delivered the opinion of the Court,

This matter is yet another chapter in the 2000 congressional redistricting controversy. Governor Rick Perry and former Secretary of State Henry Cuellar (the State defendants), Susan Weddington, several congressional members, the Associated Republicans of Texas, and Charles Babb directly appeal the final judgment from a bench trial held in Travis County, Texas. We determine two issues:

• Does the Attorney General, under the separation-of-powers doctrine, have the authority to require the trial court to adopt his redistricting plan and render judgment that his plan is the baseline state court plan for the federal court redistricting proceedings?

• Did the trial court violate the parties’ due course of law rights when it rendered a judgment based on a party’s new plans not in evidence at trial without giving the parties an opportunity for a meaningful hearing?

We conclude that the Attorney General’s assertion that he speaks for the Legislature and thus the trial court must adopt his plan violates the separation-of-powers doctrine. Moreover, under the facts here, we conclude that the manner in which the trial court rendered its judgment violated the parties’ due course of law rights. Accordingly, we vacate the trial court’s October 10, 2001 judgment and remand the case to the trial court for proceedings consistent with this opinion.

I. BACKGROUND

On September 10, 2001, this Court determined that the Travis County trial court had dominant jurisdiction to hear the plaintiffs’ claims that, under the 2000 census, the existing Texas congressional districts are unconstitutional and that the trial court must adopt a new redistricting plan. Perry et al. v. Del Rio et al., 66 S.W.3d 239. Following our decision, the Travis County trial court set the case for trial beginning September 17, 2001. The trial court received evidence and heard *88testimony and arguments from all the parties about the various proposed congressional redistrieting plans they urged the trial court to adopt. On September 28, 2001, the parties rested, closed, and presented arguments to the trial court.

On October 1, 2001, the trial court notified the parties that it was, on its own motion, appointing the Texas Legislative Council to act as the trial court’s expert in deciding the issues in this case. The trial court issued an order to this effect and required the Texas Legislative Council’s staff to maintain as confidential the trial court’s dealings and communications with the staff. The State defendants objected to this order. They argued that not only was the trial court’s order impermissible under the Texas Rules of Civil Procedure, but it also created a conflict of interest because the Texas Legislative Council’s staff are the Lieutenant Governor’s and Speaker of the House’s employees.

Two days later, on October 3, 2001, the trial court entered an order announcing its intent to adopt Plan 1065C for Texas’ congressional districts. The trial court attached Plan 1065C to the order and identified an Internet site where the parties could view the plan. Further, the trial court invited all parties to file comments, proposed changes, or requested modifications to this plan by October 9, 2001. The trial court also stated that it was preparing findings of fact and conclusions of law.

On October 9, 2001, the Democratic Congressional Interveners and the Del Rio and Cotera Plaintiffs filed objections to Plan 1065C. On that same day, Speaker Laney submitted proposed modifications to Plan 1065C and requested that the trial court incorporate his proposed new plans, different from those he offered during the trial, into Plan 1065C.

Sometime after 10:00 a.m. on October 10, 2001, the trial court notified the parties by facsimile that it was “seriously considering” adopting several changes Speaker Laney proposed. The trial court briefly explained the changes it was considering making to Plan 1065C and asked the parties to submit comments on these proposed changes by noon that day. Unlike its previous order identifying Plan 1065C as the proposed plan, the trial court did not attach a map showing what the new proposed plan looked like. Also, it did not refer the parties to an Internet site where they could view the plan.

Later that day, the trial court rendered its final judgment. And, rather than adopting Plan 1065C, the trial court adopted a new plan designated Plan 1089C. The trial court’s judgment states that Plan 1089C incorporates certain proposals the parties submitted. Moreover, the trial court’s judgment permanently enjoins further use of the existing thirty Texas congressional districts in any primary or general election.

Throughout the state-court proceedings, litigation has been pending in the United States District Court for the Eastern District of Texas (the “Tyler court”) before a three-judge panel. But the Tyler court has deferred to the state court litigation, as Growe v. Emison, 507 U.S. 25,113 S.Ct. 1075, 122 L.Ed.2d 388 (1993), requires. It previously rescheduled its redistricting proceedings to begin on October 1, 2001, and ordered the parties to file any state court plan — which the Tyler court would use as a baseline for its redistricting trial — by that date. At the Travis County trial court’s request, on October 1, the Tyler court extended the filing deadline until October 3. The Tyler court also ordered the parties to file expert reports and proposed exhibits and to submit statements of position by October 11 and 12, respectively.

*89However, after the trial court rendered its October 10 judgment that adopted a redistricting plan different from Plan 1065C, the Tyler court again extended its trial-schedule deadlines. In its October 11 order, the Tyler court recognized that when it set the October 11 and 12 deadlines, it had “not contemplated major changes to the state court plan [Plan 1065C] filed on October 3.” But it concluded that Growe compels it to use Plan 1089C, rather than Plan 1065C, as a baseline state plan. Thus, the Tyler court ordered that it would continue its trial until October 22, 2001, and extended the deadlines for the parties to file their position statements and exhibits.

On October 12, 2001, the State defendants, whom the Attorney General has represented throughout the proceedings, perfected their direct appeal with this Court. On October 15, 2001, the Associated Republicans of Texas and Charles Babb perfected their appeal. On that day, after considering the parties’ jurisdictional statements and objections, we noted probable jurisdiction under Rule 57 of the Texas Rules of Appellate Procedure and ordered the parties to file briefs on an expedited schedule. Then, on October 16, 2001, Susan Weddington, Chair of the Republican Party of Texas, and Congressmen Tom DeLay, Joe Barton, John Cul-berson, Sam Johnson, and Kevin Brady (“Congressman DeLay” collectively) perfected their direct appeal and moved to consolidate their appeal with the State defendants’ appeal. We noted probable jurisdiction and granted the motion to consolidate. The Court heard oral arguments on October 18, 2001.

II. JURISDICTION

Our State’s existing congressional redistricting plan is embodied in Article 197h of the Texas Revised Civil Statutes. In 1996, a federal district court held that three congressional districts in Article 197h were unconstitutional. Vera v. Bush, 933 F.Supp. 1341, 1347 (S.D.Tex. 1996). The Vera court entered an interim remedial order that redrew the three districts to correct the constitutional infirmities. 933 F.Supp. at 1352. Because the Legislature never enacted a new plan, the federal court’s remedial order and the unaffected districts in Article 197h remained in effect for future elections. See Vera v. Bush, 980 F.Supp. 251, 253 (S.D.Tex. 1997). However, the 2000 census demonstrates that Texas is now entitled to two additional congressional delegates. Therefore, Texas’ existing plan with thirty congressional districts is presumptively unconstitutional. See U.S. Const, amend. XIV, § 2.

This Court has direct-appeal jurisdiction from “an order of a trial court granting or denying an interlocutory or permanent injunction on the ground of the constitutionality of a statute of this state.” Tex. Gov’t.Code § 22.001(c); see Tex. Const, art. V, § 3-b. Here, the trial court’s order enjoins parties from using the State’s existing thirty congressional districts — which Article 197h and the Vera court’s 1996 remedial order reflect — in any election. Additionally, when this Court has appellate jurisdiction over any issue, it acquires “extended jurisdiction” over all other questions of law properly preserved and presented. Edgewood Indep. Sch. Dist. v. Meno, 917 S.W.2d 717, 749 n. 39 (Tex. 1995); City of Corpus Christi v. Pub. Util. Comm’n, 572 S.W.2d 290, 294 (Tex. 1978). Accordingly, we have direct-appeal jurisdiction to consider all the legal errors alleged in the various parties’ appeals. See Edgewood, 917 S.W.2d at 749 n. 39; City of Corpus Christy 572 S.W.2d at 294.

III. THE PARTIES’ CONTENTIONS

A. The State Defendants

Because the trial court failed to adopt the Attorney General’s proposed redis*90tricting plan, the State defendants contend the trial court violated our separation-of-powers doctrine. See Tex. Const, art. II, § 1. The State defendants assert that, absent a legislative redistricting plan, the Attorney General has the inherent power to fill the legislative void and present the State’s policy preferences on redistricting. The State defendants argue the trial court should have deferred to the plan the Attorney General submitted on the State defendants’ behalf, because Texas’ Attorney General is authorized to represent the State’s interests in redistricting litigation. According to the State defendants, the trial court abandoned its adjudicatory role when it adopted its own plan without finding any legal errors in the Attorney General’s or other parties’ proposed plans.

Furthermore, the State defendants argue that the trial court’s judgment adopting Plan 1089C at the eleventh hour, without providing the parties an adequate opportunity to comment and to submit evidence about this new plan, rendered the bench trial meaningless. The State defendants urge that the trial court’s order appointing the Texas Legislative Council as an expert and requiring the Council’s staff to keep its dealings with the trial court confidential further exemplifies the trial court’s irregular proceedings.

B. Weddington, CONGRESSMAN DeLay, ART, and Babb

Weddington, Congressman DeLay, ART, and Babb argue that the trial court violated their due course of law rights by adopting Plan 1089C in its final judgment without affording the parties notice and an opportunity to comment on that plan, review its statistics, or provide it to their experts for study. Relying on this Court’s decision in Terrazas v. Ramirez, 829 S.W.2d 712 (Tex. 1991), they contend that a trial court cannot adopt a redistricting plan without hearing evidence on the plan and affording interested parties a forum in which to test the plan.

Moreover, Weddington, Congressman DeLay, ART, and Babb assert that there is no evidence to support the trial court’s findings of facts and conclusions of law about Plan 1089C. They argue that in contrast to the more than nine redistricting plans the parties proposed during the bench trial — along with expert testimony to support and criticize each proposed plan — Plan 1089C was never introduced or subjected to expert scrutiny at trial.

Finally, Weddington, Congressman DeLay, ART, and Babb contend that Plan 1089C is motivated by improper criteria because it is an incumbent protection plan. Moreover, they argue that Plan 1089C violates the Voting Rights Act by not protecting minority voters.

C. Speaker Laney and Del Rio and Cotera, et al.

Speaker Laney and the Del Rio and Cotera Plaintiffs contend that all the appellants raise only factually based arguments that this Court does not have jurisdiction to consider. Specifically, they contend appellants’ arguments that Plan 1089C violates the Voting Rights Act and that the Court should adopt the plans they supported at trial require factual determinations.

Furthermore, Speaker Laney and the Plaintiffs assert that the Attorney General does not have authority to unilaterally impose a congressional redistricting plan on this State. They argue that Terrazas stands only for the proposition that the Attorney General can suggest possible redistricting remedies just as any other party can do at trial. Also, they urge that the separation-of-powers doctrine precludes extending executive branch authority to congressional redistricting. And Speaker *91Laney and the Plaintiffs refute the Attorney General’s claim that federal jurisprudence supports his position that the trial court should treat executive officials as if they speak for the Legislature.

Finally, Speaker Laney and the Plaintiffs argue that Plan 1089C was the product of a fair process. He contends that some of his post-trial proposed modifications to Plan 1065C, which the trial court adopted in Plan 1089C, answer the trial court’s question at trial about whether Speaker Laney and Lieutenant Governor Ratliff had merged their two maps together. They had not. But Speaker Laney urges that his modifications reflect “a careful merging” of the plans he and the Lieutenant Governor proposed. According to Speaker Laney, the need for an additional hearing is unfounded because the parties presented evidence to the court on an equal footing and had the opportunity to comment on the trial court’s initial proposal. Moreover, he contends, the trial court ultimately should be able to select its own remedial plan.

IY. APPLICABLE LAW

A. STANDARD OF REVIEW

Our review on direct appeal is constitutionally confined to questions of law. Tex Const, art. V, § 3-b; Tex Gov’t Code § 22.001(b); Tex.R.App. P. 57.2; O’Quinn v. State Bar of Texas, 763 S.W.2d 397, 399 (Tex. 1988). We review questions raising constitutional violations de novo. See City of College Station v. Turtle Rock Corp., 680 S.W.2d 802, 804 (Tex. 1984); State/Operating Contractors ABS Emissions, Inc. v. Operating Contractors/State, 985 S.W.2d 646, 650-51 (Tex.App.—Austin 1999, pet. denied).

B. Separation of Powers

The Texas Constitution’s separation-of-powers doctrine provides:

The powers of the Government of the State of Texas shall be divided into three distinct departments, each of which shall be confided to a separate body of magistracy, to wit: Those which are Legislative to one; those which are Executive to another, and those which are Judicial to another; and no person, or collection of persons, being of one of these departments, shall exercise any power properly attached to either of the others, except in the instances herein expressly permitted.

Tex Const, art. II, § 1.

The Legislature is the department constitutionally responsible for apportioning the State into federal congressional legislative districts. U.S. Const, amend. XIV, § 2; Tex. Const, art. III, § 28; Wise v. Lipscomb, 437 U.S. 535, 539, 98 S.Ct. 2493, 57 L.Ed.2d 411 (1978). When the Legislature does not act, citizens may sue and, then, it is the judiciary’s role to determine the appropriate redistricting plan. See Growe, 507 U.S. at 33-34, 113 S.Ct. 1075; Scott v. Germano, 381 U.S. 407, 409, 85 S.Ct. 1525, 14 L.Ed.2d 477 (1965) (per curiam); Maryland Comm. v. Tawes, 377 U.S. 656, 676, 84 S.Ct. 1429, 12 L.Ed.2d 595 (1964); Terrazas, 829 S.W.2d at 717-20.

Because of our Constitution’s explicit prohibition against one government branch exercising a power attached to another, unless specific circumstances exist, “it is only by express constitutional provision that the executive department could legitimately exercise the redistricting power.” Terrazas, 829 S.W.2d at 733 (Cornyn J., concurring) (emphasis in original). This is because “[t]he powers conferred by the Constitution upon the state officials are generally held to be exclusive, and except in the manner authorized by the Constitution, these powers cannot be en*92larged or restricted.” Garcia v. Laughlin, 155 Tex. 261, 285 S.W.2d 191, 194 (1955).

The Attorney General is a member of the Executive Department whose primary duties are to render legal advice in opinions to various political agencies and to represent the State in civil litigation. See Tex. Const, art. IV, §§ 1, 22; Tex. Gov’t Code § 402.021. We have recognized that the Attorney General, as the State’s chief legal officer, has broad discretionary power in carrying out his responsibility to represent the State. Terrazas, 829 S.W.2d at 722. But we have held that this power does not permit the Attorney General “to effectuate a valid reapportionment of senatorial districts himself,” because only the trial court’s judgment can accomplish this when the Legislature fails to act. Terrazas, 829 S.W.2d at 722. This is because the Attorney General can only act within the limits of the Texas Constitution and statutes, and courts cannot enlarge the Attorney General’s powers. Terrazas, 829 S.W.2d at 735 (Cornyn, J., concurring); State ex rel. Downs v. Harney, 164 S.W.2d 55, 59 (Tex.Civ.App.—San Antonio 1942, writ ref'd w.o.m.).

C. Due Course of Law

The Texas Constitution provides that “[n]o citizen of this State shall be deprived of life, liberty, property, privileges or immunities, or in any manner disfranchised, except by the due course of the law of the land.” Tex. Const, art. I, § 19. Typically, what course of law is due depends on several factors, including the private interests affected, the risk that the procedures used may erroneously deprive an interest, and the government’s interest, such as the burden that the procedural requirement would entail. Univ. of Tex. Med. School v. Than, 901 S.W.2d 926, 930 (Tex. 1995). We have recognized that our due course of law provision at a minimum requires notice and an opportunity to be heard at a meaningful time and in a meaningful manner. Than, 901 S.W.2d at 930; see House of Tobacco, Inc. v. Calvert, 394 S.W.2d 654, 657-58 (Tex. 1965); Freeman v. Ortiz, 106 Tex. 1, 153 S.W. 304, 304 (1913). And, under certain circumstances, the right to be heard assures a full hearing before a court having jurisdiction over the matter, the right to introduce evidence at a meaningful time and in a meaningful manner, and the right to judicial findings based upon that evidence. See Turcotte v. Trevino, 499 S.W.2d 705, 723 (Tex.Civ. App. — Corpus Christi 1973, writ refd n.r.e.). This right also includes an opportunity to cross-examine witnesses, to produce witnesses, and to be heard on questions of law. In re B— M— N— , 570 S.W.2d 493, 502 (Tex.Civ.App.—Texarkana 1978, no writ). It also involves the right to have judgment rendered only after trial. Grigsby v. Peak, 57 Tex. 142, 144 (1882); Masonic Grand Ch. of Order of E. Star v. Sweatt, 329 S.W.2d 334, 337 (Tex.Civ.App.— Fort Worth 1959, writ ref'd n.r.e.).

This Court has held in redistricting cases that Texas courts can order reapportionment “only after investigation and careful consideration of the many, diverse interests affected.... ” Terrazas, 829 S.W.2d at 718. We have also instructed that in redistricting cases Texas courts must attempt to consider the parties’ interests, as well as the public’s interest in general. Terrazas, 829 S.W.2d at 719.

V. ANALYSIS

A. Separation of Powers

We disagree with the State defendants’ argument that the trial court violated our separation-of-powers doctrine when it did not defer to, or adopt, the redistricting plan the Attorney General proposed during the trial. To the contrary, it is the *93Attorney General’s position — that in congressional redistricting controversies he steps into the Legislature’s shoes if the Legislature does not act — that violates the separation-of-powers doctrine.

The State defendants’ contentions about the Attorney General’s alleged role in this case have no basis in law. While congressional redistricting is typically a legislative function, U.S. Const. amend. XIV, § 2, courts must resolve redistricting controversies when the legislature does not do so. See, e.g., Vera, 980 F.Supp. at 252; Terrazas, 829 S.W.2d at 717. In deciding a redistricting controversy, we have stated that “Texas courts may order apportionment ... [but] that power ought to be used only after investigation and careful consideration of the many, diverse interests affected ..." Terrazas, 829 S.W.2d at 718 (emphasis added). Moreover, “the trial court must attempt to consider the interests, not only of the parties in the case, but of others who are not present.” Terrazas, 829 S.W.2d at 719. Therefore, requiring that the trial court defer to, and adopt, the Attorney General’s plan thwarts our explicit guidelines for trial courts in redistricting cases.

To accept the State defendants’ position that the Attorney General becomes the Legislature’s voice when the Legislature fails to act would condone a constitutional violation. See Tex. Const, art. II, § 1 (no branch of government “shall exercise any power properly attached to either of the others, except in the instances herein [the Constitution] expressly permitted.”). As a member of the executive branch, the Attorney General may not perform legislative functions unless expressly authorized to do so. See Tex. Const, art. II, § 1; Garcia, 285 S.W.2d at 194-95; Terrazas, 829 S.W.2d at 738 (Cornyn, J., concurring). Neither our Constitution nor Chapter 402 of the Government Code expressly authorizes the Attorney General’s position here. See Tex. Const, art. IV, §§ 1, 22; Tex. Gov’t Code § 402.021. And the State defendants provide us with no other authority giving the Attorney General the Legislature’s power to resolve the congressional redistricting controversy. As we stated in Terrazas, under these circumstances, only courts have the authority to effectuate a valid congressional reapportionment plan unless or until the Legislature acts. 829 S.W.2d at 720.

Accordingly, we agree with Speaker Laney’s and the Plaintiffs’ contentions that the Attorney General does not have the authority to act in the Legislature’s stead and dictate the remedy in a congressional redistricting case. Indeed, Terrazas and Lawyer v. Department of Justice, 521 U.S. 567, 577-78, 117 S.Ct. 2186, 138 L.Ed.2d 669 (1997), demonstrate only the Attorney General’s authority to propose and suggest remedies and settle redistricting cases. But they do not provide for the unilateral legislative authority the Attorney General now seeks. Because neither the Constitution nor a statute expressly gives the Attorney General legislative authority in redistricting cases, his claim that it exists is without merit.

B. Due Course of Law

We agree with Weddington, ART, and Babb that the manner in which the trial court arrived at its October 10 final judgment violated our Constitution’s due course of law provision. For all we know, the redistricting plan could very well dictate our State’s congressional elections for the next decade. See Vera, 980 F.Supp. at 253. Court-ordered redistricting cases require close scrutiny of several important factors and interests, including: compactness, regularity, contiguity, preservation of communities of interest, equal protection, and the integrity of natural and traditional *94county and city boundaries. See generally Bush v. Vera, 517 U.S. 952, 116 S.Ct. 1941, 135 L.Ed.2d 248 (1996). Therefore, it is imperative in these cases that a court’s procedures protect all interests involved.

Indeed, we have recognized that court-ordered redistricting could affect not only the parties to the litigation but also others in our State who should have the opportunity to intervene and be heard. Terrazas, 829 S.W.2d at 726. Also, we have recognized that courts in these cases must take careful consideration of the many, diverse interests affected. Terrazas, 829 S.W.2d at 718.

Undoubtedly, redistricting will impact nearly all Texas voters, because it will determine who they choose to represent their interests in Congress. Than, 901 S.W.2d at 930. Furthermore, there is a significant risk that all Texas citizens’ interests may not be protected in redistricting litigation. See Than, 901 S.W.2d at 930. This is why a court cannot order a reapportionment plan for the State based upon nothing more than an agreement between the Governor, the Attorney General, and a few citizens. See Terrazas, 829 S.W.2d at 714. And this is why we have held that, in redistricting cases, Texas courts may order apportionment only after investigation and careful consideration of all diverse interests affected. See Terrazas, 829 S.W.2d at 718.

The unique circumstances and constitutionally protected interests involved in court-ordered redistricting cases require that, before entering its final judgment on a redistricting plan, a trial court must afford all the parties a meaningful hearing. See Than, 901 S.W.2d at 930; Freeman, 153 S.W. at 304; Grigsby, 57 Tex. at 144; In re B—M—N—, 570 S.W.2d at 502; Sweatt, 329 S.W.2d at 337. This does not unduly burden our courts; it is simply what our Constitution and the state-wide interests a court-ordered redistricting ease affects require. See Than, 901 S.W.2d at 930; Terrazas, 829 S.W.2d at 718.

Here, the manner in which the trial court entered its final judgment did not comport with Terrazas and violated the parties’ due course of law rights. On October 3, 2001, the trial court advised the parties that it intended to adopt Plan 1065C and gave the parties a meaningful period of time to comment on this plan. To this point, the trial court provided an adequate procedure by which it could give full and careful consideration to the interests its ruling may affect. Terrazas, 829 S.W.2d at 720.

In its October 10 mid-morning facsimile to the parties, the trial court set a noon deadline for any party to comment on Plan 1089C, a plan significantly different than the plan the trial court originally proposed to adopt. The trial court’s facsimile made clear that it would not entertain any comments received after noon because the Tyler federal court’s filing deadline was 5:00 p.m. Thus, the parties not only had little time to object to the new changes, they were deprived of a meaningful opportunity to present a motion for new trial. Once the trial court determined that it intended to substantially change its proposed redistricting plan, the constitutionally-protected interests involved, Termzas, and our Constitution’s due course of law provision required the trial court to provide the parties a meaningful opportunity to be heard. See Than, 901 S.W.2d at 930; Freeman, 153 S.W. at 304; Grigsby, 57 Tex. at 144; In re B—M—N—, 570 S.W.2d at 502; Sweatt, 329 S.W.2d at 337.

We do not mean to suggest that a trial court cannot make de minimis changes to its proposed redistricting plan without reopening the evidence. The proceedings *95must end at some point. But, here, the trial court’s changes to Plan 1065C were extensive and significant — not de minimis. As the Tyler court noted in its October 11, 2001 order, the trial court made “major changes” to its initially proposed plan. Because the trial court did not provide the necessary due course of law when it made these wholesale changes, its October 10, 2001, judgment violated the parties’ due course of law rights and therefore is invalid.

Because of the procedural infirmities in the way the trial court rendered its final judgment, that judgment, which adopts Plan 1089C, is wholly invalid. Seven of the nine Justices of this Court agree that Plan 1089C is invalid, while two do not reach the issue and would dismiss on jurisdictional grounds. However, our paths diverge on the proper disposition to follow. Justice Hecht would render judgment that Plan 1065C is the state court baseline plan for the Tyler court to use. But the contortions Justice Hecht must go through to reach that result are beyond the scope of our judicial power. First, by validating Plan 1065C, Justice Hecht would breathe life into a plan that the trial court never finally adopted and therefore does not exist. Second, Justice Hecht must wholly disregard Speaker Laney’s and the Plaintiffs’ objections and proposed changes to Plan 1065C. But to disregard them, Justice Hecht must make factual determinations about the evidence and judgment calls about what the trial court’s findings might have been. This is entirely inappropriate, as our jurisdiction on direct appeal is constitutionally limited to questions of law. Tex. Const, art. V, § 3—b; Tex. Gov’t Code § 22.001(b); Tex.R.App. P. 57.2; O’Quinn, 763 S.W.2d at 399.

Additionally, although Justice Owen agrees that we must remand, she writes only to instruct the trial court how to conduct its proceedings. But her writing does nothing more than instruct the trial court on how to follow law that already exists. Moreover, it is entirely advisory, and our Constitution prohibits courts from issuing advisory opinions. Tex. Const, art. II, § 1; Texas Ass’n of Bus. v. Texas Air Control Bd., 852 S.W.2d 440, 444 (Tex. 1993).

VI. CONCLUSION

We conclude that the manner in which the trial court arrived at its judgment violated the parties’ due course of law rights. Consequently, we vacate the trial court’s October 10, 2001 judgment and remand the case to the trial court for proceedings consistent with this opinion.

Justice RODRIGUEZ issued a concurring opinion, in which Justice ENOCH and Justice JEFFERSON joined.

concurring.

I join in the Court’s decision to remand this case and its analysis of the due process issue and the scope of the Attorney General’s authority. But this Court should give guidance and instructions to the trial court on the substantive law that should govern the selection of a plan on remand if the case is to proceed. The modifications to Plan 1065C that led to Plan 1089C were based in part, if not in whole, on political considerations, including incumbent protection. I would remand *96this ease with directions to the trial court that it cannot consider political factors.

I share Justice Hecht’s frustration that these proceedings have not been more expeditious. However, I cannot assume with certainty that the three-judge federal court will conclude that the State of Texas has failed to devise a districting plan, although the federal court would be justified in coming to that conclusion. My greatest concern is that on remand of this case, the trial court will adopt a plan plagued by the same infirmities that obtained in Plan 1089C. Accordingly, I would give guidance to the trial court to preclude that possibility.

I

When courts must make redistricting decisions after a state legislature has failed to adopt a constitutional plan, those courts generally are “faced with hard remedial problems in minimizing friction between their remedies and legitimate state policies.” 1 I submit that there are few, if any, state policies that are of concern in fashioning a plan in this case. No congressional district in Texas can pass constitutional muster in the wake of the population growth and shifts that have occurred since the statute governing congressional districts was enacted in 1991. The plan, statewide, is unconstitutional. And, as de-tañed in Justice Hecht’s dissent, other significant changes have occurred in the political make-up of this State since its now unconstitutional districting plan was adopted in 1991. The failure of the Legislature to make any serious attempt in the 2001 session to adopt a new plan based on the 2000 decennial census, and its repeated failure to adopt a new plan from and after significant parts of the existing plan were declared unconstitutional in 1996, indicate the absence of any political will and the absence of any consensus on political policy on the part of the Texas Legislature in the districting context.

Even if there were discernible state apportionment policies, the United States Supreme Court has held that while it is well within the province of legislatures to formulate state policies that take political considerations into account in devising a districting plan,2 “courts by contrast possess no distinctive mandate to compromise sometimes conflicting state apportionment policies in the people’s name.”3 In this same vein, a three-judge federal court has held that “ ‘[m]any factors, such as the protection of incumbents, that are appropriate in the legislative development of an apportionment plan have no place in a plan formulated by the courts.’ ” 4 The United States Court of Appeals for the Fifth Circuit has likewise held that protection of incumbents has “no place in a plan formulated by the courts.”5

As Justice Hecht’s dissent explains, most of the reasons offered by the proponent of what came to be Plan 1089C for *97changes to Plan 1065C were for incumbency protection. I agree with Justice Hecht’s dissent that the euphemisms used were simply another way of articulating “incumbent protection.” The proponent of Plan 1089C offered only political considerations in urging the trial court to adopt that plan. No record was developed on Plan 1089C, and we cannot, therefore, determine if there was any legitimate basis for the trial court to fashion a plan that otherwise appears to place improper factors ahead of one-man-one vote considerations, compactness, respecting political subdivision boundaries, and compliance with the Voting Rights Act.

The only other reason for modifications to Plan 1065C identified by the proponent of what came to be Plan 1089C was that particular areas should retain their rural or suburban “character” or similarly, to “restore communities of interest.” These, too, are political considerations that have no bearing on compactness, the boundaries of political subdivisions, or compliance with the Voting Rights Act.

When courts must step in and fashion a redistricting plan, their task is a “sensitive one that must be ... free from any taint of arbitrariness or discrimination.”6 A plan that draws on political factors such as “communities of interest” is tainted.

II

Assuming that the federal court concludes that it will consider the new plan from the state trial court following remand, which the federal court is certainly not obliged to do,7 then there may be a window within which the state court system can devise a legitimate redistricting plan. There is a plan in existence that will not require additional evidence or an extensive hearing on remand, and there were no objections to that plan on the basis that it protected incumbents. That plan is the one the trial court originally announced that it intended to adopt. If there is indeed an opportunity for the state court system to avoid default, it would seem that this plan, with little or no modification, is the State’s best hope for arriving upon a plan that meets the criteria that courts should consider, which are adherence to the one-man-one-vote requirement, compactness, respect for political subdivisions, and compliance with the Voting Rights Act.

For the foregoing reasons, I concur in the Court’s judgment.

joined by Justice HANKINSON, dissenting.

Our Court’s direct appeal jurisdiction is limited to appeals “from an order.of a trial court granting or denying an interlocutory or permanent injunction on the ground of the constitutionality of a statute of this state.” Tex. Gov’t Code § 22.001(c). Although the final judgment in the court below included a permanent injunction, it only enjoined the “use of the existing 30 congressional districts in Texas as reflected in Plan 1000C1 in any primary or general election.”2 No party before this Court is complaining about that part of the judgment. The only dispute in the trial court or here concerns how to divide the state into the thirty-two districts that will each elect a member of the next Congress. Thus, no party in this Court is challenging or defending the constitutionality of a statute, much less challenging or defending the grant or denial of an injunction based on the constitutionality of a statute. There is therefore no appeal “from an order ... granting or denying an ... injunction on the ground of the constitutionality of a statute.” On the plain language of section 22.001(c), I would withdraw our note of probable jurisdiction and dismiss the appeal for want of jurisdiction.

The right to direct appeal did not exist until 1940, when the people of Texas adopted section 3 b of Article V of the Texas Constitution to permit the Legislature to provide for such a remedy.3 Specific constitutional authorization was necessary because Section 3 of Article V of the Constitution then limited the Supreme Court’s appellate jurisdiction to certain cases that had been decided by the courts of civil appeals.4 Pursuant to that amendment, the Legislature provided for direct appeals beginning Jan. 1, 1944. Tex. Gov’t Code § 22.001(c). As that statute directed, the Supreme Court promulgated Texas Rule of Civil Procedure 499a, effective Dec. 31, 1943, to “prescribe the necessary rule of procedure to be followed in perfecting the appeal.” Tex. Gov’t Code § 22.001(c). That rule, which remained essentially unchanged until the Court adopted the Texas Rules of Appellate Procedure in 1986, is somewhat more detailed than its successor rules, Texas Rule of Appellate Procedure 140 (eff. Sept. 1, 1986 to Aug. 31, 1997) and current Texas Rule of Appellate Procedure 57 (eff.Sept.l, 1997). In part, the original rule provided:

*99An appeal to the Supreme Court directly from such a trial court may present only the constitutionality or unconstitutionality of a statute ... or the validity or invalidity of an administrative order ... when the same shall have arisen by reason of the order of a trial court granting or denying an interlocutory or permanent injunction.

Tex.R. Civ. P. 499a(b)(eff.Dec.31,1943)(em-phasis added). While today’s rule has been shortened to address only purely procedural matters, the Court’s contemporaneous explanation of the scope of the remedy is still instructive.

In Bryson v. High Plains Underground Water Conservation Dist. No.1, 156 Tex. 405, 297 S.W.2d 117 (1956), Bryson had attempted to directly appeal a permanent injunction barring him from producing more than 100,000 gallons of water a day from his well without a permit from the water conservation district. During the trial, he had unsuccessfully attacked the constitutionality of portions of the statute that created the district. This Court ruled that it did not have direct appeal jurisdiction simply because a statute was challenged on constitutional grounds in the trial court if that question was not presented on direct appeal. In interpreting the jurisdictional statute, the Court said:

For us to have jurisdiction of a direct appeal, it must appear that a question of the constitutionality of a Texas statute ... was properly raised in the trial court, that such question was determined by the order of such court granting or denying an interlocutory or permanent injunction, and that the question is presented to this Court for decision.

Id. at 119 (emphasis added). None of the appeals to this Court in this case meet this standard.

Here, all parties concede that the old congressional district lines are unconstitutional. While the state’s population has grown by nearly twenty-three percent in the last decade, entitling Texas to two new seats in Congress, sixty-eight of the state’s 254 counties actually lost population. New districts are indisputably necessary, and the Legislature’s failure to draw those districts left the decision with the courts. If the Seventy-Seventh Legislature had passed a congressional redistricting act, any subsequent legal challenges would have involved parties objecting to the act as plaintiffs and parties supporting the act as defendants. If the trial court had enjoined the Legislature’s plan on the basis of unconstitutionality, the plan’s defenders would have had grounds for a direct appeal. Conversely, if the trial court had accepted the Legislature’s plan against a constitutional attack and denied an injunction on that basis, the plan’s objectors could have brought a direct appeal. When the Legislature has acted on redistricting, this Court can hear direct appeals on issues of law if the trial judge has granted or denied injunctive relief on constitutional grounds regarding that action. See, e.g., Richards v, Mena, 820 S.W.2d 371 (Tex. 1991); Upham v. White, 639 S.W.2d 301 (Tex. 1982); Smith v. Craddick, 471 S.W.2d 375 (Tex. 1971). But because the Legislature did not act after the 2000 Census, even if the trial court had issued a mandatory injunction in support of its plan, it would still not have been an “injunction on the ground of the constitutionality of a statute ” (emphasis added).

We have always strictly construed our direct appeal jurisdiction. See, e.g., Texas Workers’ Compensation Comm’n v. Garcia, 817 S.W.2d 60, 61 (Tex. 1991); Mitchell v. Purolator Security, Inc., 515 S.W.2d 101 (Tex. 1974); Gardner v. Railroad Comm’n, 160 Tex. 467, 333 S.W.2d 585 (1960). For example, in Boston v. Garrison, 152 Tex. 253, 256 S.W.2d 67, 70 (1953), this Court *100denied jurisdiction because the trial court order, though called an injunction, was in essence a mandamus. The Court explained that although the two actions are related, they serve different functions because injunctions are preventive while mandamus actions are remedial. Today’s decision wholly ignores this standard of strict construction.

Dismissing a case on jurisdictional grounds may be frustrating to judges and litigants alike, particularly when issues of statewide import are involved. The Supreme Court has determined that states should have the first opportunity to provide a constitutional redistricting plan, see Growe v. Emison, 507 U.S. 25, 40, 113 S.Ct. 1075, 122 L.Ed.2d 388 (1993), and state actors are understandably anxious to comply with this mandate of federalism. However, the Legislature has chosen to make direct appeal an uncommon remedy, available only in rare and specific situations. Regardless of the day’s exigencies, our highest and only duty is to respect the appropriate limits of our power.

As Justice Jackson once said for the Supreme Court: “We agree that this is a hard case, but we cannot agree that it should be allowed to make bad law.” Fed. Communications Comm’n v. WOKO, 329 U.S. 223, 229, 67 S.Ct. 213, 91 L.Ed. 204 (1946). I fear that our Court has allowed a hard case to make bad law today.

dissenting.

In my view: the Court’s remand to the district court is essentially pointless because the state judiciary is not prepared to adopt a congressional redistricting plan in time to avoid disrupting the 2002 elections, and under the United States Supreme Court’s decision in Growe v. Emison,1 the litigation should now shift once and for all to the federal court; and judgment should instead be rendered adopting Plan 1065C, in accordance with the district court’s stated intention, because the only objections to that plan, and all of the changes the court made to it, were based on an improper consideration — the protection of incumbents. Therefore I respectfully dissent.

I

Were present efforts to continue unabated, it might be possible to devise a state plan for congressional redistricting in time for the 2006 elections, maybe even with luck for the 2004 elections. But for the 2002 election cycle, which commences six weeks from Monday with the opening of the filing period,2 Texas courts, like the Texas Legislature, have now proven incapable of reapportioning the State’s congressional districts in time to avoid serious disruptions. To be sure, the 2002 congressional elections could be conducted on the kind of frantic timetable necessitated by the State’s failure to produce a lawful reapportionment plan in 1996, with a filing period opening as late as August 30, a special election among all filers in November, and any runoffs in December.3 But a plan devised at the last possible moment is hardly one adopted “ ‘within ample time ... to be utilized in the [upcoming] election’ ” within the meaning of Growe v. Emison.4

*101Litigation over the same redistricting issues has been pending in the United States District Court for the Eastern District of Texas since the spring,5 but in accordance with Growe,6 that court has deferred to state court proceedings to redraw congressional lines. On July 23, 2001, the federal court announced that its deferral to state efforts to redraw congressional districts would extend only until October 1 and then it would proceed to trial on October 15.7 None of the parties to the redistricting litigation suggested that the federal court should defer past October 15.8 The federal court’s reasonable deadline has now come and gone, and there is no state plan. Wrangling over which state district court should proceed, well underway on July 23, could not be resolved without this Court’s intervention on September 12.9 The trial in state court, commenced on September 17, was not concluded until September 28. At the state court’s request, the federal court extended the deadline for a ruling until October 3.10 On that day, the state court issued an order, in its words, “announc[ing] its intention to adopt” Plan 1065C.11 On October 10, however, the state court made what the federal court has fairly characterized as “major changes” in its ruling,12 issuing an altogether new Plan 1089C.13 The federal court immediately postponed its trial until October 22 but indicated that any further delay “would only ensure a delay in the elections.”14 I agree.

The situation is not like that in Growe, where a Minnesota court had given no indication that it “was either unwilling or unable to adopt a congressional plan in time for the elections.”15 On the contrary, the Minnesota court had adopted a legislative plan before the federal court did and was prepared to adopt a congressional plan, even though the federal court had stayed state court proceedings for a month.16 The case before us was filed on December 27, 2000, and has been litigated in earnest since the Legislature adjourned sine die on May 28, 2001,17 without enacting any redistricting legislation. Given the flaws in Plan 1089C, which the district court has now adopted, and the infirmities the Court finds in that court’s procedures, *102the state judicial process is not advancing toward a successful end any time soon. In remanding this case for further proceedings, the Court seems oblivious to the federal court’s reiterated concerns and the exigencies of the circumstances. The district court is to conduct additional hearings. Perhaps it will modify Plan 1089C, adopt another plan already in evidence, or create a new plan altogether, requiring further hearings and appeals on new issues. Perhaps the court will simply readopt Plan 1089C, only to be told on the next appeal to this Court that the plan was legally defective all along. Or perhaps this Court will insist that all further appeals come through the court of appeals. The best that can be said after months of litigation is that development of a state plan remains a remote possibility.

Growe seems to suggest that a federal court may take as a baseline a state court plan that has not been reviewed on appeal.18 I assume this accounts for the federal court’s statement that “under Growe, plan 1089C must be the only candidate for a baseline state plan in this case.”19 But even if it need not await the state appeals process to run its course, the federal court cannot take as its starting point a plan like 1089C that this Court has already declared invalid, as we do today. Were a valid plan imminent, the federal court might be pressed to defer a bit longer; but a valid plan is not imminent, or even likely at this point. If nothing else, the uncertainties left and the new ones created by today’s decision dispel any lingering hope that a Texas court can produce a valid congressional redistricting plan for 2002. Growe seems to counsel federal court deferral only so long as state courts can produce a plan in time for elections. The federal court need not wonder whether that is still possible; it is not.

Because state procedures should no longer impede the federal litigation, what happens in the state district court is no longer of immediate consequence. The court could attempt to devise a plan for the 2004 elections, in the event that the federal court plan is not preemptive and the Legislature does not revisit the issue in its next session. Or the parties may decide to abandon the state court proceedings altogether. In any event, the effect of today’s remand is to terminate the state courts’ efforts to develop a congressional redistricting plan for 2002.

II

The United States Supreme Court has instructed federal courts in redistricting cases to “follow the policies and preferences of the State, as expressed in statutory and constitutional provisions or in the reapportionment plans proposed by the state legislature, whenever adherence to state policy does not detract from the requirements of the Federal Constitution”.20 *103While the Supreme Court has not given state courts the same direction, this Court has stated that “[w]hen a court sets aside the plan chosen by the Legislature, the beginning point for fashioning a substitute plan must be those aspects of the legislated scheme which are valid.”21 The difficulty here is that the last expression of legislative will on the subject of congressional districting was in 1991. Three of the districts drawn by the 1991 Legislature were unconstitutional, resulting in changes to ten others,22 so that the existing districts are now largely an expression of the will of the federal court, not the will of the Legislature. It is inappropriate, of course, to look for direction to an illegal legislative plan. Moreover, Texas has changed dramatically since 1991, so that the legislative experience that year, separating out the legal goals from the illegal, may have limited relevance in 2001. Demographics have changed. The record reflects that the State’s population has grown 23%, from 16,986,510 to 20,851,820, of which Hispanics are now nearly one-third versus one-fourth in 1990. And politics have changed. In 1991 the congressional and legislative delegations were seventy and sixty-four percent Democrat, respectively; now they are fifty-seven and fifty-two percent Democrat. In 1991 there were six Republican statewide elected officials; in 2001 all twenty-seven statewide elected officials were Republican. These changes appear to have affected districting considerations. For example, while incumbent protection has been one. of the main aims of legislative redistricting in the past,23 under the plan adopted in July by the Legislative Redistricting Board, thirty-seven incumbents were placed in districts with one or more other incumbents. In sum, the legal deficiencies in the 1991- congressional redistricting plan enacted by the Legislature and the dramatic changes in the ethnic and political makeup of Texas’ population render that plan largely irrelevant in determining what principles should guide redistricting now.

The undeniably inherent and intense political nature of redistricting counsels that a state court drawing district lines must start with something besides a blank map, lest the court’s job become in all respects the equivalent of the Legislature’s. Courts have no business in such a non-adjudicative role. But where to start in this case? The only indication of current state policies is in the legislative redistricting completed by the Legislative Redistricting Board, a constitutional body24 charged by the people with legislative redistricting when the Legislature defaults. Arguably its work best reflects what factors should now govern redistricting, congressional as well as legislative.

But I need not resolve this question. After a two-week trial, at which many competing factors were argued and as many as twelve plans were presented, the district court on October 3, 2001, “announce[d] its intention to adopt” Plan 1065C, subject to “comments, proposed changes, or requested modifications” lodged by October 9. As I intend to show, almost all of the modifications requested in Plan 1065C, and the only ones not rejected by the district court, were based on pro-*104teeting incumbents. Because that is not a proper legal basis for judicial redistricting, Plan 1065C should not have been changed. Thus, it stands as a plan adopted by the district court to which no valid objections was raised, and this Court should order it to be the State’s plan for purposes of federal review.

A

The objections filed to Plan 1065C were of three types.

First, there were general objections to changes in historic district lines, all of which amounted to complaints that Plan 1065C did not sufficiently protect incumbents. For example, the Democratic congressional members who intervened in the case and Molly Beth Malcolm, chair of the State Democratic Party, argued that Plan 1065C did not “adhere[] to the Texas Legislature’s longstanding policies of preserving the cores of prior districts and protecting congressional incumbents.” Plan 1065C, they said, moved too far from historical district lines so as to impair the reelection of six Democratic incumbents. The existing and proposed districts of these members were analyzed district by district. Similarly, the plaintiffs objected that the district court was required to adhere as closely as possible to the 1991' legislation even though it was, they conceded, unconstitutional in significant respects and “partisan-based”. The plaintiffs themselves tell us in their brief that “the only party to propose more than technical changes to Plan 1065C was Speaker Laney.” In his objections, Speaker Laney argued for specific changes out of “respect for traditional communities of interest”, “to restore the core constituencies” of various districts, to “maintain[] the relationships between constituents and incumbents”, to return districts “closer to [their] roots and preserve more of the cores of the current districts”, to “preserve the rural character” of districts, to restore districts “closer to [their] traditional configuation”, and to return a district “to a configuration closer to its current form”.

Plaintiffs concede in their brief that these objections and incumbent-protection are “complementary”. As one court has observed, “the maintenance of the geographic and population cores of existing districts is a criterion designed primarily to protect incumbents.”25 But as argued in objection to Plan 1065C, restoring district lines and incumbent protection are completely congruent. Rhetoric aside, to grant these objections was to help incumbents, and to help incumbents was to grant these objections. Asked at oral argument here whether there were any areas covered by these objections that did not exactly overlap incumbent protection, appel-lees’ counsel was unable to identify even one. Counsel argued that preserving a “core constituency” or a “traditional configuration” might be desirable even if an incumbent were retiring, but the record does not reflect that any of the incumbents is retiring.

Second, there were objections to three districts based on the Voting Rights Act, but in each instance, the requested changes were, in context, tantamount to incumbent protection. Third, there were a few technical objections to insignificant flaws in the plan.

The district court made the exact changes urged by Speaker Laney. The court expressly said so in a letter to the parties — “The Court ... is seriously considering [changes] ... using Speaker La-ney’s proposed plan 1081C; ... based on Speaker Laney’s plan 1080C; ... based on *105Speaker Laney’s plan 1088C; ... based on Speaker Laney’s plan 1083C.” Furthermore, Plan 1089C reflects the changes urged by Speaker Laney and essentially no others. It is not clear whether the district court made any of the proposed technical changes or whether the extensive changes the court did make simply mooted the technical objections. Thus, the only objections to Plan 1065C, which the district court stated that it intended to adopt, and the only changes made in that plan to produce Plan 1089C, were to protect incumbents.

The appellees argue that Plan 1065C cannot be adopted as the state plan because the district court made no findings to support it, but the findings the court made in support of Plan 1089C were based largely on statistical materials that were never offered or admitted in evidence. The plan can be reviewed on the evidence viewed in a supporting light. The appel-lees also argue that the district court considered incumbent protection in developing Plan 1065C, and if it did it erred, but appellants did not object to that plan on that basis. Assuming they are correct, that does not justify changes to the plan based on the protection of incumbents to which there was and is vigorous objection.

B

The protection of incumbents is a legitimate redistricting goal — for the Legislature.26 The Legislature may decide to give priority to protecting relationships between representatives and constituents and benefitting from the experience and seniority incumbents have acquired. The Legislature may also decide to protect incumbents to achieve a desired partisan composition of the delegation. But all of these factors are almost purely political and for this reason are not appropriate for judicial consideration in drawing lines when the Legislature has failed to do so. A court simply has no business making the partisan political decision that particular representatives should or should not be reelected. Thus, the Fifth Circuit stated in Wyche v. Madison Parish Police Jury: “Many factors, such as the protection of incumbents, that are appropriate in the legislative development of an apportionment plan have no place in a plan formulated by the courts.”27 The three-judge federal court that drew Texas’ existing congressional district echoed this statement.28 Likewise, in Good v. Austin, a three-judge federal court devising a redistricting plan refused to consider “the maintenance of the geographic and population cores of existing districts” because it was a criteria “designed primarily to protect incumbents” and was thus “so laden with political considerations” as to be “inappropriate ... in the formulation of a judicial redistricting plan.”29

Accordingly, I would hold that it was improper for the district court to consider the protection of incumbents in adopting a redistricting plan.

C

Because the district court expressly adopted Speaker Laney’s changes in Plan 1065C, and because all of those changes were plainly for the protection of incumbents, I would hold that Plan 1089C was not properly adopted. Determining the purpose of the changes might well present *106fact questions but for their characterization in the objections that were filed. The district court appears to have overruled the other objections to Plan 1065C, and therefore it was left without any valid objection to it. For this reason, I would render judgment adopting Plan 1065C as the state congressional redistricting plan for consideration by the federal court in the pending cases.

Ill

I agree with the Court on three other issues:

I agree that a redistricting plan cannot be presumed to be the State’s plan merely because it has been proposed by the Attorney General. Without belaboring the matter, the Attorney General’s argument is simply inconsistent with the Court’s opinion in Terrazas v. Ramirez.30

I am also persuaded that the district court adopted Plan 1089C in violation of the procedures we prescribed in Terrazas. True, the district court in Terrazas ordered a redistricting plan without hearing any evidence or argument on how district lines should be drawn, and here there was a two-week trial on that issue. Appellees argue that Plan 1089C was based on evidence and arguments at trial, but they concede that the exact lines in the plan cannot be found in the evidence. Appellants were entitled to an opportunity to challenge Plan 1089C, which they had never seen, and the district court denied them this opportunity. I presume that the district court’s rush was attributable to the fact that the federal deadline had already passed, but whatever the reason, appellants were not given any reasonable opportunity to challenge Plan 1089C before it became, by virtue of the district court’s order, the baseline for the federal court litigation. In most other cases, parties may challenge a trial court’s judgment through motions to reconsider, to reopen the evidence, and for a new trial. But redistricting cases must often be litigated in exigent circumstances, and here ordinary post-judgment remedies would not afford appellants relief before their rights were severely impacted.

Finally, I agree that the Court has jurisdiction over this direct appeal. Chief Justice Phillips’s explanation why the Court lacks jurisdiction over this direct appeal seems little more than a contrivance to avoid a relatively plain statute. Section 22.001(c) of the Government Code states: “An appeal may be taken directly to the supreme court from an order of a trial court granting or denying an interlocutory or permanent injunction on the ground of the constitutionality of a statute of this state.”31 Here we have an appeal of an order of a trial court granting a permanent injunction on the ground that a statute— article 197h of the Texas Revised Civil Statutes, as later modified by the federal court — is unconstitutional. That would seem to satisfy section 22.001(c) exactly. But the Chief Justice argues that the Court does not have jurisdiction over all appeals from such orders; it has jurisdiction only over some appeals, namely where one party complains of the ruling on the constitutionality of the statute. This argument adds to the statute an additional requirement that cannot be found in its language or history. In the sixty-eight years since the predecessor to section 22.001(c) was enacted, the situation has never arisen. The Chief Justice’s argument is based on a line of dicta in Bryson v. High Plains Underground Water Con*107semotion District,32 The Court has never denied jurisdiction over a direct appeal, not in Btijson or in any other case, because the parties agreed that a statute was unconstitutional but disagreed on the relief ordered. Given the clarity of the statute, this is hardly surprising.

In most direct appeals, time is important but it is not of the essence, as it is in this case. If ever a direct appeal to this Court were appropriate, this is the kind of case for it. Few like it will ever arise, where the parties agree that a statute is unconstitutional and that injunctive relief should issue but disagree over the terms of that relief. To misapply the plain language of section 22.001(c), without a word of authority in support, in a case where haste is crucial cannot be justified.

For these reasons I respectfully dissent.

joined by Justice ENOCH and Justice JEFFERSON, concurring.

I join fully in the Court’s opinion, but I write separately because I share the concerns raised by Justice Hecht in Part 11(B) of his opinion regarding the propriety of the trial court’s consideration of incumbency protection issues in formulating its plan. 67 S.W.3d at 100 (Hecht, J.,. dissenting). The changes asked for by Speaker Laney and ultimately adopted by the trial court in plan 1089C, while not couched in terms of incumbent protection, are not so distinct from incumbent protection as to escape suspicion. And I would caution on remand that the trial court may not consider incumbent protection in drawing Congressional district lines.

That the trial court used incumbent protection as a factor in drawing the district lines for plan 1089C is an issue raised by the Attorney General in his Appellant’s Brief and by others. I believe it is appropriate for this Court to provide guidance to *98the trial court on this issue on remand. See, e.g., Dallas Mkt. Ctr. Dev. Co. v. Liedeker, 958 S.W.2d 382, 385 (Tex. 1997) (considering an issue raised by a party to provide guidance to the district court on remand); Edinburg Hosp. Auth. v. Trevino, 941 S.W.2d 76, 81 (Tex. 1997) (considering an issue “not essential to [the Court’s] disposition of th[e] case ... to provide the trial court with guidance in the retrial of [the] case”).