say? | REP. HUNTER: That is correct ․”); id. at 77 (“REP. SPILLER: ․ So, now, in Texas, one of the reasons that we're [redistricting] now is that, we feel compelled to because of the Petteway case and the ruling in the Petteway case ․ as it relates to these coalition districts, correct? | REP. HUNTER: Well, I think it's a combination, Mr. Spiller. I think you have a U.S. Supreme Court [case], Rucho. You have a Fifth Circuit [case], Petteway. The combination of both of those cases are involved in this map.”).

265.   See Brooks Prelim. Inj. Ex. 309-T, ECF No. 1327-9, at 75 (“REP. SPILLER: I would submit to you that [CD 18] is currently a coalition district; under [your proposed map], it would not be. Coalition districts are the type that are addressed in the Petteway case; and so I would submit to you that it goes from a coalition district to a majority Black CVAP district, being 58.1 [sic] percent Black. | REP. HUNTER: That is correct.”); Brooks Prelim. Inj. Ex. 316-T, ECF No. 1327-16, at 79 (“[REP. SPILLER:] [CD 18 was] one of these coalition districts, and under HB 4, [it] changes to a majority Black CVAP district. Is that correct? | REP. HUNTER: That is correct. It is now 50.71 percent Black CVAP. In 2021, it was 38.99 percent Black CVAP. | REP. SPILLER: And so, previously, Black voters in that district did not hold a majority, but under your [b]ill, under HB 4, they actually do. Is that correct? | REP. HUNTER: That is correct.”); id. at 80 (“REP. SPILLER: ․ District 9 ․ was also ․ a coalition district and the [type of] district that was addressed in the Petteway case. And now, under your HB 4, it changed from a coalition district to a majority Hispanic CVAP district. Is that correct? | REP. HUNTER: Yes. For the record, the Hispanic CVAP of Congressional District 9 under this plan ․ is 50.15 percent. In 2021, it was 25.73 percent.”).

266.   See Brooks Prelim. Inj. Ex. 309-T, ECF No. 1327-9, at 81–82 (“REP. SPILLER: ․ So, in summary, is it your testimony here today that you believe that the map created under [your bill] is in compliance with the Petteway case ․? | REP. HUNTER: Yes.”).

267.   See Prelim. Inj. Hr'g Tr. Day 1 (Afternoon), ECF No. 1337, at 51 (“REPRESENTATIVE TURNER: ․ CD18 was purposely altered to a Black CVAP majority district rather than a 38.8 percent Black CVAP district, right? | REPRESENTATIVE HUNTER: CD18 was drawn to be a 50.81 percent CVAP, which is 11.82 change plus․ | REPRESENTATIVE TURNER: ․ And similarly, the proposed CD35 was purposely changed to increase its Hispanic CVAP to be about 50 percent, correct? ․ | REPRESENTATIVE HUNTER: 51.57 percent. And it also has political performance involved ․ in all of this.”).

268.   Cf. Cooper, 581 U.S. at 308 n.7, 137 S.Ct. 1455 (“[I]f legislators use race as their predominant districting criterion with the end goal of advancing their partisan interests—perhaps thinking that a proposed district is more ‘sellable’ as a race-based VRA compliance measure than as a political gerrymander and will accomplish much the same thing—their action still triggers strict scrutiny.”).

269.   See supra notes 252–259 and accompanying text.

270.   See supra Sections II.E & III.B.3.b.ii.

271.   See supra notes 264–266; see also supra Section III.B.3.d.iii.

272.   E.g., Bethune-Hill, 580 U.S. at 189, 137 S.Ct. 788 (citation modified).

273.   See, e.g., Abbott v. Perez, 585 U.S. at 585–86, 138 S.Ct. 2305 (“The Equal Protection Clause forbids racial gerrymandering, that is, intentionally assigning citizens to a district on the basis of race without sufficient justification.” (citation modified)).

274.   See, e.g., Brooks Prelim. Inj. Ex. 309-T, ECF No. 1327-9, at 52 (“[W]e are allowed to draw congressional districts ․ based on political performance, political partisanship. That's recognized by the United States Supreme Court. These districts were drawn ․ primarily using political performance ․”); Brooks Prelim. Inj. Ex. 316-T, ECF No. 1327-16, at 28 (“You want transparency? Here's the U.S. Supreme Court legal transparency. The underlying goal of this plan is straightforward, improve Republican political performance.”); Brooks Prelim. Inj. Ex. 315-T, ECF No. 1327-15, at 4 (“[T]his map is based on partisanship, political performance ․ [I]t has enhanced and increased Republican partisanship enhanced performance [sic].”).

275.   See, e.g., Brooks Prelim. Inj. Ex. 315-T, ECF No. 1327-15, at 6 (“So based on Rucho, based on Petteway, this, Mr. Chairman, is what the Committee substitute addresses.”); id. at 29 (“I'm following Rucho, the U.S. Supreme Court [sic] in Petteway. And it allows us to do this ․”); Brooks Prelim. Inj. Ex. 316-T, ECF No. 1327-16, at 77 (“[I]t's a combination ․ I think you have a U.S. Supreme Court [case], Rucho. You have a Fifth Circuit [case], Petteway. The combination of both of those cases are involved in this map.”).

276.   See, e.g., Brooks Prelim. Inj. Ex. 316-T, ECF No. 1327-16, at 28 (“We are allowed to draw congressional districts on the basis of political performance as recognized by the U.S. Supreme Court in Rucho v. Common Cause.”); Brooks Prelim. Inj. Ex. 309-T, ECF No. 1327-9, at 68–69 (“REP. SPILLER: ․ Is it fair to say that the map in HB 4 based [sic] on political performance or partisan performance? | REP. HUNTER: The answer is, ‘Yes.’ And I want everybody to know that․ [I]t's based on Rucho, a United States Supreme Court case.”).

277.   See Prelim. Inj. Hr'g Tr. Day 2 (Morning), ECF No. 1415, at 131 (“[T]he Department of Justice letter is a letter․ That's not guiding me. I'm presenting a plan. And they can review the plan․ And if they ․ believe that I've addressed issues, good. If they believe I haven't, good. But whatever they've sent, I'm not ignoring, I'm not accepting. I'm doing this plan. So whatever their involvement is, they just sent a letter, as far as I'm concerned.”).See also Brooks Prelim. Inj. Ex. 309-T, ECF No. 1327-9, at 61 (“I don't know if [the 2025 redistricting] was caused by the Department of Justice. I keep hearing that, and I keep hearing about a letter. All I know is we're here by proclamation of the [G]overnor. Now, what the letter has to do with it, I've got no personal knowledge. I have no knowledge. And I will tell you: I don't know what that has to do with this. That wasn't part of me. All I know is we had a Special Session called and this was the topic and I agreed, by the request of [Chairman Vasut], to file this bill.”).See also Brooks Prelim. Inj. Ex. 316-T, ECF No. 1327-16, at 108–09 (“REP. GARVIN HAWKINS: ․ What was your understanding of the DOJ's letter regarding redistricting? | REP. HUNTER: Well, my answer hasn't changed one bit. There was a DOJ letter. It's out there. DOJ will get to review this. I have no criticism. I have no feedback. They do what they want. We do what we want. Nothing any different. | REP. GARVIN HAWKINS: Okay. So you've read [the DOJ Letter] now․ | REP. HUNTER: I have not ․ I just read parts of it.”).However, Chairman Hunter also made a statement suggesting that the lawyers he hired to produce the map had “t[aken the DOJ Letter] into account” when creating the map. See Brooks Prelim. Inj. Ex. 316-T, ECF No. 1327-16, at 111 (“REP. HUNTER: Look, there was a DOJ letter․ [T]he lawyers looked at it, took it all into account, and then we came up with this plan which set it dot [sic]. It mapped the threshold. It mapped the requirements.”).

278.   See, e.g., Brooks Prelim. Inj. Ex. 309-T, ECF No. 1327-9, at 95 (“REP. PIERSON: ․ This has been redrawn, as you stated in your opening statement, to reflect political performance but also compactness; is that right? | REP. HUNTER: Yes.”).

279.   See Brooks Prelim. Inj. Ex. 316-T, ECF No. 1327-16, at 344 (“I didn't go at coalition districts. I had the lawyers come up with five seats and enhance the Republican performance, and that's what we did. I didn't go at a coalition.”).

280.   See supra Section II.A.4.

281.   See supra Section II.C (recounting that requests to redistrict for purely partisan reasons went nowhere).

282.   See Rucho v. Common Cause, 588 U.S. 684, 139 S.Ct. 2484 (decided June 27, 2019).

283.   See Petteway v. Galveston County, 111 F.4th 596, 600 (5th Cir. 2024) (en banc) (decided August 1, 2024); see also supra Section II.B.

284.   See, e.g., Prelim. Inj. Hr'g Tr. Day 5 (Afternoon), ECF No. 1341, at 77.

285.   See, e.g., Prelim. Inj. Hr'g Tr. Day 8 (Afternoon), ECF No. 1344, at 60.

286.   See, e.g., Prelim. Inj. Hr'g Tr. Day 5 (Afternoon), ECF No. 1341, at 80 (“Q. What significance did [the DOJ L]etter play in Texas redistricting in 2025? | A. Well, I can't speak for everyone else in the Legislature, but for me it didn't really carry any significance.”); Prelim. Inj. Hr'g Tr. Day 1 (Morning), ECF No. 1414, at 107 (“[M]y support ․ of [the redistricting bill] does not in any way take into account the DOJ letter.”); Prelim. Inj. Hr'g Tr. Day 8 (Morning), ECF No. 1421, at 139 (“I honestly never took the [DOJ L]etter into account. I didn't think it mattered.”).See also Brooks Prelim. Inj. Ex. 319-T, ECF No. 1327-12, at 208 (Chairman King's statement on the Senate floor that he “thought the DOJ Letter ․ unnecessarily confused the redistricting process”).

287.   See, e.g., Prelim. Inj. Hr'g Tr. Day 5 (Afternoon), ECF No. 1341, at 111 (“Q. ․ Did you review any racial data associated with [the redistricting bill]? | A. No, I didn't look at any racial data.”); Prelim. Inj. Hr'g Tr. Day 2 (Morning), ECF No. 1415, at 32 (“I have not taken racial data into consideration in drawing the map.”).

288.   See, e.g., Prelim. Inj. Hr'g Tr. Day 5 (Afternoon), ECF No. 1341, at 111 (“Q. To your knowledge, was race used in the drawing of the map? | A. It was not.”).

289.   See, e.g., id. at 85.

290.   See, e.g., id. at 115 (“Q. And did the map that ultimately passed both houses of the Legislature, did it meet all three of your stated goals? | A. Yes. It was a legal map, it should elect more Republican members to the U.S. House, and it did improve compactness in some districts.”); Prelim. Inj. Hr'g Tr. Day 8 (Afternoon), ECF No. 1344, at 52–53 (similar).

291.   See, e.g., Prelim. Inj. Hr'g Tr. Day 5 (Afternoon), ECF No. 1341, at 77 (“Q. And what is your understanding of why ․ redistricting was being considered in Texas? | A. Well, it was absolutely to create more Republican seats in the U.S. Congress.”); id. at 99–100 (“Q. And so was Texas Congressional Redistricting, and the reasons for it, widely publicized both prior to it being placed on the call and during the redistricting effort? | A. Oh, yes. I think it was apparent to everyone the purpose of it was partisan ․”).

292.   See supra notes 155–158 and accompanying text.

293.   Prelim. Inj. Hr'g Tr. Day 5 (Afternoon), ECF No. 1341, at 91.See also, e.g., id. at 121–22 (“The Lieutenant Governor ․ had told me that [he and the Speaker had] divided up all the major issues between the House and the Senate․ He informed me that the House would take the lead [on redistricting] ․”).

294.   See id. at 91 (“Q. Did you play any role in drawing the map for [the Senate counterpart to the House redistricting bill] during the first Special Session? | A. No, none whatsoever. | Q. And did you draw any map for redistricting in 2025? | A. No, I did not. | Q. Did you open any map-drawing software? | A. No, I did not.”).See also Prelim. Inj. Hr'g Tr. Day 8 (Morning), ECF No. 1421, at 127 (“Q. ․ [Y]ou only saw the final product, right? You only saw the versions that were filed in the House that you then filed during each of the special sessions, correct? | A. That is correct. | Q. You weren't involved in any interim steps of the map, true? | A. That is correct.”).

295.   See, e.g., Prelim. Inj. Hr'g Tr. Day 5 (Afternoon), ECF No. 1341, at 91 (“A. ․ [The House] passed their bill out of committee and then, before it got to the floor, the Democrats broke quorum and left the state. And so at that point I went ahead and filed the companion bill, which was SB4. | Q. Where did you get the map that was associated with Senate Bill 4? | A. Well, it was the same map that was being considered by the House.”).

296.   Prelim. Inj. Hr'g Tr. Day 1 (Morning), ECF No. 1414, at 108.

297.   See Prelim. Inj. Hr'g Tr. Day 8 (Morning), ECF No. 1421, at 140–41 (“Q. ․ Do you have any idea when it is that the map that Mr. Kincaid drew landed with the lawyers for Chairman Hunter? | A. No. | Q. The testimony here is that that took place on July the 23rd. And it sounds to me like you were out of the loop in terms of the delivery of that map. Is that fair to say? | A. Yes.”).See also Prelim. Inj. Hr'g Tr. Day 1 (Morning), ECF No. 1414, at 119 (“CHAIRMAN KING: ․ [The mapdrawer, Adam Kincaid] called me and asked me if I was aware that the House was going to be putting out a map that had some changes from the original H.B. 4. And I said, no, I wasn't.”).See also Prelim. Inj. Hr'g Tr. Day 1 (Morning), ECF No. 1414, at 120 (“SENATOR GUTIERREZ: ․ There were some changes between the final version of H.B. 4 and the committee sub[stitute]? ․ My understanding of that is those changes were made at the behest of incumbent congresspeople. Is that accurate? | CHAIRMAN KING: I do not know.”).

298.   See infra Section III.B.4.d.

299.   See, e.g., Prelim. Inj. Hr'g Tr. Day 5 (Afternoon), ECF No. 1341, at 82 (Chairman King's testimony); Prelim. Inj. Hr'g Tr. Day 6 (Afternoon), ECF No. 1342, at 20 (Mr. Kincaid's testimony).

300.   See Prelim. Inj. Hr'g Tr. Day 5 (Afternoon), ECF No. 1341, at 82 (“I told him at that—when we met that I would not—or that I would probably be chairing the Redistricting Committee and that I preferred that we not discuss the redistricting maps.”); id. at 118 (“Q. ․ [W]hat you stated here today is that you told Mr. Kincaid you didn't want to hear anything about the Texas Redistricting Map. Did I hear that correctly? | A. Yes.”); id. at 119 (“Q. ․ Why did you tell Adam Kincaid you didn't want to know anything about the Texas map that you were about to facilitate the passage of? | A. ․ I wanted all information that came to me to come in a public forum.”); id. (“I said, ‘Let's not talk about the map.’ ”).See also Prelim. Inj. Hr'g Tr. Day 1 (Morning), ECF No. 1414, at 117–18 (Chairman King's floor statement that he “specifically told” Mr. Kincaid: “Don't tell me anything you are doing with regard to map drawing. Don't tell me about the details of any map if you are involved in it.”); Prelim. Inj. Hr'g Tr. Day 5 (Afternoon), ECF No. 1341, at 128 (Chairman King's floor statement that he “specifically told” Mr. Kincaid: “Don't tell me anything about the maps you're drawing. I don't want to discuss that.”).

301.   See Prelim. Inj. Hr'g Tr. Day 6 (Afternoon), ECF No. 1342, at 20–22 (“Q. And what did you discuss? | [MR. KINCAID:] ․ He said, ‘How many seats are we talking?’ I said, ‘Five seats. It's going to be a five-seat pickup.’ ․ | Q. ․ But you did talk about the map? | A. Broadly, yes. There was kind of open questioning at that point in time whether or not we would actually be able to pick up five seats․ | Q. And he was curious about that? | A. Yeah. He was curious, like, ‘Is it actually five seats?’ And I said, ‘Yes, five seats.’ | Q. And you confirmed that for him? | A. I believe so․ | Q. Do you remember anything else he said to you in that meeting? | A. He mentioned something about, you know, getting the map done—or, you know, working together to get the map done, something along those lines.”).

302.   See Prelim. Inj. Hr'g Tr. Day 5 (Afternoon), ECF No. 1341, at 119 (“Q. You left that meeting with not a bit of knowledge over what this map would look like? | [CHAIRMAN KING:] I don't recall us discussing any details of the map․ I said, ‘Let's not talk about the map.’ | Q. He didn't tell you how many Republican seats might be harvested? | A. Not that I recall.”).

303.   See Prelim. Inj. Hr'g Tr. Day 6 (Afternoon), ECF No. 1342, at 20–22 (“[ADAM KINCAID:] ․ He said, ‘How many seats are we talking?’ I said, ‘Five seats. It's going to be a five-seat pickup.’ ․ | Q. And he was curious about that? | A. Yeah. He was curious, like, ‘Is it actually five seats?’ And I said, ‘Yes, five seats.’ ”).

304.   See Prelim. Inj. Hr'g Tr. Day 8 (Morning), ECF No. 1421, at 131–32 (“Q. ․ I specifically asked you if you were told [during the ALEC] meeting whether or not the map was going to make changes to five districts․ And you said, no, I didn't want to know anything about the map. That was your testimony here. | [CHAIRMAN KING:] My recollection of the meeting was that when we sat down and I told Adam it looks like I'm going to be the chairman of the committee and so I don't want to talk anything about the map. | Q. And so if it's been stated under oath here in this courtroom in that chair by a different witness that ․ you specifically asked about the number of districts that would be affected and were told five would be affected, that testimony was false, in your opinion? | A. It's either incorrect or I'm remembering incorrectly.”).

305.   Prelim. Inj. Hr'g Tr. Day 5 (Afternoon), ECF No. 1341, at 82–83.

306.   See Prelim. Inj. Hr'g Tr. Day 8 (Afternoon), ECF No. 1344, at 110 (testimony that the first-called special session adjourned on August 15, 2025).

307.   Prelim. Inj. Hr'g Tr. Day 5 (Afternoon), ECF No. 1341, at 83 (“I had been repeatedly asked on the floor and in hearings if racial data was used to draft the map. I had always answered that, to my knowledge, it was not. I finally just picked up the phone and called Adam [Kincaid] and said, ‘Adam, I just have one question to ask you. Did you use racial data in drawing the map?’ ”).See also Prelim. Inj. Hr'g Tr. Day 8 (Morning), ECF No. 1421, at 128–29 (“I did call [Mr. Kincaid] and ask him if he used racial data because I had been asked so many times on the floor and in committee. And I finally thought, well, I'll just call him and ask him. And so I picked up the phone and I said, [‘]Mr. Kincaid, just one question for you. I don't want to talk about the map. Did you use racial data in drawing this map?[’]”).

308.   See Prelim. Inj. Hr'g Tr. Day 8 (Morning), ECF No. 1421, at 129 (“[H]e responded, [‘]no, I did not.[’]”).

309.   See id. at 176; see also Brooks Prelim. Inj. Ex. 319-T, ECF No. 1327-19, at 14.

310.   See Prelim. Inj. Hr'g Tr. Day 8 (Morning), ECF No. 1421, at 176 (“SENATOR GUTIERREZ: And you don't know whether [the mapdrawer] looked at race in creating these maps, do you? | SENATOR KING: What I—no.”); see also Brooks Prelim. Inj. Ex. 319-T, ECF No. 1327-19, at 14.

311.   See Prelim. Inj. Hr'g Tr. Day 8 (Morning), ECF No. 1421, at 176; see also Brooks Prelim. Inj. Ex. 319-T, ECF No. 1327-19, at 14.

312.   Respectfully, we disbelieve Chairman King's assertion that his conversation with Mr. Kincaid about whether he used racial data simply slipped his mind during his exchange with Senator Gutierrez because Chairman King was drained from a lengthy legislative debate. Contra Prelim. Inj. Hr'g Tr. Day 8 (Morning), ECF No. 1421, at 129 (“[T]hat was in the middle or toward the end ․ of a four- to six-hour debate where I had been standing on the floor as the sole member representing that map, that bill. And, you know, it's just easy to make a mistake when you have been through that long a debate.”). We find it unlikely that Chairman King would have forgotten about a particularly recent conversation that he personally initiated with one of the key participants in the redistricting process about an issue critical to the map's legality. See id. (“Q. ․ [I]t seems like when Senator Gutierrez asked you about your contacts with Kincaid ․ that might be the first one at the front of your lobe that you would think of. Don't you agree? | A. I don't disagree with that ․”).

313.   Contrast, e.g., Prelim. Inj. Hr'g Tr. Day 8 (Morning), ECF No. 1421, at 130 (Chairman King's testimony at the preliminary-injunction hearing that he and Mr. Kincaid “bumped into each other” at ALEC), and Prelim. Inj. Hr'g Tr. Day 5 (Afternoon), ECF No. 1341, at 82 (similarly testifying that he and Mr. Kincaid “ran into each other at the ALEC ․ conference”), and Prelim. Inj. Hr'g Tr. Day 1 (Morning), ECF No. 1414, at 117 (Chairman King's floor statement that he “ran into [Adam Kincaid] at the ALEC Annual Conference”), with Prelim. Inj. Hr'g Tr. Day 6 (Afternoon), ECF No. 1342, at 21–22 (Q. And you just, like, happened to run into each other or had you made a plan to— | [ADAM KINCAID:] We planned to meet. | Q. Okay. How did that planning process happen? Did he call you, text you? | [ADAM KINCAID:] I think we spoke briefly the day before and said, ‘Hey, let's meet up at ALEC.’ | Q. Okay. And that was a phone call that he made? | [ADAM KINCAID:] Yeah. Or I made. I can't remember ․ who called who.”).See also Prelim. Inj. Hr'g Tr. Day 8 (Morning), ECF No. 1421, at 130–31 (“Q. And so if it's been stated under oath in this courtroom that you didn't run into Mr. Kincaid, you had a phone call with him the day before to arrange a meeting with Mr. Kincaid, that testimony is false, in your opinion? | [CHAIRMAN KING:] I don't remember a phone call with Adam Kincaid ․ during the ALEC.”).

314.   Contrast Prelim. Inj. Hr'g Tr. Day 5 (Afternoon), ECF No. 1341, at 142 (Chairman King's testimony at the preliminary-injunction hearing that he never “call[ed] up Adam Kincaid” to “ask him to come give his testimony to the Senate” because he'd “already sent him a letter formally inviting [Mr. Kincaid] to do so”), with Prelim. Inj. Hr'g Tr. Day 6 (Afternoon), ECF No. 1342, at 23 (“[ADAM KINCAID:] [Chairman King] called me one time during the hearings ․ He wanted to make sure ․ I had received the invitation to testify. | Q. Okay. And what did you say? | A. ‘Yes.’ | Q. And what else did you say? | A. ‘I couldn't make it to Austin.’ | Q. And how did he respond to that? | A. ‘Okay.’ | Q. And so ․ the general nature of that phone call was just calling you to ․ ask if you'd gotten the invitation? | A. He wanted to make sure I knew I was invited to come․ He made a point to say that he had made a promise to the Democrat he was working with to, you know—he would do that, so he did.”).See also Prelim. Inj. Hr'g Tr. Day 8 (Morning), ECF No. 1421, at 147–48 (“Q. Now, you recall the testimony here on Monday, I asked you ․ did it ever occur to you, since you had [Mr. Kincaid's] number and your colleagues were asking for it, to just call him up and ask him to come down and talk to the committee? ․ | [CHAIRMAN KING:] I do. | Q. And you said nobody ever asked me to do that. Do you remember that? | A. That sounds correct. Nobody did ever ask me to do that. | Q. And so if it's been the testimony here that in fact you did call Mr. Kincaid and ask him to come to the committee and testify, and he told you he was too busy and couldn't spare three days, that testimony, in your view, is false? ․ | A. It would be incorrect. I sent him a letter as an invitation.”).

315.   See Prelim. Inj. Hr'g Tr. Day 7 (Afternoon), ECF No. 1343, at 67–70 (“[L]et's stop pretending that this is all about race. It is about values. It is about representation—real representation. The fact that we are redrawing the maps is to ensure that ․ the people are able to have representation that reflects their values, not their last name, not their skin color․ And with that, members, I proudly stand and look forward to casting my vote in favor of House Bill 4.”); see also Defs.’ Prelim. Inj. Ex. 1325, ECF No. 1357-5, at 63–66.

316.   See Prelim. Inj. Hr'g Tr. Day 7 (Afternoon), ECF No. 1343, at 78–79 (“Q. This record reflects that at no point prior to [your speech on the floor] had you engaged in the legislative process on the map. Isn't that true? | A. Right, drawing maps or anything like that, no. | Q. There was [sic] no public comments from you in committee, either on the dais or as a participant, as a witness, or in any of the Senate floor proceedings on this map until that speech that we saw here in Court today. Is that fair to say? | A. Fair to say.”); id. at 80 (“Q. ․ [Y]ou weren't involved in the drawing of the lines that are made up of this new congressional map. Is that fair to say? | A. That's correct, sir.”).See also id. at 65 (Senator Hinojosa's testimony that he didn't serve on the Senate Redistricting Committee in 2025).

317.   See Fusilier, 963 F.3d at 466.

318.   United States v. O'Brien, 391 U.S. 367, 384, 88 S.Ct. 1673, 20 L.Ed.2d 672 (1968); see also, e.g., Fusilier, 963 F.3d at 466 (indicating that the quoted language from O'Brien remains good law).

319.   See supra note 237 and accompanying text.

320.   See, e.g., Prelim. Inj. Hr'g Tr. Day 8 (Afternoon), ECF No. 1344, at 117 (“I have not seen any evidence that this map was racially based. What I have seen is evidence that this map was politically based.”).

321.   See id. at 118 (“I disagree with the assumption that this process had anything to do with the DOJ letter. Yeah, they sent us a letter, but as you know, the proclamation called us in to do congressional redistricting, and we did congressional redistricting when we passed HB4 based off of political performance. So I frankly don't care what the DOJ letter said—and I think it's pretty clear that no one does․ So this bill was not based off of that DOJ letter. That bill was based off of improving political performance.”).See also id. at 81 (Chairman Vasut's testimony that the “DOJ [L]etter did not factor into [his] decision to make any vote on” the 2025 Map).

322.   See id. at 93–94 (“REPRESENTATIVE WU: Do you know whether the Governor's true intent is to remove coalition districts from Texas maps? ․ Would you be surprised if the Governor specifically said, point blank, quote, We have the ability now to draw maps that don't have coalition districts, end quote? ․ | REPRESENTATIVE VASUT: I'm aware of the Governor making remarks ․ [b]ut it's not the [C]hair's intention to be taking action based off the ․ expressed words of the Governor in a private setting. The Governor has given a proclamation, and, as the [C]hair has indicated, the [C]hair is going to act on that proclamation.”).

323.   See supra note 286 and accompanying text.

324.   See H.B. 4, 89th Leg., 1st Spec. Sess. (Tex. 2025); H.B. 4, 89th Leg., 2d Spec. Sess. (Tex. 2025).

325.   The Legislature made certain changes to the introduced map that Mr. Kincaid didn't draw. See, e.g., Prelim. Inj. Hr'g Tr. Day 6 (Morning), ECF No. 1419, at 159 (“Q. ․ [D]id the border you drew that we see in [the introduced version of the 2025 Map] between [CDs] 16 and 23 make it into the final map? | A. It did not. | Q. Did you draw the change between 16 and 23 between [the introduced map] and [the enacted map]? | A. I did not.”); id. at 173 (“The ․ change was in El Paso․ [T]hat was a change that had come from the Texas House. I did not draw that.”).No Plaintiff Group challenges those non-Kincaid-drawn districts on racial-gerrymandering grounds, see Chart of Claims, ECF No. 1208-1, at 2–4, so nothing about Mr. Kincaid's non-involvement with those districts affects our legal conclusions.

326.   See Prelim. Inj. Hr'g Tr. Day 6 (Morning), ECF No. 1419, at 33–34.

327.   See id. at 59–62.

328.   See id. at 76–191.We leave undetermined the issue of whether Mr. Kincaid's testimony amounted to undisclosed expert testimony that we must exclude from the preliminary-injunction record. See id. at 6–32 (the parties’ arguments on that issue). Either way the Court were to rule on that issue would not substantively change the Court's determination of the preliminary-injunction motions.

329.   See id. at 43 (“Q: Is the census data that comes preloaded in ․ your redistricting software, your map drawing software, is there racial data in there? | A. Yes.”); id. at 45 (“Q. Can you help the Court understand whether you can ever see racial data on this screen? How that happens? | A. Sure. So ․ [the software] has at the top left corner is a ․ demographics tab. You click on that․ [I]t will have all the census data that's provided by the [B]ureau ․ So you can select or not select ․ whatever datasets you are looking to work with.”).See also, e.g., Prelim. Inj. Hr'g Tr. Day 7 (Morning), ECF No. 1420, at 54 (“Q. ․ [I]f you had [CVAP] by race on your platform ․ you could also set it up in [a display box on the screen] so that every time you moved geography into and out of the district, even if you are using shading on political performance, you could watch those numbers changing as you are adding or taking out geography with respect to, for example, Hispanic [CVAP]? | A. You could do that, yes.”).

330.   See, e.g., Prelim. Inj. Hr'g Tr. Day 6 (Morning), ECF No. 1419, at 46 (“[W]hen you draw a map ․ do you have racial data visible? | A. I do not.”).See also id. at 57–58 (“Q. Do you ever become aware of racial data after you draw a map? | A. Yes. | Q. Do you then incorporate that racial data into your next draw of the map? ․ So let's say—have you ever been in a situation where you drew a map without looking at race? | A. Uh-huh. | Q. And then found out the racial makeup of a given district and then gone back and made changes to that district based on that racial understanding? | A. No.”); id. at 191 (“Q. Did you make any changes as a result of becoming aware of the racial demographic character of the districts in [the first version of the 2025 Map you drew]? | A. I did not. | Q. Why not? | A. I don't draw based off of race.”).

331.   See id. at 64.See also Prelim. Inj. Hr'g Tr. Day 7 (Morning), ECF No. 1420, at 125–26 (testifying that the requirement that “incumbent Republicans who lived in their seats stayed in their seats” was an “instruction[ ] from the White House”).

332.   See Prelim. Inj. Hr'g Tr. Day 6 (Morning), ECF No. 1419, at 65.See also, e.g., id. (“I was not allowed to take any incumbent Republican who was above 60 below 60.”); Prelim. Inj. Hr'g Tr. Day 7 (Morning), ECF No. 1420, at 125–26 (testifying that “the 60 percent threshold for incumbent [Republican] members of [C]ongress” was an “instruction[ ] from the White House”).

333.   See Prelim. Inj. Hr'g Tr. Day 6 (Morning), ECF No. 1419, at 65.See also Prelim. Inj. Hr'g Tr. Day 7 (Morning), ECF No. 1420, at 125–26 (testifying that the requirement not to “decrease [the partisan performance of] the districts that were under 60 percent” was an “instruction[ ] from the White House”).See also Prelim. Inj. Hr'g Tr. Day 6 (Morning), ECF No. 1419, at 65–66 (defining PVI); Prelim. Inj. Hr'g Tr. Day 9 (Morning), ECF No. 1422, at 59–61 (expert testimony further explaining how PVI is measured).

334.   See Prelim. Inj. Hr'g Tr. Day 6 (Morning), ECF No. 1419, at 67 (“Another [criterion] was the five pickup opportunities․ five districts that Republicans could gain that we currently did not hold in the 2026 midterms.”).Mr. Kincaid testified that he was free to decide which specific districts to flip, and that he based those decisions on the “political realities as [he] worked through the map.” See Prelim. Inj. Hr'g Tr. Day 7 (Morning), ECF No. 1420, at 129–30.

335.   See Prelim. Inj. Hr'g Tr. Day 6 (Morning), ECF No. 1419, at 67 (“[T]he five [new districts], at a minimum, every single one of them had to be a district that President Trump carried by ten points or more ․”); id. at 68–69 (“[T]he 10 points was a minimum result. He had to win it by a minimum of 10 percent. It didn't mean I couldn't draw a district at Trump plus 20 ․”); id. at 69 (“Q. ․ If you had the opportunity to draw a district that was more Republican than Trump plus 10 in ’24, did you try to take that opportunity? | A. Absolutely.”).

336.   See, e.g., id. at 68 (“[E]very one of those seats had to be carried by Ted Cruz in 2024. There was no set amount of range on how much he had to win it by, but he had to win each of those five seats.”).Where possible, Mr. Kincaid also configured those districts such that Governor Abbott carried the district by a comfortable margin in 2018 and 2022. See id. at 72 (“I also looked at Governor Abbott's performance in 2022 and 2018. We wanted to make sure that all of those districts, or at least most of them, were seats that he carried by as decent a margin as possible within the criteria in [20]22 and [20]18 because, obviously, the first test of this map would be in a midterm election versus a presidential election.”). Mr. Kincaid occasionally deviated from that criterion, however. See id. at 161 (“Q. Did you look at the Abbott 2022 numbers when you were drawing District 28? | A. I did. | Q. How, if at all, did that inform the way that you drew it? | A. Governor Abbott didn't carry those districts down there, but I was able to get them the Cruz and Trump numbers that did. So that's what I looked at.”).

337.   See id. at 73 (“[O]ne thing that I did is I went back and I did a durability test on all of these districts․ We have a national redistricting dataset that has disaggregated results down to the block level going back ․ decades. So what I was able to do is, with Texas, look at the 2012 Romney results. And so I looked at every presidential, [U.S.] senate, and governor's race in Texas ․ from 2012 through 2024. And the reason I did that is[,] obviously, Texas has been ․ politically ․ volatile for ․ several years now. It's been ․ wide Republican wins, narrow Republican wins, wide Republican wins again. And the coalition[ ] that Republicans have been winning elections with has changed significantly from 2012 to now. And so what I wanted to do is look at how those districts performed over the last three iterations of the Republican coalition.”).

338.   See, e.g., id. at 66–67 (“I wanted to improve the overall compactness of the map. That was another criteri[on]․ I just wanted to take [the districts in the 2021 Map] and make them cleaner, more compact, more city-based, more county-based, where I could than the previous one. That's more of a personal preference more than anything else. I like, when I can, to draw clean districts.”).See also, e.g., id. at 70–72 (discussing how Mr. Kincaid assesses compactness both visually and numerically when drawing maps); id. at 74–75 (exploring how Mr. Kincaid accounts for geographical boundaries when drawing maps).

339.   See, e.g., Abrams, 521 U.S. at 98, 117 S.Ct. 1925 (“[T]he constitutional guarantee of one person, one vote ․ requires congressional districts to achieve population equality as nearly as is practicable.” (citation modified)).

340.   See, e.g., Prelim. Inj. Hr'g Tr. Day 6 (Morning), ECF No. 1419, at 54 (“I have to balance the population of every district across the state ․ perfectly. Because we're not allowed to deviate from perfect population. So every district has to be about the same.”); id. at 75–76 (“Q. ․ [Y]ou mentioned earlier that drawing the maps with the appropriate equality in population was part of the process. Generally, is that something you did when you drew the Texas maps? | A. Yes. I equalized the populations when drawing the maps, yes.”).

341.   See, e.g., id. at 89–90 (“[MR. KINCAID:] ․ [A] nonnegotiable for Texas 5 was that I had to keep Kaufman, Van Zandt, and Henderson Counties whole. I could not split those. So they had to remain the core of Texas 5. | Q. Is that, again, the instruction from the Texas Republican congressional delegation? | A. Yes.”).

342.   See, e.g., id. at 95 (“[T]he city of Addison is slightly split there; and that was to make sure that the district office for Texas 24 stayed in Texas 24.”).

343.   See supra note 331 and accompanying text.

344.   See, e.g., Prelim. Inj. Hr'g Tr. Day 6 (Morning), ECF No. 1419, at 97–98 (“Q. What consideration, if any, did you give to keeping Democratic incumbents in the districts where they were under the 2021 map? | A. I didn't.”).

345.   See, e.g., id. at 129–30 (“Q. As the map drawer, did you consider core retention more closely when dealing with districts with a Republican incumbent or did that—did that partisan consideration not matter? | A. I was definitely trying to minimize the disruption in the Republican incumbent seats, yes. | Q. What about the Democratic incumbent seats? | A. No. I was trying—I had to rework most of the Democrat seats to create new pickup opportunities. So that wasn't a consideration.”).

346.   See id. at 76–191.

347.   See id. at 76–191.

348.   See Prelim. Inj. Hr'g Tr. Day 7 (Morning), ECF No. 1420, at 101.

349.   See Prelim. Inj. Hr'g Tr. Day 9 (Afternoon), ECF No. 1345, at 134 (observing that Mr. Kincaid testified “totally without notes”).

350.   See 581 U.S. at 313–15, 137 S.Ct. 1455.

351.   See id. at 295, 137 S.Ct. 1455.

352.   See id. at 313–14, 137 S.Ct. 1455.

353.   See id. at 313, 137 S.Ct. 1455.

354.   See id. at 315, 137 S.Ct. 1455 (“Whether the racial make-up of the county was displayed on his computer screen or just fixed in his head, the court thought, [the mapmaker]’s denial of race-based districting rang hollow.” (citation modified)).

355.   See id. at 313, 137 S.Ct. 1455.

356.   See id. at 314, 137 S.Ct. 1455.We recognize that part of the reason why the district court disbelieved the outside mapmaker's testimony in Cooper was because he gave “self-contradictory testimony” at his deposition and at trial. See id. at 314–15, 137 S.Ct. 1455. In our view, nothing that Mr. Kincaid said at the preliminary-injunction hearing was self-contradictory; it was instead inconsistent with the testimony of other witnesses and the enacted map's raw racial demographics. Nonetheless, Cooper remains illustrative.

357.   See id. at 316, 137 S.Ct. 1455.

358.   Contrast Brooks Prelim. Inj. Ex. 258, ECF No. 1326-5, at 1, with Brooks Prelim. Inj. Ex. 265, ECF No. 1326-12, at 1. See also supra Section II.G.We have purposefully omitted CD 22 from this list of “suspicious” districts. CD 22 went from being just 0.8% below 50% White to just 0.8% above. Contrast Brooks Prelim. Inj. Ex. 258, ECF No. 1326-5, at 1, with Brooks Prelim. Inj. Ex. 265, ECF No. 1326-12, at 1. That's the sort of negligible variation that could easily happen by chance.For that reason, we conclude that the Plaintiff Groups haven't shown a sufficient likelihood of success on the merits of their challenge to CD 22.

359.   See Prelim. Inj. Hr'g Tr. Day 2 (Afternoon), ECF No. 1338, at 123–24 (eliciting that one of the Plaintiff Groups’ expert cartographers once drew a 50.1% Black district without purposefully trying to do so).

360.   Cf. Cooper, 581 U.S. at 313–14, 137 S.Ct. 1455.

361.   Cf. id. at 313, 137 S.Ct. 1455.

362.   Contrast Brooks Prelim. Inj. Ex. 258, ECF No. 1326-5, at 1, with Brooks Prelim. Inj. Ex. 265, ECF No. 1326-12, at 1. See also supra Section II.D.

363.   See Prelim. Inj. Hr'g Tr. Day 6 (Afternoon), ECF No. 1342, at 51–52, 54–55.

364.   See supra Section II.E.

365.   Brooks Prelim. Inj. Ex. 309-T, ECF No. 1327-9, at 105–07.We agree with the State Defendants that the “nothing's a coincidence” comment is not direct evidence of racial intent. See also Defs.’ Post-Hr'g Br., ECF No. 1284, at 28–29. In context, Chairman Hunter's “nothing's a coincidence” comment was not an admission of racial motives, but rather a preface to a discussion of traditional districting criteria. See Brooks Prelim. Inj. Ex. 309-T, ECF No. 1327-9, at 107.

366.   See supra note 329 and accompanying text (establishing that Mr. Kincaid had the ability to display both racial and partisan data in his mapmaking software and base his districting decisions on race accordingly).

367.   See supra Section III.B.4.ii.

368.   See, e.g., Prelim. Inj. Hr'g Tr. Day 7 (Morning), ECF No. 1420, at 125–26 (discussing “the instructions from the White House” regarding how to draw the map).

369.   See, e.g., Prelim. Inj. Hr'g Tr. Day 6 (Morning), ECF No. 1419, at 89–90 (discussing a mapdrawing instruction Mr. Kincaid received “from the Texas Republican congressional delegation”).

370.   See Prelim. Inj. Hr'g Tr. Day 6 (Afternoon), ECF No. 1342, at 43 (“Q. ․ [W]hen you were drawing the map ․ there were no legislators present for that process? | [MR. KINCAID:] When I was drawing the map? No. | Q. ․ [T]he Governor wasn't there? | A. He was not looking over my shoulder, no.”); id. at 46 (“Q. ․ So no legislators present for the map drawing. You did not speak directly to any member of the House. You did not speak to anyone directly in the Senate other than Senator King. Is that right? | A. That's correct, as far as ․ during the map-drawing process.”).

371.   See supra Section III.B.4.d.iii.

372.   Cf. Brnovich, 594 U.S. at 689–90, 141 S.Ct. 2321 (emphasizing that “the legislators who vote to adopt a bill are not the agents of the bill's ․ proponents,” as “legislators have a duty to exercise their judgment” when deciding whether to vote for a particular piece of legislation).

373.   See generally 227 F.3d 504 (5th Cir. 2000).

374.   See id. at 510 (“Judge Turner, formerly a lawyer in and unsuccessful candidate for an at-large judgeship in the 23rd [Judicial District Court (“JDC”)] ․ drew the district lines ․ for the 23rd JDC, and the legislature adopted his proposed subdistricting scheme.”); id. (“Judge Turner was not a member of the state legislature.”).

375.   See id. (“Judge Turner averred that race did not predominate over traditional districting principles; he stated that, while following traditional districting principles, he drew the district lines to accommodate his candidacy.”).

376.   See id. at 508.

377.   Id. at 510; see also id. at 514 (“Although Judge Turner's affidavit provides some insight into the legislature's intent, it is far from determinative.”).

378.   See id. at 510 (emphasizing that “Judge Turner was not a member of the state legislature,” and that a factfinder could plausibly infer “that the legislature was ready to adopt whatever proposal would satisfy its objective of creating a black subdistrict”).

379.   See supra note 378.

380.   See 227 F.3d at 510 (“To end [litigation with voters over the state's system for electing judges], and to address the Justice Department's [objections to preclearance], the state agreed to implement a subdistrict election plan ․ that would contain at least one subdistrict with a majority black voter registration.” (citation modified)); id. at 511 (“[O]ne could readily infer that the state was motivated to pass [the challenged plan] by the desire to secure Section 5 preclearance, which, under DOJ's policy, meant creating racially-based subdistricts.”).

381.   See id. at 508.

382.   See id. at 510.

383.   See id. at 510 n.8 (noting that the non-racial “factors [the mapmaker] considered in redrawing the district lines” included “contiguity, non-splitting of precincts, the one-person/one-vote principle, protection of incumbents, the political preference of incumbents to include parts of each parish in each subdistrict, and the location of [the mapdrawer]’s own [political] supporters”).

384.   See id. at 511; see also id. (noting that “the state forthrightly declared that the reason for the change ․ was to reapportion” the challenged district to have “a majority black population”).

385.   See id.

386.   See id. at 514.

387.   Prejean arose in a summary judgment posture. See id. at 508. The Prejean court was therefore “required to view the evidence and all inferences therefrom in the light most favorable to” the plaintiffs challenging the map. Id. at 510. Here, by contrast, the Plaintiff Groups face a much heavier burden to show a sufficient likelihood that they'll ultimately succeed on the merits. See supra Section II.B.1. Our analysis accounts for that procedural distinction.We also recognize that the mapmaker's affidavit in Prejean constituted far weaker evidence than Mr. Kincaid's extensive and detailed testimony. See 227 F.3d at 514 (“There is no supporting documentation showing who [the mapdrawer's] supporters were, and where they would be found—or not found—in the proposed subdistrict. No evidence of his previous candidacies’ vote distribution was offered. Yet [the mapdrawer's] statement [that he drew the district lines to include his political supporters from his previous attempts at elective office] cries out for objective verification.”). We've thus been careful not to read more into Prejean than is supportable.

388.   See 227 F.3d at 510 (refusing to treat the mapmaker's “affidavit describing his intent in drawing the subdistricts ․ as conclusive proof of the legislature's intent”); id. at 514 (“Although Judge Turner's affidavit provides some insight into the legislature's intent, it is far from determinative.”).

389.   See id. (opining that the record permitted a “plausible inference ․ that the legislature was ready to adopt whatever proposal would satisfy its objective of creating a black subdistrict”).

390.   Cf. id. at 511; see also Section III.B.3.

391.   Cf. 227 F.3d at 514.

392.   Cf. id.

393.   Common Cause Fla., 726 F. Supp. 3d at 1364–65.Cf. Reno v. Bossier Par. Sch. Bd., 520 U.S. 471, 487, 117 S.Ct. 1491, 137 L.Ed.2d 730 (1997) (“[T]he impact of an official action is often probative of why the action was taken in the first place since people usually intend the natural consequences of their actions.”).

394.   Cf. Prejean, 227 F.3d at 511.

395.   See Cooper, 581 U.S. at 313, 137 S.Ct. 1455.

396.   See id. at 300, 137 S.Ct. 1455 (citation modified); see also supra Sections III.B.3.a & III.B.4.d.ii.

397.   See Brooks Prelim. Inj. Ex. 258, ECF No. 1326-5, at 2 (indicating that CD 37 was 60.7% White by CVAP under the 2021 Map).

398.   See Prelim. Inj. Hr'g Tr. Day 3 (Morning), ECF No. 1416, at 39 (“In [the 2021] version of CD 37, White voters voted for Democratic candidates. On average they voted 80 percent for Democrats.”); see also, e.g., Tex. NAACP Prelim. Inj. Ex. 208, ECF No. 1384-8, at 9.

399.   See, e.g., Prelim. Inj. Hr'g Tr. Day 3 (Morning), ECF No. 1416, at 40 (“Q. So did the legislature change the nature of CD 37 as a majority White Democratic voting district? | A. No.”); id. (“In new CD 37 the Whites ․ prefer Democratic candidates.”); see also, e.g., Tex. NAACP Prelim. Inj. Ex. 208, ECF No. 1384-8, at 9.

400.   See, e.g., Brooks Prelim. Inj. Ex. 265, ECF No. 1326-12, at 2 (indicating that CD 37 is 54.0% White by CVAP under the 2025 Map); Prelim. Inj. Hr'g Tr. Day 3 (Morning), ECF No. 1416, at 39 (“New CD 37 remains a White majority district.”).

401.   See Brooks Prelim. Inj. Ex. 267, ECF No. 1326-14, at 6.

402.   See id. at 2.

403.   See, e.g., Tenn. State Conf., 746 F. Supp. 3d at 494 (opining that if a map “treat[s] minority voters of one party worse than white voters of the same party,” “that could undercut the possibility that partisan politics were to blame for the decision” (citation modified)).While Mr. Kincaid provided putatively partisan and practical rationales for drafting CD 37 the way he did, see Prelim. Inj. Hr'g Tr. Day 6 (Morning), ECF No. 1419, at 146–48, we discredit that testimony for the reasons given above. See supra Section III.B.4.d.ii.The Legislature also left CD 7 in the Houston area largely untouched. See Brooks Prelim. Inj. Ex. 267, ECF No. 1326-14, at 1–2 (indicating that 74.6% of the voters in the old CD 7 remain in the new CD 7). Though CD 7 was not a majority-White district, see Brooks Prelim. Inj. Ex. 258, ECF No. 1326-5, at 1, it generally elected candidates preferred by White Democrats under the 2021 Map, and it will likely continue to do so under the 2025 Map. See, e.g., Tex. NAACP Prelim. Inj. Ex. 208, ECF No. 1384-8, at 9. That shows the Legislature radically transformed districts that elected Democratic candidates preferred by voters of color while leaving districts that elected Democrats preferred by White voters mostly unchanged. See, e.g., Tex. NAACP Prelim. Inj. Ex. 208, ECF No. 1384-8, at 9 (indicating that the Legislature changed the political performance of CD 9 but not CD 7). That reinforces that racial concerns predominated over partisanship. See, e.g., Tenn. State Conf., 746 F. Supp. 3d at 494.We likewise discredit Mr. Kincaid's proffered race-neutral rationales for CD 7's configuration. Contra Prelim. Inj. Hr'g Tr. Day 6 (Morning), ECF No. 1419, at 140–44. See also supra Section III.B.4.d.ii.

404.   Compare Brooks Prelim. Inj. Ex. 258, ECF No. 1326-5, at 1, with Brooks Prelim. Inj. Ex. 265, ECF No. 1326-12, at 1; see also supra Section II.G.5.See also Prelim. Inj. Hr'g Tr. Day 3 (Morning), ECF No. 1416, at 41 (testimony affirming that CD 27 remains a district that “Republican candidates will consistently win”).Here too, Mr. Kincaid provided partisan and race-neutral rationales for CD 27's boundaries. See Prelim. Inj. Hr'g Tr. Day 6 (Morning), ECF No. 1419, at 146–51. We discredit that testimony too. See supra Section III.B.4.d.ii.

405.   See, e.g., Prelim. Inj. Hr'g Tr. Day 5 (Morning), ECF No. 1418, at 56–60; Tex. NAACP Prelim. Inj. Ex. 208, ECF No. 1384-8, at 14 & n.7.

406.   See, e.g., Tex. NAACP Prelim. Inj. Ex. 208, ECF No. 1384-8, at 23.

407.   See, e.g., id. at 22–23 (“Partisanship favoring Republican candidates in general [elections] is accounted for with a score based on the number of Republican district wins across a set of 29 general elections ․”); id. at 23 (“Partisanship specific to the performance of Donald Trump is accounted for in two ways: counting the number of Trump district wins in three elections (2016, 2020, 2024) and by simply considering the most recent election ․”).

408.   See, e.g., id. at 22; Prelim. Inj. Hr'g Tr. Day 5 (Morning), ECF No. 1418, at 58 (“[T]he basic method creates plans that take into account population balance [and] ensure contiguity and that prioritize compactness ․”).

409.   See Prelim. Inj. Hr'g Tr. Day 5 (Morning), ECF No. 1418, at 62–63 (“Q. So these parameters, do they generate a large number of maps? | [DR. DUCHIN:] Under these parameters I then generate a very large number of maps, correct. | Q. And do you winnow them down? | A. Right․ The second stage is to filter it. So by winnowing, ․ I mean I'll take all those maps and I'll filter them down by whether they meet some checklist of other conditions.”).See also Tex. NAACP Prelim. Inj. Ex. 208, ECF No. 1384-8, at 23 (winnowing down to only include maps in which “Republicans overall have at least as many wins” as they do in the enacted map); id. (further winnowing down to only include maps in which “at least as many districts have a plurality win for Donald Trump from the 2024 election as in” the enacted map); id. (further winnowing down to only include maps in which “the double-bunking of incumbents ․ is no greater than in” the enacted map).See also Prelim. Inj. Hr'g Tr. Day 5 (Morning), ECF No. 1418, at 64 (explaining that the program winnows down the universe of randomly generated maps to only include maps in which “the number [of districts] won by Republicans” is “at least as high as in” the Enacted Map); id. (explaining that “the winnowing, the filter, ensures that [the surviving maps] are getting at least as strong Republican performance as the [enacted] plan”).

410.   See, e.g., Tex. NAACP Prelim. Inj. Ex. 208, ECF No. 1384-8, at 23.

411.   See, e.g., id. at 22–23.

412.   See supra notes 155–158 and accompanying text.

413.   See, e.g., Prelim. Inj. Hr'g Tr. Day 5 (Morning), ECF No. 1418, at 68 (explaining that Dr. Duchin's method permits her “to compare the racial attributes of the [enacted] map to a baseline that's been constructed according to [the] parameters” discussed above).

414.   See, e.g., id. at 57 (“The point of this is just to show you what plans look like when created by known rules. So it lets you assess whether a proposed plan behaves as though it was created by the stated rules.”).

415.   See, e.g., id. at 66.

416.   See id. at 72 (explaining that if “race-blind comparators ․ don't reproduce [the] racial composition” of the enacted map, that would suggest “that race was used in making” the map).

417.   See id. at 68.

418.   See, e.g., Alexander, 602 U.S. at 6, 144 S.Ct. 1221.

419.   See Prelim. Inj. Hr'g Tr. Day 5 (Morning), ECF No. 1418, at 68.

420.   See Tex. NAACP Prelim. Inj. Ex. 208, ECF No. 1384-8, at 2, 14.The boxplots don't correspond with specific district numbers. See Prelim. Inj. Hr'g Tr. Day 5 (Morning), ECF No. 1418, at 69 (“[T]hey are arranged not in order of the district numbers but from lowest to highest POC CVAP.”).

421.   See id. at 14; see also Prelim. Inj. Hr'g Tr. Day 5 (Morning), ECF No. 1418, at 68 (“I'm showing you what is abbreviated POC CVAP, which means the minority citizen voting age percentage in each of the districts.”).

422.   See Tex. NAACP Prelim. Inj. Ex. 208, ECF No. 1384-8, at 14.

423.   The black figures represent “a 40,000 plan subsample” without “filtering conditions” like “rural composition and various kinds of tests that the partisanship matches or exceeds that in the State's plan.” Prelim. Inj. Hr'g Tr. Day 5 (Morning), ECF No. 1418, at 68. We're more interested in the orange figures, which “only include plans that meet the full checklist of districting principles.” See Tex. NAACP Prelim. Inj. Ex. 208, ECF No. 1384-8, at 14.

424.   See Prelim. Inj. Hr'g Tr. Day 5 (Morning), ECF No. 1418, at 69.

425.   See id.

426.   See Tex. NAACP Prelim. Inj. Ex. 208, ECF No. 1384-8, at 14.

427.   See Prelim. Inj. Hr'g Tr. Day 5 (Morning), ECF No. 1418, at 69.

428.   See Tex. NAACP Prelim. Inj. Ex. 208, ECF No. 1384-8, at 14.

429.   See Prelim. Inj. Hr'g Tr. Day 5 (Morning), ECF No. 1418, at 69.

430.   See id. at 68 (explaining that “the blue dots” represent the “districts drawn by the State”).

431.   See id. at 70 (testifying that “if the dot is outside the whiskers altogether,” that means “that no plan [that Dr. Duchin] generated in the sample ever had as low [or high] of a minority CVAP”).

432.   See id. at 72 (“[T]hat is suggestive that race was used in making these plans because these race-blind comparators, even made with layer upon layer of different assumptions about partisanship and other principles, don't reproduce that kind of racial composition.”).

433.   See Tex. NAACP Prelim. Inj. Ex. 208, ECF No. 1384-8, at 15.

434.   See supra Section II.D.

435.   See Tex. NAACP Prelim. Inj. Ex. 208, ECF No. 1384-8, at 2, 15.

436.   See id. at 15.See also Prelim. Inj. Hr'g Tr. Day 5 (Morning), ECF No. 1418, at 73 (“The second [column] from the [right] is off the charts in the direction of packing. Where you would expect POC CVAP in the 60 to 70[%] range; instead, it's over 80 percent.”).

437.   See Tex. NAACP Prelim. Inj. Ex. 208, ECF No. 1384-8, at 14 (concluding that “the racial composition of the districts is highly atypical of random plans whose partisan performance is at least as favorable to Republicans generally and to Donald Trump in particular”).See also Prelim. Inj. Hr'g Tr. Day 5 (Morning), ECF No. 1418, at 73 (“Q. So does this mean that the racial composition of the district was something you did not see in any of your maps? | [DR. DUCHIN:] Right. In several of these instances, it's past anything ever observed.”).

438.   See Tex. NAACP Prelim. Inj. Ex. 208, ECF No. 1384-8, at 2, 14.

439.   See id. at 2, 15.

440.   See id. at 14 (“[T]wo of the eight districts [in the Dallas/Fort Worth area]—both where we would expect districts near the 50% mark—show that the POC CVAP is outlyingly low. In the next district, it is outlyingly high.”).See also Prelim. Inj. Hr'g Tr. Day 5 (Morning), ECF No. 1418, at 70 (“A. ․ There are two districts where the minority citizen voting age population is really anomalously low. You can see that ․ in the fourth and the third column from the end ․ In one case, the blue dot is at the whisker, which means it's at the 1st percentile. In the other case, it's below the whisker, suggesting that it is lower than whatever is observed in this large generation process to make plans under the assumptions reported earlier. | Q. What does it mean if ․ the dot is in the 1st percentile? | A. That means that ․ only 1 percent of the plans have a lower minority CVAP.”).

441.   See Tex. NAACP Prelim. Inj. Ex. 208, ECF No. 1384-8, at 15 (“The signs of packing and cracking are less severe in the [Travis/Bexar County area], but the characteristic pattern is still present: one district near an expected 50% POC CVAP status has markedly diminished minority citizen share, while the next district is elevated to over 60%.”).See also Prelim. Inj. Hr'g Tr. Day 5 (Morning), ECF No. 1418, at 73–74 (“[W]hile directionally the same, [the Travis/Bexar County area] doesn't show as extreme or as strong of a pattern. However, you can see that in one district there is what looks like about a 5th percentile level of cracking. And in that top district there is what looks to be about a 5th percentile showing of packing. So you see directionally the same pattern, never the reverse. But the evidence here isn't as strong as in the previous two clusters.”).

442.   See, e.g., Prelim. Inj. Hr'g Tr. Day 5 (Morning), ECF No. 1418, at 66 (“[T]he State's plan is an outlier in its racial composition.”).

443.   See, e.g., id. at 30 (concluding that “there is strong evidence that race was used in the creation of” the 2025 Map, and that the 2025 Map is not “consistent with ․ the race neutral pursuit of pure partisan aims”); id. at 72 (“[The results are] suggestive that race was used in making [the Enacted Map] because these race-blind comparators, even made with layer upon layer of different assumptions about partisanship and other principles, don't reproduce that kind of racial composition.”).

444.   See Prelim. Inj. Hr'g Tr. Day 9 (Afternoon), ECF No. 1345, at 46–47, 164; Prelim. Inj. Hr'g Tr. Day 5 (Morning), ECF No. 1418, at 8.

445.   See 602 U.S. at 33, 144 S.Ct. 1221.See also Defs.’ Post-Hr'g Br., ECF No. 1284, at 12 (“[The Plaintiff Groups’] case depends on the very methods the Supreme Court rejected in Alexander ․ and even some of the same experts. Alexander contains a section labeled ‘Dr. Moon Duchin’ that finds a district court clearly erred in relying on her opinions. Yet here, Plaintiffs come to this Court with Dr. Moon Duchin and ask it to discredit [Mr. Kincaid's testimony] based on her work.” (emphases omitted) (citation modified)).

446.   See 602 U.S. at 33, 144 S.Ct. 1221.

447.   See Tex. NAACP Prelim. Inj. Ex. 208, ECF No. 1384-8, at 22 (“Core retention with respect to the State's new plan is implemented with a surcharge of 0.2 on edges that span across two of the State's new enacted congressional districts.”); id. at 22–23 (“Partisanship favoring Republican candidates in general is accounted for with a score based on the number of Republican district wins across a set of 29 general elections ․”); id. at 23 (“Partisanship specific to the performance of Donald Trump is accounted for in two ways: counting the number of Trump district wins in three elections (2016, 2020, 2024) and by simply considering the most recent election ․”); id. (listing winnowing conditions that explicitly take partisanship into account).

448.   See 602 U.S. at 33, 144 S.Ct. 1221.See also, e.g., Ala. Legis. Black Caucus v. Alabama, 575 U.S. 254, 262, 135 S.Ct. 1257, 191 L.Ed.2d 314 (2015) (“A racial gerrymandering claim ․ applies to the boundaries of individual districts. It applies district-by-district. It does not apply to a State considered as an undifferentiated ‘whole.’ ”).

449.   See Tex. NAACP Prelim. Inj. Ex. 208, ECF No. 1384-8, at 2, 14–15.See also Prelim. Inj. Hr'g Tr. Day 5 (Morning), ECF No. 1418, at 32 (explaining that focusing on these geographic clusters “make[s] [the analysis] local,” while still “acknowledg[ing] that the drawing of lines in one district has an impact on neighboring districts”).

450.   See, e.g., Prelim. Inj. Hr'g Tr. Day 5 (Morning), ECF No. 1418, at 118–20 (“Q. When you are putting the parameters in your [computer program] to draw maps, you are putting those in there because you want for the maps the [program] draws to match your understanding of the stated intent of the map, right? | A. I am testing versions of that. That's right․ | Q. So the similarities between the maps you draw and the enacted map matter for the precision of your analysis? | A. The similarities between my parameters and the stated intent are important. I agree with that.”).

451.   See Defs.’ Post-Hr'g Br., ECF No. 1284, at 51–52.

452.   See Tex. NAACP Prelim. Inj. Ex. 208, ECF No. 1384-8, at 23.See also Prelim. Inj. Hr'g Tr. Day 5 (Morning), ECF No. 1418, at 65 (“[DR. DUCHIN:] [O]ne set of runs were done under just simple Trump wins. Did Trump have more votes?”).

453.   See Tex. NAACP Prelim. Inj. Ex. 208, ECF No. 1384-8, at 23.See also Prelim. Inj. Hr'g Tr. Day 5 (Morning), ECF No. 1418, at 65 (“[DR. DUCHIN:] But later, as a check, I also sought out plans in which Trump's percentage was at least 55 percent, to make sure that that 50 percent line wasn't guiding the findings.”); id. at 67 (“[I]t's my understanding that when trying to execute partisan gerrymandering, you don't just want to win narrowly. You would like it to be durable and withstand some swing in partisan performance. So 55[%] is a threshold that tells you that even if the vote were to swing by 5 percent you would still win.”).

454.   See Tex. NAACP Prelim. Inj. Ex. 208, ECF No. 1384-8, at 23.See also Prelim. Inj. Hr'g Tr. Day 5 (Morning), ECF No. 1418, at 66 (“[DR. DUCHIN:] [S]ometimes layering in additional principles can change the observed range. But it never changes the finding that the State's plan is an outlier in its racial composition. And that includes the Trump 55 plus.”).

455.   See Defs.’ Post-Hr'g Br., ECF No. 1284, at 52 (“Applying a 55% or 50%-plus-one threshold is too low to fairly model the political performance of the 2025 Plan ․”).

456.   See id.

457.   See LULAC Prelim. Inj. Ex. 1202, ECF No. 1402-6, at 5 (CD 9 = 59.5%); id. at 13 (CD 22 = 59.9%); id. at 16 (CD 27 = 60.0%).

458.   See id. at 19 (CD 32 = 57.7%).

459.   See id. at 20 (CD 34 = 54.6%); id. at 21 (CD 35 = 54.6%).

460.   See Prelim. Inj. Hr'g Tr. Day 5 (Morning), ECF No. 1418, at 141–42 (“Q. And you executed a run seeking to match the number of districts with Trump's 2024 major party vote share over 55 percent, right? | [DR. DUCHIN:] Right. | Q. Does that mean that 55 percent was a minimum? | A. That's what that is. | Q. And so the districts that achieved more than 55 percent would be accounted for in that run? | A. That's right. That would include districts that achieve 60 percent or more.”).

461.   See supra note 444 and accompanying text.

462.   See Defs.’ Post-Hr'g Br., ECF No. 1284, at 53 (“While Dr. Duchin attempted to model core retention by having her [program] surcharge simulated districts with a lower core retention, it did not differentiate between core retention of Republican-held districts versus Democratic-held districts ․” (citation modified)).See also id. at 54 (“[W]hile Dr. Duchin required the algorithm to draw simulated plans that did not pair more incumbents than [the enacted map], she failed to consider whether the simulated plans paired Republican or Democrat incumbents with each other. But incumbents are not fungible—and given the Legislature's partisan goal of flipping five Democrat-held seats to Republican-held seats, it is not reasonable to assume that a plan that paired two sets of Republican incumbents would be equally preferred to a plan that paired two sets of Democrats. Nor is Dr. Duchin's assumption consistent with [Mr. Kincaid's] testimony in this case that only Republican incumbents were not paired together in the mapmaking process.” (citations omitted)).See also supra notes 343–345 and accompanying text.

463.   See Defs.’ Post-Hr'g Br., ECF No. 1284, at 54 (“Dr. Duchin performed her incumbency analysis using an out-of-date list of incumbent addresses ․ Dr. Duchin did not dispute that ten of the incumbents on the list she used were not in Congress in 2024–2025. Former members of Congress are not incumbents the Legislature would want to protect in 2025; therefore Dr. Duchin's use of outdated incumbent addresses severely impacts her analysis.”).

464.   See Prelim. Inj. Hr'g Tr. Day 5 (Morning), ECF No. 1418, at 108–09 (“[DR. DUCHIN:] I have been aware for some time that these incumbent addresses are out of date and have been requesting updated incumbent addresses for months.”).

465.   See, e.g., id. at 73 (“Q. So does this mean that the racial composition of the district was something you did not see in any of your maps? | [DR. DUCHIN:] Right. In several of these instances, it's past anything ever observed.”).

466.   See, e.g., id. at 72 (“[The results are] suggestive that race was used in making [the Enacted Map] because these race-blind comparators, even made with layer upon layer of different assumptions about partisanship and other principles, don't reproduce that kind of racial composition.”).

467.   Compare Brooks Prelim. Inj. Ex. 258, ECF No. 1326-5, at 1 (reflecting that, under the 2021 Map, CD 33 was 43.6% Hispanic, 25.2% Black, 5.7% Asian, and 23.4% White), with Brooks Prelim. Inj. Ex. 265, ECF No. 1326-12, at 1 (reflecting that, under the 2025 Map, CD 33 is 38.2% Hispanic, 19.6% Black, 4.4% Asian, and 35.5% White).See also supra Section II.B.4.

468.   See, e.g., Bethune-Hill, 580 U.S. at 191, 137 S.Ct. 788 (citation modified).

469.   E.g., Alexander, 602 U.S. at 7, 144 S.Ct. 1221 (citation modified).

470.   E.g., id. at 8, 144 S.Ct. 1221.

471.   E.g., id. (citation modified).

472.   See, e.g., Prelim. Inj. Hr'g Tr. Day 5 (Morning), ECF No. 1418, at 78–80.

473.   See, e.g., id. at 81.

474.   See, e.g., Alexander, 602 U.S. at 8, 144 S.Ct. 1221.

475.   E.g., Bethune-Hill, 580 U.S. at 190, 137 S.Ct. 788.See also, e.g., id. (“Of course, a conflict or inconsistency [with traditional districting principles] may be persuasive circumstantial evidence tending to show racial predomination, but there is no rule requiring challengers to present this kind of evidence in every case.”).

476.   E.g., id. at 189, 137 S.Ct. 788 (citation modified).

477.   E.g., id. at 191, 137 S.Ct. 788.

478.   See supra note 272 and accompanying text.

479.   See supra Sections III.B.3 & 5.

480.   Cooper, 581 U.S. at 317, 137 S.Ct. 1455.

481.   Id.

482.   Id.

483.   Id.

484.   602 U.S. at 10, 144 S.Ct. 1221; see also supra notes 189–190 and accompanying text.

485.   602 U.S. at 35, 144 S.Ct. 1221 (citation modified).

486.   Id.

487.   Id.

488.   The map that counsel produced while fiddling with map-drawing software in front of the State Defendants’ expert for several hours doesn't qualify as a proper Alexander map. See Prelim. Inj. Hr'g Tr. Day 9 (Morning), ECF No. 1422, at 82–141; see also Prelim. Inj. Hr'g Tr. Day 9 (Afternoon), ECF No. 1345, at 29 (“[H]e's trying to draw an Alexander district through me.”) If the Plaintiff Groups intended that to be their Alexander map, they should have presented it through expert testimony during their case-in-chief. See Prelim. Inj. Hr'g Tr. Day 9 (Afternoon), ECF No. 1345, at 50 (“Q. To your knowledge, did Plaintiffs offer an expert to draw an alternative map, an Alexander map, as you discussed on cross-examination? | A. I have a feeling I am their Alexander witness.”).Nor do any of Dr. Duchin's randomly generated maps qualify as an Alexander map for our purposes, since none of those maps were introduced into evidence (as opposed to a pictorial representation of their racial demographics). See Tex. NAACP Prelim. Inj. Ex. 208, ECF No. 1384-8, at 14–15; see also supra Section III.5.e.

489.   See 602 U.S. at 35, 144 S.Ct. 1221.

490.   See id. at 18, 144 S.Ct. 1221.

491.   See supra Section III.3.

492.   Contra Alexander, 602 U.S. at 9, 144 S.Ct. 1221.

493.   See id. at 35, 144 S.Ct. 1221 (“Nor is an alternative map difficult to produce. Any expert armed with a computer can easily churn out redistricting maps that control for any number of specified criteria, including prior voting patterns and political party registration.”).

494.   See id. at 34, 144 S.Ct. 1221 (citation modified).See also id. at 35, 144 S.Ct. 1221 (“A plaintiff's failure to submit an alternative map—precisely because it can be designed with ease—should be interpreted by district courts as an implicit concession that the plaintiff cannot draw a map that undermines the legislature's defense that the districting lines were based on a permissible, rather than a prohibited, ground.” (citation modified)); id. (“The Challengers enlisted four experts who could have made these maps at little marginal cost.” (emphasis omitted)).

495.   See id. at 15, 144 S.Ct. 1221.

496.   See id. at 13, 144 S.Ct. 1221.

497.   Cf. Tenn. State Conf., 746 F. Supp. 3d at 482, 497 (“Alexander arose after a trial. This case, by contrast, remains at the pleadings stage․ We agree that the Challengers do not have to satisfy any alternative-map obligation at this stage.”).

498.   Cf., e.g., Prelim. Inj. Hr'g Tr. Day 3 (Morning), ECF No. 1416, at 81, 116–19 (another expert's testimony that, due to the “limited time” he had to prepare his analysis, he had to restrict his focus to six prior elections); Prelim. Inj. Hr'g Tr. Day 7 (Afternoon), ECF No. 1343, at 139–40 (“Q. Now, you said in your report that you did not have enough time to run ecological inference analysis yourself, right? | [DR. JEFFREY LEWIS:] That's right․ [F]rom the time that ․ I was asked to provide opinions on the matters that I described, I think I had more on the order of ten days.”).

499.   See supra note 488.

500.   See supra Section III.5.e.

501.   E.g., Alexander, 602 U.S. at 11, 144 S.Ct. 1221; see also supra note 191 and accompanying text.

502.   E.g., Cooper, 581 U.S. at 292, 137 S.Ct. 1455; see also supra note 192 and accompanying text.

503.   See, e.g., Defs.’ Post-Hr'g Br., ECF No. 1284, at 17 (insisting that the Plaintiff Groups “cannot” “demonstrate [any] use of race in the development of the map”); id. at 23 (“Race was not used here.”).

504.   See generally Defs.’ Resp. Intervenors’ & Tex. NAACP's Prelim. Inj. Mot., ECF No. 1195; Defs.’ Resp. Gonzales Pls.’ Prelim. Inj. Mot., ECF No. 1199; Defs.’ Resp. J. Prelim. Inj. Mot., ECF No. 1200; Defs.’ Post-Hr'g Br., ECF No. 1284.

505.   See Brooks Prelim. Inj. Ex. 253, ECF No. 1326, at 2.

506.   Miller, 515 U.S. at 920, 115 S.Ct. 2475 (citation modified).

507.   Id. at 922, 115 S.Ct. 2475.

508.   Id.

509.   See supra Section II.B; see also supra Section II.F.

510.   See supra Section II.B.

511.   E.g., Cooper, 581 U.S. at 301, 137 S.Ct. 1455.

512.   See Brooks Prelim. Inj. Ex. 253, ECF No. 1326, at 2 (“It is well established that so-called ‘coalition districts’ run afoul the [sic] Voting Rights Act ․”).

513.   See Cooper, 581 U.S. at 306, 137 S.Ct. 1455; see also Wis. Legis. v. Wis. Elections Comm'n, 595 U.S. 398, 404, 142 S.Ct. 1245, 212 L.Ed.2d 251 (2022) (explaining that “State have breathing room to make reasonable mistakes” regarding whether the VRA requires the State to enact a particular compliance measure).

514.   See Cooper, 581 U.S. at 306, 137 S.Ct. 1455.

515.   See supra Section II.D.

516.   See Miller, 515 U.S. at 921, 115 S.Ct. 2475.

517.   See id. at 922, 115 S.Ct. 2475.

518.   Winter, 555 U.S. at 20, 129 S.Ct. 365.

519.   SO Apartments, L.L.C. v. City of San Antonio, 109 F.4th 343, 353 (5th Cir. 2024) (quoting Janvey, 647 F.3d at 600 (quotation marks omitted)).

520.   TX NAACP's Mot. Prelim. Inj., ECF No. 1142, at 22–23; Congr. Intervenors’ Mot. Prelim. Inj., ECF No. 1143, at 14–15; Gonzales Pls.’ Mot. Prelim. Inj., ECF No. 1149, at 24–25; Brooks, LULAC, and MALC Pls.’ Joint Mot. Prelim. Inj., ECF No. 1150, at 44–45. See also U.S. Const. amend. XIV § 1; id. amend. XV § 1.

521.   BST Holdings, L.L.C. v. OSHA, 17 F.4th 604, 618 (5th Cir. 2021) (citation modified) (quoting Elrod v. Burns, 427 U.S. 347, 373, 96 S.Ct. 2673, 49 L.Ed.2d 547 (1976)). See also DeLeon v. Perry, 975 F. Supp. 2d 632, 663 (W.D. Tex. 2014) (“Federal courts at all levels have recognized that violation of constitutional rights constitutes irreparable harm as a matter of law.”), aff'd sub nom. DeLeon v. Abbott, 791 F.3d 619 (5th Cir. 2015); Obama for Am. v. Husted, 697 F.3d 423, 436 (6th Cir. 2012) (“When constitutional rights are threatened or impaired, irreparable injury is presumed ․ A restriction on the fundamental right to vote therefore constitutes irreparable injury.”).

522.   1st Prelim. Inj. Op., 601 F. Supp. 3d at 182; see also Deerfield Med. Ctr. v. City of Deerfield Beach, 661 F.2d 328, 338 (5th Cir. 1981); Obama for Am., 697 F.3d at 436.

523.   See generally Defs.’ Resp. to Texas NAACP and Congr. Intervenors’ Mtn. for Prelim. Inj., ECF No. 1195; Defs.’ Resp. to Gonzales Pls.’ Mtn. for Prelim. Inj., ECF No. 1199; Defs.’ Resp. to Brooks, LULAC, and MALC Pls.’ Joint Mtn. for Prelim. Inj., ECF No. 1200. See also Defs.’ Post-Hr'g Br., ECF No. 1284, at 88.

524.   See the sources cited supra note 523.

525.   TitleMax, 142 F.4th at 328; see also Texas v. United States, 809 F.3d 134, 150 (5th Cir. 2015), aff'd by an equally divided court, 579 U.S. 547, 136 S.Ct. 2271, 195 L.Ed.2d 638 (2016) (per curiam).

526.   1st Prelim. Inj. Op., 601 F. Supp. 3d at 183 (quotation marks omitted) (quoting Def. Distributed v. U.S. Dep't of State, 838 F.3d 451, 459 (5th Cir. 2016)).

527.   Id. (citation modified) (quoting Winter, 555 U.S. at 24, 129 S.Ct. 365).

528.   Texas v. United States, 524 F. Supp. 3d 598, 663 (S.D. Tex. 2021) (citing Texas v. United States, 809 F.3d at 187).

529.   Nken v. Holder, 556 U.S. 418, 435, 129 S.Ct. 1749, 173 L.Ed.2d 550 (2009).

530.   1st Prelim. Inj. Op., 601 F. Supp. 3d at 183 (citation modified) (quoting Veasey v. Abbott, 870 F.3d 387, 391 (5th Cir. 2017) (per curiam)).

531.   State Defs.’ Resp. to Gonzales Pls.’ Mtn. for Prelim. Inj., ECF No. 1196-1, at 36–39; State Defs.’ Post-Hr'g Br., ECF No. 1284, at 91.

532.   Merrill v. Milligan, ––– U.S. ––––, 142 S. Ct. 879, 879, ––– L.Ed.2d –––– (2022) (mem.) (Kavanaugh, J., concurring).

533.   Petteway v. Galveston Cnty., 87 F.4th 721, 723 (5th Cir. 2023) (per curiam) (Oldham, J., concurring) [hereinafter Petteway Purcell Op.].

534.   League of Women Voters of Fla., Inc. v. Fla. Sec'y of State, 32 F.4th 1363, 1371 (11th Cir. 2022) (quoting Republican Nat'l Comm. v. Democratic Nat'l Comm., 589 U.S. 423, 424, 140 S.Ct. 1205, 206 L.Ed.2d 452 (2020) (per curiam)).

535.   Kim v. Hanlon, 99 F.4th 140, 160 (3d Cir. 2024). See Milligan, 142 S. Ct. at 880 (Kavanaugh, J., concurring) (collecting cases); Petteway Purcell Op., 87 F.4th at 723 (Oldham, J., concurring) (collecting cases); McClure v. Jefferson Cnty. Comm'n, Nos. 25-13253, 25-13254, 2025 WL 2977740, at *2 (11th Cir. Oct. 16, 2025) (per curiam) (collecting cases).

536.   See Jacksonville Branch of NAACP v. City of Jacksonville, No. 22-13544, 2022 WL 16754389, at *2–3 (11th Cir. Nov. 7, 2022) (per curiam).

537.   State Defs.’ Resp. to Gonzales Pls.’ Mtn. for Prelim. Inj., ECF No. 1196-1, at 37–38; State Defs.’ Post-Hr'g Br., ECF No. 1284, at 89–91.

538.   For an exhaustive discussion of this development, see Callais v. Landry, 732 F. Supp. 3d 574, 585–87 (W.D. La. 2024).

539.   See generally id.

540.   Robinson v. Callais, ––– U.S. ––––, ––––, 144 S. Ct. 1171, 1171, 219 L.Ed.2d 1224 (2024) (mem.).

541.   See id.

542.   ––– U.S. ––––, 142 S. Ct. 879, ––– L.Ed.2d –––– (2022).

543.   See generally Singleton v. Merrill, 582 F. Supp. 3d 924 (N.D. Ala. 2022).

544.   Singleton v. Merrill, No. 2:21-cv-1291, 2022 WL 272636, at *11 (N.D. Ala. Jan. 27, 2022).

545.   See Milligan, 142 S. Ct. at 879.

546.   See id.

547.   See id.

548.   See id. at 879–82 (Kavanaugh, J., concurring) (basing his vote on Purcell).

549.   Petteway Purcell Op., 87 F.4th at 723 (Oldham, J., concurring).

550.   See id. at 724 (Oldham, J., concurring).

551.   State Defs.’ Resp. to Gonzales Pls.’ Mtn. for Prelim. Inj., ECF No. 1196-1, at 37–38; State Defs.’ Post-Hr'g Br., ECF No. 1284, at 89–91.

552.   State Defs.’ Resp. to Gonzales Pls.’ Mtn. for Prelim. Inj., ECF No. 1196-1, at 37–38; State Defs.’ Post-Hr'g Br., ECF No. 1284, at 89–91.

553.   See Robinson, 144 S. Ct. at 1171; see also H.B. 4, 89th Leg., 2d Spec. Sess. (Tex. 2025) (signed on August 29, 2025).

554.   See Milligan, 142 S. Ct. at 879 (Kavanaugh, J., concurring) (“The District Court declined to stay the injunction for the 2022 elections even though the primary elections begin (via absentee voting) just seven weeks from now ․”). Primary absentee voting begins January 17, 2026, in the 2026 Texas congressional election. Seven weeks before then is November 29, 2025.

555.   Robinson v. Ardoin, 37 F.4th 208, 228–29 (5th Cir. 2022) (per curiam) (collecting cases).

556.   Get Loud Ark. v. Thurston, 748 F. Supp. 3d 630, 665 (W.D. Ark. 2024).

557.   Milligan, 142 S. Ct. at 881 (Kavanaugh, J., concurring).

558.   Id.

559.   Id.

560.   See Tenn. Conf. of the NAACP v. Lee, 105 F.4th 888, 897 (6th Cir. 2024) (“As others have recognized, the Supreme Court has not adopted any categorical answer to the question of ‘how close is too close?’ The answer might depend on injunction-specific factors about the nature of the required changes and the burdens they will impose.” (citation modified)). See also Milligan, 142 S. Ct. at 881 n.1 (Kavanaugh, J., concurring) (“How close to an election is too close may depend in part on the nature of the election law at issue, and how easily the State could make the change without undue collateral effects.”); Jacksonville Branch of NAACP, 2022 WL 16754389, at *2–3.

561.   McClure, 2025 WL 2977740, at *2; cf. Milligan, 142 S. Ct. at 880 (Kavanaugh, J., concurring) (noting that primary elections by absentee voting began seven weeks from the date of the Supreme Court's stay).

562.   Prelim. Inj. Hr'g Tr. Day 7 (Morning), ECF No. 1420, at 152:9–154:13.

563.   Id. at 154:14–155:21.

564.   See Thompson v. Dewine, 959 F.3d 804, 813 (6th Cir. 2020) (“Moving one piece on the game board invariably leads to additional moves.”).

565.   Prelim. Inj. Hr'g Tr. Day 7 (Afternoon), ECF No. 1343, at 18:16–19:24.

566.   “Abbott sets Jan. 31 runoff for special election to replace U.S. Rep. Sylvester Turner.” Texas Tribune. Nov. 17, 2025. https://www.texastribune.org/2025/11/17/texas-18th-congressional-district-special-election-runoff-date-jan-31-houston/. (Accessed Nov. 17, 2025).

567.   Id.

568.   Prelim. Inj. Hr'g Tr. Day 7 (Afternoon), ECF No. 1343, at 20:10–18.

569.   Brooks, LULAC, and MALC Post-Hr'g Br., ECF No. 1281, at 39–40; Prelim. Inj. Hr'g Tr. Day 7 (Afternoon), ECF No. 1343, at 17:19–18:11.

570.   Contra La Union Del Pueblo Entero v. Abbott, 119 F.4th 404, 408 (5th Cir. 2024) (determining a stay pending appeal was warranted in part because the district court issued the injunction after counties had started to mail absentee ballots); Pierce v. North Carolina State Bd. of Elections, 97 F.4th 194, 226 n.11, 227 (4th Cir. 2024) (affirming the district court's denial of a preliminary injunction in part under Purcell because the election at issue was “well underway,” including the primary election results having already been certified by the time the opinion was publicly released).

571.   Cf. Chancey v. Ill. State Bd. of Elections, 635 F. Supp. 3d 627, 645 (N.D. Ill. 2022) (“And to the extent the State claims any prejudice, the problem is in large measure self-inflicted; the State, not the plaintiffs, enacted these amendments, which raise substantial constitutional concerns, less than a year before the election.”).

572.   Milligan, 142 S. Ct. at 881 (Kavanaugh, J., concurring).

573.   Id.

574.   Id.

575.   Democratic Nat'l Comm. v. Bostelmann, 977 F.3d 639, 642 (7th Cir. 2020) (“The Justices have deprecated but not forbidden all change close to an election. A last-minute event may require a last-minute reaction.”).

576.   Milligan, 142 S. Ct. at 880 (Kavanaugh, J., concurring).

577.   See Jacksonville Branch of NAACP, 2022 WL 16754389, at *3 (affirming the district court's finding that “the primary reason for applying [Purcell’s heightened] standard—risk of voter confusion—[is] lacking”).

578.   Milligan, 142 S. Ct. at 880 (Kavanaugh, J., concurring).

579.   See, e.g., ECF Nos. 1205, 1226.

580.   Contra League of Women Voters of Fla., Inc., 32 F.4th at 1371; Thompson, 959 F.3d at 813; Tenn. Conf. of the NAACP v. Lee, 105 F.4th at 898.

581.   League of Women Voters of Fla., Inc., 32 F.4th at 1372 n.7 (citation modified); see Milligan, 142 S. Ct. at 881 (Kavanaugh, J., concurring); see La Union Del Pueblo Entero, 119 F.4th at 409 (noting that Purcell is not “absolute”).

582.   Milligan, 142 S. Ct. at 881 (Kavanaugh, J., concurring); La Union Del Pueblo Entero, 119 F.4th at 409.

583.   Milligan, 142 S. Ct. at 881 (Kavanaugh, J., concurring); La Union Del Pueblo Entero, 119 F.4th at 409.

584.   Brooks, LULAC, and Gonzales Pls. Mtn. to Schedule Prelim. Inj. Hearing, ECF No. 1127, at 2.

585.   Order Scheduling Status Conf., ECF No. 1128.

586.   See generally Texas NAACP's Mot. for Prelim. Inj., ECF No. 1142; Congr. Intervenors’ Mot. for Prelim. Inj., ECF No. 1143; Gonzales Pls.’ Mot. for Prelim. Inj., ECF No. 1149; and Brooks, LULAC, and MALC Pls.’ Joint Mot. for Prelim. Inj., ECF No. 1150.

587.   See Aug. 27, 2025, Minute Entry, ECF No. 1145.

588.   See id.

589.   See Aug. 28, 2025, Minute Entry.

590.   The lawyers in this case have exhibited exemplary legal acumen, advocacy skills, and professionalism, all under intense pressure. The Court is not surprised. Throughout this years-long litigation, the lawyers on both sides have conducted themselves in the ways we hope all lawyers will, including during this case's 18-day full merits trial only five months ago. All of the advocates and parties in this matter have earned this sincere commendation by the Court.

591.   Carey v. Wis. Elections Comm'n, 624 F. Supp. 3d 1020, 1035 (W.D. Wis. 2022).

592.   Milligan, 142 S. Ct. at 881 n.1 (Kavanaugh, J., concurring).

593.   See supra Section III.D.1.

594.   Prelim. Inj. Hr'g Tr. Day 7 (Morning), ECF No. 1420, at 153:13–18.

595.   Brooks, LULAC, and MALC Joint Mot. for Prelim. Inj., ECF No. 1150, at 45; Brooks, LULAC, and MALC Post-Hr'g Br., ECF No. 1281, at 40.

596.   See supra Section III.C.

597.   See DeLeon v. Perry, 975 F. Supp. 2d at 663 (“Federal courts at all levels have recognized that violation of constitutional rights constitutes irreparable harm as a matter of law.”).

598.   See supra Section III.B.

599.   See Milligan, 142 S. Ct. at 881 n.2 (Kavanaugh, J., concurring) (finding the underlying merits “not clearcut in favor of the plaintiffs” in part because “[e]ven under the ordinary stay standard outside the election context, the State has at least a fair prospect of success on appeal—as do the plaintiffs, for that matter”). But see Lower Brule Sioux Tribe v. Lyman Cnty., 625 F. Supp. 3d 891, 933 (D.S.D. 2022) (“What is ‘entirely clearcut’ is somewhat in the eye of the beholder, and here the probability of a VRA violation is sufficiently clearcut to allow for relief as discussed above.”).

600.   See, e.g., Brnovich, 594 U.S. at 689–90, 141 S.Ct. 2321 (“[T]he legislators who vote to adopt a bill are not the agents of the bill's sponsor or proponents. Under our form of government, legislators have a duty to exercise their judgment and to represent their constituents. It is insulting to suggest that they are mere dupes or tools.”); see also supra Section III.B.4.D.iii.

601.   Milligan, 142 S. Ct. at 881 (2022) (Kavanaugh, J., concurring); cf. La Union Del Pueblo Entero, 119 F.4th at 409 (applying the conditions under which Purcell can be overcome to a permanent injunction at the district court level).

602.   League of Women Voters of Fla., Inc., 32 F.4th at 1371, 1372 n.7 (citation modified).

603.   See supra Section III.D.1.

604.   See Graham v. Nat'l Collegiate Athletic Ass'n, 804 F.2d 953, 959 (6th Cir. 1986) (collecting cases).

605.   Milligan, 142 S. Ct. at 882 (2022) (Kavanaugh, J., concurring) (quotation marks omitted) (quoting Riley v. Kennedy, 553 U.S. 406, 426, 128 S.Ct. 1970, 170 L.Ed.2d 837 (2008)).

606.   Riley, 553 U.S. at 426, 128 S.Ct. 1970 (citing Purcell, 549 U.S. at 5–6, 127 S.Ct. 5).

607.   Brooks Prelim. Inj. Ex. 254, ECF No. 1326-1, at 3.

608.   See State Defs.’ Resp. to Gonzales Pls.’ Mtn. for Prelim. Inj., ECF No. 1196-1, at 38 (first citing Wise v. Circosta, 978 F.3d 93, 98 (4th Cir. 2020) (per curiam); then citing Pierce, 97 F.4th at 226–27).

609.   Milligan, 142 S. Ct. at 881 (2022) (Kavanaugh, J., concurring).

610.   League of Women Voters of Fla., Inc., 32 F.4th at 1372 n.6.

611.   Id. at 1372.

612.   Aug. 27, 2025, Minute Entry, ECF No. 1145.

613.   Republican Nat'l Comm. v. Democratic Nat'l Comm., 589 U.S. at 424, 140 S.Ct. 1205 (per curiam) (citations omitted).

614.   See Petteway Purcell Op., 87 F.4th at 723 (Oldham, J., concurring) (collecting cases).

615.   See id. (staying orders issued by Judge Jeffrey V. Brown affecting the maps of Galveston County Commissioners Court precincts).

616.   “[T]he loss of constitutional freedoms for even minimal periods of time unquestionably constitutes irreparable injury.” BST Holdings, 17 F.4th at 618 (citation modified) (emphasis added).

617.   See generally Texas NAACP's Mot. for Prelim. Inj., ECF No. 1142; Congr. Intervenors’ Mot. for Prelim. Inj., ECF No. 1143; Gonzales Pls.’ Mot. for Prelim. Inj., ECF No. 1149; and Brooks, LULAC, and MALC Pls.’ Joint Mot. for Prelim. Inj., ECF No. 1150.

618.   See In re Landry, 83 F.4th 300, 303 (5th Cir. 2023) (collecting cases).

619.   See Veasey v. Perry, 769 F.3d 890, 895 (5th Cir. 2014) (noting that a law's “use[ ] in at least three previous elections” was a key fact in determining and “maintaining the status quo”).

620.   Milligan, 142 S. Ct. at 881 n.1 (Kavanaugh, J., concurring).

621.   Id.

622.   State Defs.’ Post-Hr'g Brief, ECF No. 1284, at 90. See In re Landry, 83 F.4th at 303.

623.   In re Landry, 83 F.4th at 303.

624.   Id. (quoting Wise v. Lipscomb, 437 U.S. 535, 540, 98 S.Ct. 2493, 57 L.Ed.2d 411 (1978)).

625.   See Veasey v. Abbott, 830 F.3d 216, 270 (5th Cir. 2016) (en banc) (collecting cases).

626.   Id. at 271.

627.   Contra Perry v. Perez, 565 U.S. at 395–97, 132 S.Ct. 934.

628.   See Veasey v. Perry, 769 F.3d at 895 (noting that a law's “us[e] in at least three previous elections” was a key fact in determining and “maintaining the status quo”).

1.   Bette Davis (as Margo Channing), All About Eve (20th Century Fox 1950).

2.   When misbehavior, or even irregular procedural behavior, occurs, there's ample precedent for bringing it to the attention of the public. See Grutter v. Bollinger, 288 F.3d 732, 810-14 (6th Cir. 2002) (en banc) (Boggs, J., dissenting) (describing the misbehavior of the Chief Judge in manipulating en banc court proceedings); see also Dunn v. Price, 587 U.S. 929, 933 (2019) (arguing that the Supreme Court's decision to vacate a stay without full discussion was improper); see also id. (“To proceed in this matter in the middle of the night without giving all Members of the Court the opportunity for discussion tomorrow morning is, I believe, unfortunate.”); see Department of State v. Aids Vaccine Advocacy Coalition, 606 U.S. ___ (2025) (contending that a stay should not be granted “with scant briefing, no oral argument, and no opportunity to deliberate in conference.”).

3.   Jackson v. Tarrant Cnty., No. 25-11055, --- F.4th ---, ---, 2025 WL 3019284, at *14 (5th Cir. Oct. 29, 2025) (citing Justice Stevens).

4.   Perez v. Abbott, 253 F. Supp. 3d 864, 988 (W.D. Tex. 2017) (three-judge redistricting court) (Smith, J., dissenting), affirmed in part and reversed in part, 138 S. Ct. 2305 (2018).

5.   Just a few days ago, plaintiff Congressman Al Green described the 2025 redistricting as “corrupt racist election rigging.” Houston Chronicle, Nov. 12, 2025, at A1.

6.   Brooks Ex. 269 (Barreto-CV 8) (receiving a $2.5 million Open Society Foundation Grant over a 36-month term ending in February 2026).

7.   Open Society Foundations, opensocietyfoundations.org/who-we-are. The Open Society Foundation was founded by George Soros, and Alex Soros is the chair of its Board of Directors.

8.   Tr. 10/4/2025 AM 22:7-8 (acknowledging that Barreto is the faculty director of the UCLA Voting Rights Project).

9.   Pierce v. N.C. State Bd. of Elections, 713 F. Supp. 3d 195, 229 (E.D.N.C.), affirmed but criticized, 97 F.4th 194 (4th Cir. 2024) (noting “profound discrepancies between the methods of analysis [Barreto] performed in his initial report and in his supplemental declaration” and finding his “belated explanation” to be “unpersuasive”).

10.   Before describing the connections of these attorneys, I emphasize that all of them serve, as officers of this court, with integrity and professionalism. Their partisan circumstance does not detract from the fact that they meet the highest standards of the profession and assist this court in the administration of justice. The same is true of the State's counsel in this case.

11.   Tr. 10/4/2025 AM 26:3-11.

12.   Jackson v. Tarrant Cnty., No. 25-11055, --- F.4th ---, ---, 2025 WL 3019284, at *14 (5th Cir. Oct. 29, 2025) (noting that “[t]he most obvious reason for mid-cycle redistricting, of course, is partisan gain”) (citing Justice Stevens).

13.   Petteway v. Galveston Cnty., 698 F. Supp. 3d 952 (S.D. Tex. 2023), reversed and remanded, Petteway v. Galveston Cnty., 111 F.4th 596 (5th Cir. 2024) (en banc).

14.   How the Open Society Foundations Support Election Integrity, opensocietyfoundations.org/newsroom/how-the-open-society-foundations-support-election-integrity.

15.   Vogel Kenneth P., Democratic Lawyer Stymied Trump in 2020. Other Efforts Played into G.O.P. Hands, www.nytimes.com/2024/10/30/us/politics/democratic-lawyer-stymied-trump-in-2020-other-efforts-played-into-gop-hands.html.

16.   Id.

17.   The point is that it's all about politics. These plaintiffs, and their counsel, and their experts, are welcome, in this court, to present their partisan views, as is the State of Texas. But if we are to tell it like it is, we must recognize that the well-funded machinery that I have just identified is all about that political crusade that these parties are free to pursue under the First Amendment. And the public is entitled to know who's really driving this bus.“The most obvious reason for mid-cycle redistricting, of course, is partisan gain.” That is the core of this case, and I will repeat it ad nauseum. Judge Brown won't tell you that. I just did.

18.   I suppose someone will say that in making these comments about the Soros connections, I'm expressing a political view, not the proper role of a federal judge. To the contrary: As I say above, the political branches engage in policy and politics. It's our job as judges to let that happen, but it's also our duty to recognize the societal and political effects of what we do, regardless of whether we approve of those downstream results. Today's ruling has dramatic political consequences by meddling in the orderly processes of a duly-elected state government. It's not “political” for me to point that out by describing the political dynamics that are inherent in the litigation of redistricting cases.

19.   Deguzman, Colleen, “You woke us up”: California Gov. Gavin Newsom, energized by Prop 50 redistricting win, thanks Texas, https://www.texastribune.org/2025/11/08/texas-california-gavin-newsom-congress-redistricting-map/

20.   Id.

21.   John C. Moritz, Texas' GOP-drawn Congressional map blocked by court in stunning blow to Republican hopes for 2026, THE HOUSTON CHRONICLE, Nov. 18, 2025 (last updated at 2:00 pm) (https://www.houstonchronicle.com/news/politics/elections/article/texas-congress-redistricting-court-case-21118138.php).

22.   Louisiana v. Callais, 3:24-cv-00122-DCJ-CES-RRS (W.D. La. Aug. 1, 2024), probable jurisdiction noted, 145 S. Ct. 434 (2025), restored to the calendar for reargument, 145 S. Ct. 2608 (2025), argued Oct. 15, 2025.

23.   Landis v. N. Am. Co., 299 U.S. 248, 254 (1936).

24.   Brown Op. at 53.

25.   Jackson v. Tarrant County, --- F.4th ---, ---, 2025 WL 3019284, at *3 (5th Cir. Oct. 29, 2025) (a mid-decade redistricting case with a preliminary-injunction posture).

26.   See Brown Op. at 53 n.159 (emphasis in original).

27.   See Jackson, --- F.4th at ---, 2025 WL 3019284, at *8 n.19 (emphasis added).

28.   Mock v. Garland, 75 F.4th 563, 587 n.60 (5th Cir. 2023) (“There is authority that the first factor—likelihood of success on the merits—is the most important of the preliminary injunction factors.”).

29.   Id. at 577 (noting that the moving party must satisfy four factors, the first of which is “a substantial likelihood of success on the merits”) (emphasis added).

30.   Id.

31.   Indeed, the language “substantial likelihood of success on the merits” is not a new formulation. It is supported by decades of precedent in the Fifth Circuit, including the case Judge Brown's opinion quotes (Brown Op. at 53 n.161). See Canal Authority Auth. of Fla. v. Callaway, 489 F.2d 567, 572 (5th Cir. 1974) (noting that the first perquisite for the extraordinary relief of preliminary injunction is “a substantial likelihood that plaintiff will prevail on the merits”) (emphasis added).

32.   To be clear, I do not deny that a sliding scale exists. I want to highlight Judge Brown's inconsistent and disjointed reasoning.

33.   See Brown Op. at 53 n.159 (emphasis in original).

34.   Brown Op. at 55.

35.   See Jackson, --- F.4th at ---, 2025 WL 3019284, at *32 n.33 (citing Justice Stevens) (emphasis added).

36.   Brown Op. at 134.

37.   Plaintiffs, during the preliminary injunction hearing, presented the testimony of six experts. However, Judge Brown, in his 161-page opinion, omits any discussion of the following five plaintiffs' experts: David Ely, Stephen Ansolabehere, Loren Collingwood, Matt Barreto, and Daniel Murray. Their collective testimony spanned several days, and they submitted hundreds of pages of expert reports. Yet, Judge Brown, despite his best efforts, fails to make a single, fleeting reference to these five experts in his lengthy opinion. This dissent, in a footnote, tells you more about these plaintiffs' experts than does Judge Brown's entire opinion does. And the reason is obvious—their testimony is unhelpful at best, or their analysis is flawed at worst. Judge Brown won't tell you that. I just did.

38.   Brown Op. at 54 (emphasis added).

39.   Unsurprisingly, that's not the law. See Mock, 75 F.4th at 587 n.60 (“Still, even with a strong likelihood of success, a district court cannot give the other factors short shrift.”).

40.   Mock, 75 F.4th at 587 n.60 (“Still, even with a strong likelihood of success, a district court cannot give the other factors short shrift.”).

41.   Brown Op. at 55 (emphasis added).

42.   Veasey v. Perry, 769 F.3d 890, 895 (5th Cir. 2014).

43.   See id. (quoting TitleMax, Inc. v. City of Dallas, 142 F.4th 322, 328 (5th Cir. 2025) (emphasis added)).

44.   See Merrill v. Milligan, 142 S.Ct. 879, 881 (Kavanaugh, J., concurring).

45.   See id.

46.   See id.

47.   Brown Op. at 146.

48.   Brown Op. at 147.

49.   See Alexander v. S.C. State Conf. of the NAACP, 602 U.S. 1, 6 (2024).

50.   Brown Op. at 145.

51.   See Tex. All. for Retired Americans v. Hughs, 976 F.3d 564, 568 (5th Cir. 2020) (noting that it was the “district court's eleventh-hour injunction that alter[ed] the status quo, not the Texas legislature's 2017 duly enacted law”) (emphasis in original).

52.   Brown Op. at 157 n.619; Brown Op. at 159 n.628.

53.   Veasey v. Perry, 769 F.3d 890, 895 (5th Cir. 2014) (emphasis added).

54.   Merrill v. Milligan, 142 S. Ct. 879, 881 (2022) (Kavanaugh, J., concurring) (noting that “[l]ate judicial tinkering with election laws can lead to disruption and to unanticipated and unfair consequences for candidates, political parties, and voters”).

55.   Veasey, 769 F.3d at 895.

56.   Judge Brown's lengthy opinion uses the word “quorum” only twice, thus giving this significant interruption—which erased the first called session—scant mention. Judge Brown makes no effort to discuss the significance of that break. I just did.

57.   See https://senate.texas.gov/cmtes/87/c625/SB6-plan-C2193.pdf, Available in interactive format and therefore much greater visual detail at https://dvr.capitol.texas.gov/Congress/56/PLANC2193 (DistrictViewer is a website maintained by the Texas Capitol).

58.   https://dvr.capitol.texas.gov/Congress/73/PLANC2308

59.   Gonzales Plaintiffs' Second Supplemental Complaint, 3:21-cv-00259-DCG-JES-JVB, ECF No. 1147, pg. 30 (August 28, 2025) (“Second Supplemental Complaint”).

60.   https://dvr.capitol.texas.gov/Congress/0/PLANC2331; Second Supplemental Complaint at 33.

61.   https://dvr.capitol.texas.gov/Congress/89/PLANC2333; Second Supplemental Complaint at 33-34.

62.   Plan C2333's summary statistics including VAP and CVAP breakdowns are also available at https://capitol.texas.gov/tlodocs/892/districtplanrpts/pdf/HB00004H_PLANC2333.pdf. (“C2333 summary statistics”).

63.   602 U.S. at 6.

64.   Cooper v. Harris, 581 U.S. 285, 291 (2017) (citation modified).

65.   Alexander v. S.C. State Conf. of the NAACP, 602 U.S. 1, 8 (2024).

66.   Cooper, 581 U.S. at 291.

67.   Brnovich v. Democratic Nat'l Comm., 594 U.S. 647, 689–90 (2021).

68.   All subsequent transcript citations in this section refer to Tr. 10/7/25 AM, unless otherwise noted.

69.   Brown Op. at 79-104.

70.   To avoid ambiguity, it is important to note that Representative Hunter was Chairman of the Special Select Committee on Redistricting, while Chairman Vasut is Chairman of the overall Redistricting Committee.

71.   Brown Op. at 66-79.

72.   Brown Op. at 59-66.

73.   See Alexander, 602 U.S. 1, 10 (2024) (“This presumption of legislative good faith directs district courts to draw the inference that cuts in the legislature's favor when confronted with evidence that could plausibly support multiple conclusions.”) (citing Abbott v. Perez, 585 U.S. 579, 610-612 (2018)).

74.   Brown Op. at 56.

75.   See, e.g., Antonin Scalia, Common-Law Courts in a Civil-Law System: The Role of the United States Federal Courts in Interpreting the Constitution and Laws, A MATTER OF INTERPRETATION: FEDERAL COURTS AND THE LAW, 16-17 (Amy Gutmann, ed., 1997).

76.   Brown Op. at 15-35.

77.   Alexander, 602 U.S. at 10.

78.   I did tell you to buckle up, didn't I?

79.   Brown Op. at 67-69.

80.   Brown Op. at 79-90.

81.   See Brown Op. at 67-79 (covering Hunter's recitation of demographic statistics and mentions of Petteway and Rucho).

82.   Miller v. Johnson, 515 U.S. 900, 916 (1995).

83.   Brown Op. at 74-75.

84.   Brown Op. at 77; contra Prelim. Inj. Hr'g Tr. Day 8 (Afternoon), ECF No. 1344, at 115:2-7 (Hunter: “As based in my previous commentary on Rucho, this map is based on partisanship, political performance. And for all of you here, it has enhanced and increased Republican partisanship, enhanced performance. The intent of the changes was to increase Republican political performance in existing Republican districts from the proposed plan.”).

85.   Brown Op. at 74-75; see Brooks Prelim. Inj. Ex. 282, ECF No. 1326-28 at 1 (Burrows: “I want to thank Representative Todd Hunter for carrying this bill and for his tireless efforts ensuring the new map is not only constitutional, but secures Republican representation in Congress ․ Today's passage of the congressional map has ushered in a new chapter of Republican unity․”).

86.   How can you avoid talking about Petteway? If representatives asked about Petteway had said, for example, “I don't want to talk about that,” Judge Brown's motivated reasoning could twist such a response into concealing their racist intent. That style is conspiracy-theorist thinking.

87.   Brooks Prelim. Inj. Ex. 327-T, ECF No. 1327-27 at 3 (“So I am on the main redistricting committee also, but I'm not on the special select committee that's reviewing these particular maps ․ I think what I would say is that I know that we certainly have the right to look at the maps and make changes. I think the courts have consistently held that redistricting for purposes of political performance by either party is acceptable.”).

88.   Brooks Prelim. Inj. Ex. 339-T, ECF No. 1411-5, at 1-2.

89.   Brown Op. at 79-90.

90.   See Prelim. Inj. Hr'g Tr. Day 8 (Afternoon), ECF No. 1344, at 117:11 – 118:18 (Vasut: “I see no evidence that this was racially drawn. This is a political performance map. I haven't looked at those. The question I had when I, you know, looked at this – and I was evaluating it myself, was – does this improve the political performance of Republicans in Texas? Which is where we have been trending and what we need to do to respond nationally. This is not just a Texas issue. It's a nationwide issue, it's perhaps one of the biggest issues that we're taking up. And when we've seen all of these blue states over-perform with their maps and Texas is underperforming, that puts Republicans at a distinct disadvantage nationwide, and it's right for Texas to step up. So I have not seen any evidence that this map was racially based. What I have seen is evidence that this map was politically based. And that's totally legal, totally allowed, totally fair․ I disagree with the assumption that this process had anything to do with the DOJ letter. Yeah, they sent a letter, but as you know, the proclamation called us in to do congressional redistricting, and we did congressional redistricting when we passed HB4 based off of political performance. So I frankly don't care what the DOJ letter said – and I think it's pretty clear that no one does. And I ought to probably prepare to sign this bill. So this bill was not based off of that DOJ letter. That bill was based off of improving political performance.”).

91.   See Prelim. Inj. Hr'g Tr. Day 7 (Afternoon), ECF No. 1343, at 67–70 (“[L]et's stop pretending that this is all about race. It is about values. It is about representation—real representation. The fact that we are redrawing the maps is to ensure that ․ the people are able to have representation that reflects their values, not their last name, not their skin color․ And with that, members, I proudly stand and look forward to casting my vote in favor of House Bill 4.”).

92.   Brown Op. at 66-69.

93.   The earliest that Hunter was involved with the maps was apparently July 23. See Prelim. Inj. Hr'g Tr. Day 8 (Morning), ECF No. 1420, at 140-141.

94.   See Prelim. Inj. Hr'g Tr. Day 1 (Morning), ECF No. 1414, at 127:18-128:9; 129:1-3.

95.   Asked when he became aware that the White House was having conversations about redistricting, Kincaid answered, “It would have been earlier in 2025․ I was aware that people were meeting with White House officials on redistricting probably [in] February or March.” Morning Transcript, 10/7/2025, 58:13-17 (Direct Exam of Adam Kincaid). And when asked when he first began speaking with a Texas national committeeman about redistricting in Texas, Kincaid answered, “I believe it was in March was when I first had a conversation with Robin [Armstrong] about this.” Id. at 59:22-23.

96.   In response to defendant counsel's question, “How often would you say you draw maps․ ?” Kincaid replied, “We do a lot of different things in [the National Republican Redistricting Trust]. But when it's quiet, I'll sit down and I'll look at a map and see what I can do in different places. So it's regularly that part of my job is to look at maps and see what is possible or what would have been possible, yeah, across the entire country.” Prelim. Inj. Hr'g Tr., 10/7/2025, 36:24 – 37:4 (Direct Exam of Adam Kincaid).

97.   See id. at 58-59.

98.   See Prelim. Inj. Hr'g Tr. Day 6 (Afternoon), ECF No. 1342, at 20–22. Senator King either had a lapse of memory or was concealing the number of conversations he had with Kincaid. Given Kincaid's remarkably lucid, rapid-fire, and forthright demeanor on the stand—compared to King's calculated demeanor—I think it is obvious that Kincaid is telling the truth. Additionally, Kincaid's was entirely consistent with Senator Hinojosa, who had a sober demeanor and was another sponsor of the bill.

99.   Prelim. Inj. Hr'g Tr. Day 2 (Morning), ECF No. 1415, at 90–91 (“Q. “Now, it's been stated by others that redistricting was in the conversation prior to [the DOJ Letter discussed below] ․ What do you say to that? | [REPRESENTATIVE THOMPSON:] I heard it all during the session, and I made inquiries about it. And I asked [Chairman Hunter] ․ if they were going to be redistricting․ And subsequent he said he didn't know. You know, I think he told me he was unaware of any redistricting. And he kind of brushed it off as though it just might have been just a rumor or something, you know.”); Morning Transcript, 10/7/2025, 62:1-3 (“I think the final phase of the redistricting for 2025 probably started late June or early July”).

100.   Brown Op. at 17 n.48.

101.   See Brown Op. at 15-17; also Defs.' Resp. Intervenors' & Tex. NAACP's Prelim. Inj. Mot., ECF No. 1195, at 23–24 (“Given the danger to President Trump's legislative agenda posed by [the] 2026 elections and the historical trend of the presidential party doing poorly in non-presidential election years, there was a great deal of political pressure placed on the State of Texas to match the political gerrymandering of Democrat states. This pressure only intensified when other states, especially California, pledged to perform mid-decade redistricting to make their already one-sided congressional maps even more favorable to Democrats․ None of those factors indicate race was involved ․”).

102.   227 U.S. F.3d 504 (5th Cir. 2000).

103.   Id. at 510.

104.   Id.

105.   Supra, Kincaid testimony at 31-32.

106.   See Fusilier v. Landry, 963 F.3d 447, 466 (5th Cir. 2020).

107.   United States v. O'Brien, 391 U.S. 367, 384 (1968); also Fusilier v. Landry, 963 F.3d 447, 466 (5th Cir. 2020) (discussing O'Brien); N. & S. Rivers Watershed Ass'n v. Town of Scituate, 949 F.2d 552, 555 n.6 (1st Cir. 1991) (“[S]tatements of individual legislators, even the sponsors of legislation, should not be given controlling effect.”), overruled on other grounds by Blackstone Headwaters Coal., Inc. v. Gallo Builders, Inc., 32 F.4th 99 (1st Cir. 2022).

108.   594 U.S. 647 (2021).

109.   Id. at 689-90.

110.   Brown Op. at 100-104.

111.   Brown Op. at 3.

112.   Brown Op. at 105.

113.   Brown Op. at 38.

114.   See, e.g., Brooks Prelim. Inj. Ex. 335-T, ECF No. 1328-1.

115.   Brown Op. at 105.

116.   Brown Op. at 106-7.

117.   Brown Op. at 106.

118.   See Brooks Prelim. Inj. Ex. 267, ECF No. 1326-14 at 5-6.

119.   Id. at 6.

120.   Brown Op. at 107 n.403.

121.   Brown Op. at 107.

122.   Tr. 10/7/25 AM 148:10-11.

123.   Brown Op. at 3 (emphasis added).

124.   Judge Brown states the law correctly here, Brown Op. at 89, but later misapplies it.

125.   Brown Op. at 41-50.

126.   Alexander, 602 U.S. at 6.

127.   Supra, Kincaid testimony at

128.   Supra, Kincaid testimony at 36, 48.

129.   Infra at 64 et seq.

130.   Bethune-Hill v. Va. State Bd. of Elecs., 580 U.S. 178, 191-192 (2017) (quoting Ala. Leg. Black Caucus v. Alabama, 575 U.S. 254, 262 (2015)).

131.   Brown Op. at 107, n.403.

132.   Supra at Kincaid Testimony, 36-37.

133.   Alexander v. S.C. State Conference of the NAACP, 602 U.S. 1, 7 (2024) (quoting Miller v. Johnson, 515 U.S. 900, 915–16 (1995)).

134.   Id. (quoting Miller v. Johnson, 515 U.S. 900, 915 (1995)).

135.   726 F. Supp. 3d 1322, 1364–65 (N.D. Fla. 2024).

136.   Brown Op. at 105.

137.   Alexander, 602 U.S. at 7 (quoting Miller v. Johnson, 515 U.S. 900, 916.)

138.   Alexander, 602 U.S. at 8.

139.   Brown Op. at 108-127.

140.   See Brown Op. at 108, lauding her credentials.

141.   Schulz v. Pennsylvania R.R. Co., 350 U.S. 523, 526 (1956).

142.   There are “causes of action in which the law predicates recovery upon expert testimony.” Salem v. U.S. Lines Co., 370 U.S. 31, 35 (1962).

143.   Daubert v. Merrell Dow Pharm., Inc., 509 U.S. 579, 595 (1993) (citation omitted).

144.   Federal Rules of Evidence 702, 703.

145.   Alexander, 602 U.S. at 15.

146.   Id. at 33, citing Bethune-Hill, 580 U.S. at 191 (“the basic unit of analysis for racial gerrymandering claims ․ is the district”); Ala. Legislative Black Caucus, 575 U.S. at 262-63 (a racial gerrymandering claim “does not apply to a State considered as an undifferentiated ‘whole’ ”); see also Alexander at 45 (Thomas, J., concurring in part) (“A legislature seeking to gerrymander a district will often proceed by “packing” or “cracking” groups of minority voters ․ But, in areas where ‘political groups ․ tend to cluster (as in the case with Democratic voters in cities)’ apparent packing or cracking can simply reflect ‘adherence to compactness and respect for political subdivision lines’ or ‘the traditional criterion of incumbency protection.’ This case exemplifies the problem—Judge Brown observes that Dr. Moon Duchin's report failed to ‘account for’ the traditional districting principles of ‘partisanship or core retention’ in ‘assessing whether the Enacted Plan ‘cracks’ black voters among multiple districts․ The difference between illegitimate packing and the legitimate pursuit of compactness is too often in the eye of the beholder.”) (quoting Vieth v. Jubelirer, 541 U.S. 267, 278)).

147.   Id. at 33.

148.   We therefore pass over her analysis of effective minority representation, which was disputed at the hearing due to her changes of denominators between two recent editions of her report to include more past elections in Austin related to Rep. Lloyd Doggett. See Expert Report of Dr. Moon Duchin, September 7, 2025 (“Duchin Report”), ECF No. 1384-8 at 9.Also, between the August and September editions of her report, she made several material changes to her box plot histograms, compare Expert Report of Dr. Moon Duchin, August 25, 2025 (“Duchin's August Report”), ECF 1142-6, pg. 14-15; with Duchin Report at 14-15. These changes generally ramped up her estimates of outlier behavior. But if she was so certain of her report's results in August, what can explain her materially changed results in September? What is to suggest that her results may not change again, if an out-of-state academic again needs to fly into Texas to override the will of tens-of-millions of voters in the state?

149.   Expert Report of Dr. Moon Duchin (“Duchin Report”), ECF 1142-6, pg. 5-6.

150.   Duchin Report at 1-2, 14-15.

151.   I use the term random walk here to refer to Markov Chain analysis, which is a step-wise outcome generation process where the prior state probabilistically influences the subsequent state. A demonstrative thought experiment is the “drunk at the lamppost” scenario. In this experiment, a drunk moves randomly from the lamppost, in any direction. Where is the most likely place for him to end up after an hour? Right back at the lamppost.As to spanning tree analysis, this is a topological exercise that in Euclidean space collapses to geometric connection of vertices. Put simply, this is connect-the-dots, with probabilistic weights that affect the probability of the spanning tree's connecting to the next vertex. Duchin describes these weights as “surcharges” geared towards compactness. Duchin Report at 19.

152.   Duchin Report at 14-15, 22-23 (the three histograms and Appendix E).

153.   See also Brown Op. at 108-122.

154.   In statistics, an internal threat to validity is a factor that can undermine the proposed relationship between a variable an outcome. The simplest example is “omitted variable bias,” where a third factor C drives the relationship between observed factors A and B. Since Hume, we have all been aware that correlation does not imply causation. Omitted variable bias is one of the phenomena that drives this distinction.

155.   An external threat to validity limits the relationship between a research study and its application to the external world. While the most famous examples typically come from the medical literature, as in placebo trials affecting patients' behavior, which demands the double-blind protocol, a simpler example is that external conditions may change during and after the time of the study. Instantly, this could include Hispanic voters shifting their preferences to the Republican party in Texas, rather than remaining a constant figure, as was developed by Dr. Lewis's expert testimony and report. See Expert Report of Jeffrey B. Lewis, 3:21-cv-00259-DCG-JES-JVB ECF No. 1386 at pg. 4-6 (Exhibit 570) (“Lewis Report”).

156.   See Duchin Report at 14–15, lacking any labels of the blue dots in her histograms. She was invited to clarify the meaning of these actual outcomes, compared to the ensemble simulation, during her oral testimony. Morning Transcript 10/4/2025 at *130, ll.10-24. However, she failed to do so. Neither she nor plaintiffs' counsel ever clarified which dot corresponds to which real outcome district's composition. It is best to scrutinize these small points, which compared to the broad labelling on the Y-axis provide little guidance, and compare them to the available statistics in the C2333 tables to figure out exactly what she means. Cf. C2333 summary statistics at pg. 13-15, https://capitol.texas.gov/tlodocs/892/districtplanrpts/pdf/HB00004H_PLANC2333.pdf.

157.   For example, Duchin bizarrely asserts that congressional districts CD-29, CD-18 and CD-9 were rotated in their name assignment so as to confuse any reviewing body. Duchin Report at 6. But the reason is not particularly confusing: CD-18, by virtue of having been Sheila Jackson Lee's former seat, the “Barbara Jordan district,” while also being the easternmost seat in central Harris County, could not have been moved outside the county without provoking greater uproar. So, it made sense to move CD-9, at least in terms of name, even though CD-9 substantially swapped locations with CD-18 measured by core retention, such that CD-18 remained a safe Democrat seat. Duchin's assertion that the name change was made for conspiratorial and racist reasons suggests her motivated reasoning, as contrasted with dispassionate expert testimony.

158.   Federal Rule of Evidence 704; see generally, Molly Treadway Johnson et Al., Expert Testimony in Federal Civil Trials, a Preliminary Analysis, FEDERAL JUDICIAL CENTER (2000).

159.   See Duchin Report at 6, showing improved scored on Polsby-Popper, Reock, and Block Cut Edges in Plan C2333 as compared to Plan C2193. Higher scores on Polsby-Popper and Reock are better, reflecting greater “circle-like” nature to a district, where a perfect circle would have the highest score. Lower scores are better on Block Cut Edges, reflecting the total ‘scissoring’ or serration in the plan.

160.   Duchin Report at 5-6.

161.   Id.; Duchin Report at 16 (conclusions).

162.   Supra Kincaid testimony at 28-29; compare Prelim. Inj. Hr'g Tr. Day 6 (Morning), at 37-39; with Prelim. Inj. Hr'g Tr. Day 5 (Morning), at 84:15-23.

163.   See Defs.' Post-Hr'g Br., ECF No. 1284, at 51–52.

164.   Brown Op. at 126-127.

165.   Supra, Kincaid testimony at 30.

166.   See Duchin Report at 14-15, considering the core targets and tails of the generated ensembles.

167.   Duchin Report at 1-2.

168.   Supra, Kincaid testimony at 30.

169.   Duchin Report at 16.

170.   Id. at 1-2. Judge Brown discusses her cluster method, Brown Op. at 108-110, but fails to consider its constraints on variance and how those may drive skew.

171.   602 U.S. at 33.

172.   See Duchin Report at 14-15.

173.   Id.

174.   Duchin Report at 22-23.

175.   Id. at 22 (“Population balance is enforced by requiring each step to leave districts within 1% of ideal population.”).

176.   Id. at 22-23.

177.   Supra, Kincaid testimony at 28-29.

178.   See Duchin Report at 20 (“Republican performance: Republicans overall have at least as many wins in each cluster as in C2333”). But the map-drawer did not care about 50.1% wins—he cared about safe wins. Cf. Morning Transcript 10/4/2025 at *63, ll. 5-12.

179.   Supra, Kincaid testimony at 28-29.

180.   Morning Transcript 10/4/25, at 9:4-10; see also Defs.' Post-Hr'g Br., ECF No. 1284, at 54.

181.   Brown Op. at 112-116.

182.   As to CD29 and CD9, answering the question: What is the significance of altering the urban, rural -- the urban-rural demographics of the county? Duchin's answer: Well, this is a kind of configuration that's often consistent with taking, as I said earlier, pieces of more diverse urban population and combining them with more rural population. This is the kind of reconfiguration you would often see when trying to change the partisan composition of a district. This is consistent with partisanship, but it also has demographic markers. (Transcript Morning 10/4/25,*51, ll. 9-17).

183.   See Duchin Report at 22, 23 (on “filtering down to maps that meet all of these conditions․”). Note, this is not filtering in the sense of strictly pulling a subset—this is filtering in the sense of a matrix transformation, as is shown by the fact that the curves are sometimes non-overlapping, see Id. at 15, Figure 9, Plot 4 (non-overlap).

184.   See Duchin Report at 15, figure 8 (“The results of the algorithmic runs are shown in the boxplots in black, where the whiskers span from the 1st to the 99th percentile in each case.”).

185.   See Duchin Report at 15, figure 10.

186.   See Duchin Report at 14, figure 8.

187.   An additional wrinkle here is that Duchin employs the pre-Petteway “all persons of color” approach, meaning that she aggregates together Black, Hispanic, Asian and Pacific Island, Native American, and the growing group of ‘other’ and mixed-race voters. Therefore, her high outlier in Harris-Fort Bend may itself be artificially inflated through the inclusion of these voters who are not material to Judge Brown's theory and should properly have been accounted for in the ensemble analysis.Additionally, it neglects the possibility that through the census counting of non-citizen voters, for example Hispanic voters in Colony Ridge in Liberty County, i.e. CD-9 in C2333, there is a deflation in the CVAP figure. Kincaid was required to draw maps equi-populously based on census results, so any counting of non-citizens may correspondingly deflate the CVAP figure for that district. This dynamic may negate the inference of cracking in at least one of Duchin's Harris/Fort Bend outliers, see Duchin Report at 15, Figure 9.

188.   Morning Transcript 10/4/2025 at *51-53.

189.   Id. at *52.

190.   See Plan C2333 summary statistics at pg. 14, https://capitol.texas.gov/tlodocs/892/districtplanrpts/pdf/HB00004H_PLANC2333.pdf.

191.   See Alexander, 602 U.S. at 1.

192.   Generally, note that County > Voting Precinct > Census Block > House, in terms of levels of partisan (or racial) voter data.

193.   Supra Kincaid testimony at 28-29.

194.   Prelim. Inj. Hr'g Tr. Day 6 (Morning), at 37-39.

195.   Brown Op. at 108-127.

196.   Brown Op. at 112.

197.   See Duchin Report at 14 (“The orange boxplot shows the statistics once we have filtered the ensembles to only include plans that meet the full checklist of districting principles.”); Id. at 22-23 (Appendix E) (explaining the first round of district generation and the second round of winnowing).

198.   Brown Op. at 113.

199.   Nevertheless, within orthodox social science, this would be a fine measure to identify outlying outcomes with a p-certainty value below 0.05.

200.   Brown Op. at 116.

201.   Id.

202.   Id. at 118-20.

203.   Id. at 119, 122.

204.   Id. at 122.

205.   Id. at 112.

206.   Id. at 128.

207.   Also, where Judge Brown claims that “not one of them had racial demographics that looked anything like the enacted map,” Brown Op. 127, this is flatly without logical foundation, given that 400 maps were off either end of the tails, supra at 76.

208.   Id. at 125 (quoting Tex. NAACP Prelim. Inj. Ex. 208, ECF No. 1384-8, at 23) (Duchin Report at 23).

209.   Id. at 127.

210.   Id. at 127-38.

211.   In probability theory, the variance space on any larger set is larger than the variance space on a smaller set of elements using the same draw.

212.   Brown Op. at 130-34.

213.   See id. at 133.

214.   Tr. 10/6/2025 AM 75:25-77:5.

215.   See Brown Op. at 134.

216.   Alexander v. S.C. State Conf. of the NAACP, 602 U.S. 1, 34 (2024) (quoting Cooper v. Harris, 581 U.S. 285, 337 (2017) (emphasis added)).

217.   Plaintiffs' counsel cited Duchin's new job as reason for the delay in immediately turning over Duchin's materials to the State. Tr. 10/4/2025 14:24-15:1. When asked, plaintiffs' counsel sidestepped the question whether Duchin, a well-paid, purported “expert,” did nothing between September 26th and the day of her testimony. See id. 15:2-14. Indeed, plaintiffs' counsel did not explicitly foreclose the possibility that nothing was done. See id.

218.   Plaintiffs, during the preliminary injunction hearing, presented the testimony of six experts. However, Judge Brown, in his 161-page opinion, omits any discussion of the following five plaintiffs' experts: David Ely, Stephen Ansolabehere, Loren Collingwood, Matt Barreto, and Daniel Murray. Their collective testimony spanned several days, and they submitted hundreds of pages of expert reports. Yet, Judge Brown, despite his best efforts, fails to make a fleeting reference to these five experts. If Judge Brown could, he would. For what it's worth, this dissent, in a footnote, tells you more about these plaintiffs' experts than Judge Brown's entire opinion does. And the reason is obvious—their testimony is unhelpful at best, or their analysis is flawed at worst.Judge Brown won't tell you that. I just did.

219.   Plaintiffs' top expert Matt Barreto is a Soros operative. His CV confirms it. He expects to receive $2.5 million in his Soros piggybank. Soros has been pumping money into Barreto's UCLA Voting Rights Project for years. And this steady supply of money will not stop until the new year, at the earliest. Unsurprisingly, Barreto has been on quite the road show, parading across the country opposing Republican redistricting. Judge Brown could not plausibly conjure up anything helpful from Barreto's testimony, which lasted from 9:20 AM – 3:43 PM (including breaks) on October 4th. If Judge Brown could, he would. His testimony is untouchable (and not in the good way).Judge Brown won't tell you that. I just did.

220.   Tr. 10/6/2025 AM 137:14-138:10.

221.   Duchin Decl., Ex. F at 31. (Document 1142-6).

222.   To be clear, it is perfectly appropriate for someone to file an amicus brief. In fact, amici often help judges understand complex issues and background information. I note her involvement in the Rucho case because she remarked, at the preliminary injunction hearing, that partisan gerrymandering is not her “motivating influence.” Tr. 10/6/2025 AM 137:14-138:10.

223.   Rucho v. Common Cause, 588 U.S. 684, 742 (2019) (Kagan, J., dissenting) (citing the Brief for Mathematicians et al. as Amici Curiae).

224.   Rucho Brief for Mathematicians et al. as Amici Curiae at *12 (arguing that vote dilution, on the basis of partisanship, is problematic).

225.   Judge Brown won't tell you that. I just did.

226.   Allen v. Milligan, 599 U.S. 1, 34 (2023).

227.   Judge Brown won't tell you that. I just did.

228.   Alexander v. S.C. State Conf. of the NAACP, 602 U.S. 1, 33 (2024).

229.   Id. at 33.

230.   Id.

231.   Id. (emphasis added).

232.   Judge Brown relies exclusively on the testimony of one of the six plaintiffs' experts, Duchin. The testimony of the other five is anywhere on the spectrum between unusable at best to deeply flawed at worst. It speaks for itself.

233.   Compare Brown Op. 35-49; with C2333 summary statistics at 13-15, https://capitol.texas.gov/tlodocs/892/districtplanrpts/pdf/HB00004H_PLANC2333.pdf.This Anglo language itself tramples over any nuance between sub-groups of Hispanic Americans like Cubans, sub-groups of Whites such as Jewish or Arab Americans, or the growing populations of Asian, multi-racial and “other” Americans. The tri-racial vision advanced by plaintiffs, of an Anglo vs Black vs Hispanic political climate, embraces the coalitional logics overturned by Petteway, and defies any nuanced and mature conversation about Texas politics and its complex demographic evolution.

234.   There are also major VAP vs CVAP distinctions observable in the Hispanic population. For example, in CD-9, one of the districts analyzed by Judge Brown and Dr. Duchin, there are more than 100,000 non-citizen residents, see C2333 summary statistics at 15. https://capitol.texas.gov/tlodocs/892/districtplanrpts/pdf/HB00004H_PLANC2333.pdf. This distinction may account for Duchin's allegation about CD9's being “cracked” or “packed,” supra at Duchin discussion, 73-75.

235.   See Lewis Report at 4, figure 1, panel 3 (“Trump Support (G24)” y-axis, “Percent Hispanic Voters” x-axis).

236.   CDs 28 and 34 appear in one footnote quoting Chairman Hunter with zero further commentary from Judge Brown. Brown Op. at 79.

237.   Alexander, 602 U.S. at 10.

238.   Brown Op. at 76.

239.   See Alexander, 602 U.S. at 10.

240.   Tr. 10/7/25 AM 37:20–24.

241.   Tr. 10/7/25 AM 61:20–24.

242.   Brown Op. at 128.

243.   See Texas Trends 2025, Univ. of Houston, Oct. 2025, https://www.uh.edu/hobby/txtrends/2025/ (finding that 68% of Texans believe partisan gerrymandering is a major problem and 21% believe it's a minor problem); see also Alexander Rossell Hayes, Large majorities of Americans say gerrymandering is a major problem, unfair, and should be illegal, YouGov, https://today.yougov.com/politics/articles/52740-large-majorities-americans-say-gerrymandering-major-problem-unfair-should-be-illegal-redistricting-texas-california-poll (finding that 76% of Americans thinks gerrymandering is a major problem).

244.   Supra at 75-77.

245.   Brown Op. at 158-159.

246.   Relating to the composition of the districts for the election of members of the United States House of Representatives from the State of Texas, Tex. H.B. 4, 89th Leg. 2d Spec. Sess., Art. III § 3 (2025).

247.   See Printz v. U.S., 521 U.S. 898, 912 (1997) (“[S]tate legislatures are not subject to federal direction.”) (citing New York v. U.S., 505 U.S. 144, 112 (1992).

248.   The Tenth Amendment imposes the same anti-commandeering limit on federal courts and the federal legislature, see Murphy v. NCAA, 584 U.S. 453, 471 (The legislative powers granted to Congress are sizable, but they are not unlimited. The Constitution confers on Congress not plenary legislative power but only certain enumerated powers. Therefore, all other legislative power is reserved for the States, as the Tenth Amendment confirms.”).

249.   Bush v. Vera, 517 U.S. 952, 1003 (1996) (Thomas, J., concurring).

250.   In re Landry, 83 F.4th 300, 303 (5th Cir. 2023).

251.   Brown Op. at 160.

252.   See Brown Op. at 158-59.

253.   This parallels the U.S. Code, 1 U.S.C. § 108 (repeal of repealing statute does not reinstate the former statute).

254.   Cf. Shelby Cnty. v. Holder, 570 U.S. 529, 557 (ending the § 5 coverage and preclearance requirement).

255.   606 U.S. 831, 837 (2025); see also Jonathan F. Mitchell, The Writ-of-Erasure Fallacy, 104 VA. L. REV. 933, 936 (2018).

256.   209 U.S. 123, 159-160 (1908).

257.   527 U.S. 308, 318 (1999).

258.   Id. at 332.

259.   Id. at 333.

260.   606 U.S. at 846 (quoting Grupo Mexicano, 527 U.S. at 332).

261.   Id. at 861.

262.   This matters because of the diagonal federalism relationship between a federal court and a state legislature, infra at 95.

263.   Pidgeon v. Turner, 538 S.W.3d 73, 88 n.21 (Tex. 2017).

264.   State by & Through Off. of Att'y Gen. of Texas v. City of San Marcos, 714 S.W.3d 224, 244 (Tex. App. 2025), review denied (Sept. 12, 2025).

265.   See City of San Marcos, 714 S.W.3d at 244 (“The Texas Constitution vests the City of San Marcos, not the Court, with authority to adopt and repeal ordinances.”) (quoting also Ex parte E.H., 602 S.W.3d 486, 502 (Tex. 2020) (Blacklock, J., dissenting) (“Courts are not legislatures. The Texas Constitution reserves the law-making and law-rescinding powers to the Legislature, and it prohibits the judiciary from ‘exercis[ing] any power properly attached to either of the other [ ] [branches].’ ” (quoting Tex. Const. art. II, § 1))).

266.   See Civil Rights Cases, 109 U.S. 3, 10 (1883) (The Fourteenth Amendment, “nullifies and makes void all State legislation, and State action of every kind, which impairs the privileges and immunities of citizens of the United States, or which injures them in life, liberty, or property without due process of law, or which denies to any of them the equal protection of the laws.”); Id. at 25 (of the Civil Rights Act of 1875, “we are of opinion that no countenance of authority for the passage of the law in question can be found in either the Thirteenth or Fourteenth Amendment of the Constitution; and no other ground of authority for its passage being suggested, it must necessarily be declared void․); Plessy v. Ferguson, 163 U.S. 537, 546 (1896) (“In the Civil Rights Cases ․ it was held that an act of congress entitling all persons within the jurisdiction of the United States to the full and equal enjoyment of the accommodations ․ was unconstitutional and void, upon the ground that the fourteenth amendment was prohibitory upon the states only.”).

267.   636 S.W.3d 247, 263 (Tex. App. 2021), rev'd on other grounds sub nom. Lilith Fund for Reprod. Equity v. Dickson, 662 S.W.3d 355 (Tex. 2023) (“When a legislative act is declared to be unconstitutional, the act is ‘absolutely null and void,’ and has ‘no binding authority, no validity [and] no existence.’ ”) (citing Ex parte Bockhorn, 138 S.W. 706, 707 (Tex. Crim. App. 1911) (an unconstitutional law should be viewed as “lifeless,” as “if it had never been enacted,” given that it was “fatally smitten by the Constitution at its birth.”).

268.   Id.

269.   Tex. Const. art. II § 1.

270.   Tex. Const. art. III § 1.

271.   Martinez v. State, 503 S.W.3d 728, 733-34 (Tex. App.—El Paso 2016, pet. ref'd).

272.   Compare In re Texas Dep't of Fam. & Protective Servs., 660 S.W.3d 161, 171-172 (Tex. App. 2022) (“the trial court unduly interfered with the powers of the legislative branch when it ordered the Department [of Family and Protective Services] to submit [certain detailed] written offers to specific child-placing agencies”); with INS v. Chadha, 462 U.S. 919, 952 (1983) (“Whether actions taken by either House are, in law and fact, an exercise of legislative power depends not on their form but upon whether they contain matter which is properly to be regarded as legislative in its character and effect.”) (internal quotation omitted).

273.   See, e.g., Voting for Am., Inc. v. Steen, 732 F.3d 382 (5th Cir. 2013) (vacating preliminary injunction entered against Texas voter-registration laws); Planned Parenthood Ass'n, Inc. v. Suehs, 692 F.3d 343 (5th Cir. 2012) (vacating preliminary injunction entered against the enforcement of a law excluding Planned Parenthood from the Texas Women's Health Program); Tex. Med. Providers Performing Abortion Servs. v. Lakey, 667 F.3d 570 (5th Cir. 2012) (vacating preliminary injunction entered against Texas informed-consent law).

274.   558 U.S. 310 (2010).

275.   See Mitchell at 989-92 (comparing Citizens United with the still-extant 52 U.S.C. § 30118(a) (2012) (“It is unlawful ․ for any corporation ․ to make a contribution or expenditure in connection with any election․”)).

276.   See CASA, 606 U.S. at 837; also Brown v. Plata, 563 U.S. 493, 550 (2011) (Scalia, J., dissenting) (criticizing an overbroad structural injunction); generally Missouri v. Jenkins, 515 U.S. 70, 88 (1995) (limiting a school segregation structural injunction and remarking “the ‘principle that the nature and scope of the remedy are to be determined by the violation means simply that federal-court decrees must directly address and relate to the constitutional violation’ ” (quoting Milliken v. Bradley, 433 U.S. 267, 281-282 (1977)).

277.   Cf. South Bay United Pentecostal Church v. Newsom, 140 S. Ct. 1613, 1613-14 (Mem.) (Roberts, C.J., concurring) (balancing political and health considerations during the Covid Era's shutdown “should not be subject to second-guessing by an ‘unelected federal judiciary,’ which ․ is not accountable to the people.” (quoting Garcia v. San Antonio Metro. Transit Auth., 469 U.S. 528, 545 (1985)).

278.   Learned Hand, THE BILL OF RIGHTS: THE OLIVER WENDELL HOLMES LECTURES, 70 (1958).

279.   141 S. Ct. 28, 29 (Gorsuch, J., concurring).

280.   Id. at 33 (Kavanaugh, J., concurring).

281.   141 S. Ct. 9, 10 (2020) (Mem.) (Kavanaugh, J., concurring).

282.   See Landry, 83 F.4th at 303 (“the Supreme Court has repeatedly reminded lower federal courts that if legislative districts are found to be unconstitutional, the elected body must usually be afforded an adequate opportunity to enact revised districts before the federal court steps in to assume that authority ․ [such that] that ‘legislative reapportionment is primarily a matter for legislative consideration and determination.’ ” (quoting Reynolds v. Sims, 377 U.S. 533, 586 (1964); Wise v. Lipscomb, 437 U.S. 535, 540 (1978)).

283.   LULAC Second Supplemental Complaint at *6, No. 3:21-cv-00259-DCG-JES-JVB, ECF No. 1147 (August 28, 2025, W.D. Tex. – El Paso).

284.   § 11, 1 Stat. 78.

285.   See CASA, 606 U.S. at 841 (2025) (“Though flexible, this equitable authority is not freewheeling. We have held that the statutory grant encompasses only those sorts of equitable remedies ‘traditionally accorded by courts of equity’ at our country's inception.”) (quoting Grupo Mexicano de Desarrollo, S.A. v. Alliance Bond Fund, Inc., 527 U.S. 308, 319 (1999)).

286.   Petteway v. Galveston County, 111 F.4th 596 (5th Cir. 2024) (en banc).

287.   Cf. Brown Op. at 96-99 (refusing to credit Kincaid's testimony).

288.   See, e.g., Guy Marzorati, California voters OK new congressional lines, boosting Democrats ahead of midterms, NATIONAL PUBLIC RADIO (Nov. 4, 2025) (last accessed November 16, 2025) (https://www.npr.org/2025/11/04/nx-s1-5587742/election-results-california-proposition-50-redistricting).

289.   Supra at 21.

290.   142 S. Ct. 879, 880–81 (2022)

291.   Id. (Kavanaugh, J., concurring)

292.   See id. at 881 (“Late judicial tinkering with election laws can lead to disruption and to unanticipated and unfair consequences for candidates, political parties, and voters, among others.”).

293.   See id. (“It is one thing for a State on its own to toy with its election laws close to a State's elections. But it is quite another thing for a federal court to swoop in and re-do a State's election laws in the period close to an election.”).

294.   Brown Op. at 154.

295.   See Milligan, 142 S. Ct. at 880 (Kavanaugh, J., concurring) (“It is one thing for a State on its own to toy with its election laws close to a State's elections. But it is quite another thing for a federal court to swoop in and re-do a State's election laws in the period close to an election.”).

296.   Brown Op. at 154.

297.   See Democratic Nat'l Comm. v. Wis. State Legislature, 141 S. Ct. 28, 31 (2020) (Kavanaugh, J., concurring) (“It is one thing for state legislatures to alter their own election rules in the late innings and to bear the responsibility for any unintended consequences. It is quite another thing for a federal district court to swoop in and alter carefully considered and democratically enacted state election rules when an election is imminent.”).

298.   See id. (“That important principle of judicial restraint not only prevents voter confusion but also prevents election administrator confusion—and thereby protects the State's interest in running an orderly, efficient election and in giving citizens (including the losing candidates and their supporters) confidence in the fairness of the election.”).

299.   See Tex. Gov't Code § 312.007.

300.   In re Landry, 83 F.4th 300, 303 (5th Cir. 2023) (quoting Wise v. Lipscomb, 437 U.S. 535, 540 (1978); see also Landry, 83 F.4th at 303 (stating that the above “is the law today as it was forty-five years ago.”).

301.   See id. at 303 n.2 (providing myriad Supreme Court citations for the primacy of the legislature in redistricting).

302.   Id. (collecting cases).

303.   At the close of the preliminary-injunction trial, the State explicitly invoked its right to redraw the map should this court decide to grant relief.

304.   Brown Op. at 159.

305.   See 2022 State Legislative Session Calendar, National Conference of State Legislatures, https://www.ncsl.org/ about-state-legislatures/2022-state-legislative-session-calendar (last visited Nov. 13, 2025) (stating that the Alabama regular session convened on January 11th and Adjourned on April 7th).

306.   Tex. Elec. Code § 172.023 (“An application for a place on the general primary election ballot must be filed not later than 6 p.m. on the second Monday in December of an odd-numbered year unless the filing deadline is extended under Subchapter C.”); see also Tex. Elec. Code § 172.054 (allowing the filing deadline to be extended only due to death, withdrawal of an incumbent, or incapacity).

307.   See Merrill v. Milligan, 142 S. Ct. 879, 881 n.1 (Kavanaugh, J., concurring).

308.   Tr. 10/8/25 AM 151:18–24.

309.   Tr. 10/8/25 AM 146:18–147:2.

310.   Tr. 10/8/25 AM 155:11–17.

311.   Tr. 10/8/25 AM 155:25–156:4.

312.   Tr. 10/8/25 PM 6:18–24.

313.   Tr. 10/8/25 PM 9:2–15.

314.   Tr. 10/8/25 PM 9:16–25.

315.   Tr. 10/8/25 PM 9:1–6; see also 52 U.S.C. § 20302(a)(8).

316.   52 U.S.C. § 20302(g).

317.   Tr. 10/8/25 PM 14:16–19.

318.   Tr. 10/8/25 PM 10:16–19.

319.   Tr. 10/8/25 AM 149:19–150:5.

320.   Tr. 10/8/25 PM 8:14–23.

321.   John C. Moritz, Texas candidate filings open with a big question: Will Republicans' new map stick?, Houston Chronicle, Nov. 10, 2025, https://www.houstonchronicle.com/news/politics/state/article/redistricting-candidate-primaries-21151780.php. See also id. (describing a series of candidates who may not run or who may run elsewhere due to the alteration of the map by the court).

322.   Brown Op. at 151.

323.   Tr. 10/8/25 AM 153:13–18.

324.   See Milligan, 142 S. Ct. at 880 (Kavanaugh, J., concurring).

325.   Id.

326.   See id. at 881.

327.   Brown Op. at 2.

328.   Brown Op. at 16.

329.   Brown Op. at 16-17.

330.   Brown Op. at 55.

331.   Brown Op. at 60.

332.   Brown Op. at 79.

333.   Brown Op. at 98.

334.   Brown Op. at 122.

335.   Brown Op. at 126.

336.   Brown Op. at 146.

337.   Brown Op. at 151.

JEFFREY V. BROWN, United States District Judge:1 FN1. U.S. District Judge Jeffrey V. Brown delivers the opinion of the Court, which Senior U.S. District Judge David C. Guaderrama joins. U.S. Circuit Judge Jerry E. Smith will file a dissenting opinion.