Martinez v. Bush

234 F.Supp.2d 1275 (2002)

Honorable Mayor Raul L. MARTINEZ, Bishop Victor T. Curry, Hattie D. Harden, Southwest Voter Registration Education Project, Inc., Robert Pope, James O. Brown, State Senator Betty Holzendorf, Dr. Carl Warren, State Representative Denise Lee, State Representative Gary Siplin, State Representative Matthew Meadows, State Representative, Dorothy Bendross-Mindingall, Rev. Richard P. Dunn, III, Enid Pinkney, Jacqui Colyer, Alexander Kelly, Maria B. Reckley, Councilman Philippe Derose, Unrepresented People's Positive Action Council, Inc., Nury Molina, State Representative Cindy Lerner, Patrick Vilar, Lorenzo S. Ruiz, Jose A. Paez, and Lavita Holmes Plaintiffs
v.
John Ellis "Jeb" BUSH, Katherine Harris, Tom Feeney, John McKay, and Robert Butterworth Defendants
Peter R. Deutsch, etc., et al. Intervenor-Plaintiffs
v.
John Ellis "Jeb" Bush, et al. Defendants
George Maurer Plaintiff
v.
State of Florida, Jeb Bush, Katherine Harris, Tom Feeney, and John McKay Defendants

No. 02-20244-CIV-JORDAN. No. 02-10028-CIV-JORDAN.

United States District Court, S.D. Florida, Miami Division.

December 3, 2002.

Norman C. Powell, Miami, FL, Thomasina H. Williams, Miami, FL, J. Gerald Herbert, Alexandria, VA, Ephraim Roy Hess, Colleen Kathryn O'Loughlin, Ft. Lauderdale, FL, Ronald A. Kalain, Jeremy B. Bash, Goodwin Liu, Washington, DC, for plaintiffs.

Paul F. Hancock, Maxine S. Ryan, Deputy Atty. Gen., Office of Atty. Gen., Fort Lauderdale, FL, George Waas, Senior Assist. Atty. Gen., Office of Atty. Gen., Tallahassee, FL, for State of Florida.

Dean Colson, Roberto Martinez, Colson Hicks Eidson, Coral Gables, FL, Charles Canady, Tallahassee, FL, for Governor Jeb Bush.

Deborah K. Kearney, Tallahassee, FL, for Secretary of State.

Christopher Carver, Miami, FL, J. Thomas Cardwell, Orlando, FL, Miguel Angel De Grandy, Miami, FL, Jason Lawrence Unger, George N. Meros, Jr., Tallahassee, FL, Joseph W. Hatchett, Tallahassee, for Speaker Feeney.

Edward Joseph Pozzuoli, III, James A. Scott, Alexis M. Yarbrough, Fort Lauderdale, FL, Thomas Emerson Scott, Jr., Miami, FL, for president McKay.

George Maurer, Key West, FL, pro se.

Before TJOFLAT, Circuit Judge, and HINKLE and JORDAN, District Judges.^[*]

MEMORANDUM OPINION

PER CURIAM

INTRODUCTION

Every ten years, after the census, the Florida legislature is required to redraw the State's congressional districts and the State Senate and House districts to adjust for population shifts. The 2000 census revealed that Florida's population grew by more than 3 million people during the 1990s. As a result, Florida was apportioned two additional representatives in the United States House of Representatives, and the Florida legislature, which was controlled by a substantial Republican majority in each house, adopted a redistricting plan for the State's Congressional districts to adjust for the shifts in population and the increase in the number of representatives. On March 27, 2002, Governor John Ellis "Jeb" Bush signed into law House Bill 1993, which established the lines for Florida's 25 congressional districts.

Following the 2000 census, the Republican-controlled Florida legislature also adopted redistricting plans for the State Senate and House districts to adjust for the population changes since the 1990 census. On March 22, 2002, the Florida legislature passed House Joint Resolution 1987, which established the lines for Florida's 40 Senate districts and 120 House districts.

Following the passage of the redistricting legislation, plaintiffs and intervenors brought this action for declaratory and injunctive relief against the Speaker of the Florida House of Representatives, the President of the Florida Senate, the Governor of Florida, the Florida Secretary of State, and the Florida Attorney General.^[1] Plaintiffs alleged (1) that the process used to arrive at the redistricting plan was inadequate under the Equal Protection Clause of the Fourteenth Amendment and Title VI of the Civil Rights Act of 1964, 42 U.S.C. § 2000d, because minorities were not given sufficient opportunity to participate; (2) that the reapportionment plan led to the dilution of black^[2] voting power^[3] in violation of section 2 of the Voting Rights Act of 1965, 42 U.S.C. § 1973; and (3) that the Florida legislature politically gerrymandered the congressional districts to favor Republicans, thereby violating the rights of Democratic voters guaranteed by the Equal Protection Clause of the Fourteenth Amendment.^[4] We address each of these claims in turn.^[5]

In part I, we address the plaintiffs' "process" claims. First, we present detailed findings of fact regarding the redistricting process and the opportunity afforded Florida citizens to participate in that process. We conclude that plaintiffs have failed to present sufficient evidence to show that defendants were motivated by discrimination against blacks or Hispanics in deciding what redistricting software or allocation method to use, where to hold public hearings, when to hold public hearings, what type of notice to provide, or whether to consider input from Florida citizens and Democratic legislators in drawing the redistricting plans. Absent such a showing, plaintiffs' process claims must fail.

In part II, we address plaintiffs' vote dilution claims. We present detailed findings of fact comparing voting behavior by race in performing black districts existing before and after the redistricting process at issue in this litigation. We conclude that in the districts under challenge in this litigation, black candidates of choice^[6] will continue to prevail in most elections. We outline the three Gingles preconditions for prevailing on a claim of vote dilution under section 2 of the Voting Rights Act: (1) geographic compactness, (2) political cohesiveness, and (3) bloc voting on behalf of the majority group. Thornburg v. Gingles, 478 U.S. 30, 50-51, 106 S.Ct. 2752, 2766-67, 92 L.Ed.2d 25 (1986). Because it is undisputed that these preconditions are met, section 2 clearly required the drawing of at least some number of districts likely to perform for black candidates of choice. We conclude that the legislature satisfied its section 2 obligation by drawing performing black districts which are, in fact, likely to perform for black candidates of choice. Therefore, under the redistricting plans, black voters will not have "less opportunity than other members of the electorate to participate in the political process and to elect representatives of their choice," 42 U.S.C. § 1973(b)—i.e., black voting power has not been diluted— and plaintiffs have failed to establish a violation of section 2 of the Voting Rights Act.

Finally, in part III, we address plaintiffs' political gerrymandering claims. We begin by holding that the three Gingles preconditions apply to equal protection claims in the political gerrymandering context just as they do in the racial vote dilution context. Next, we address the two-prong Bandemer test for unlawful political gerrymandering—legislative intent to discriminate and discriminatory effects. Davis v. Bandemer, 478 U.S. 109, 127, 106 S.Ct. 2797, 2808, 92 L.Ed.2d 85 (1986). Although we conclude that the Florida legislature intended to draw the congressional districts in a way that favors Republicans, we conclude that plaintiffs have failed to demonstrate discriminatory effects that are serious enough to warrant federal court intervention. Plaintiffs failed to establish the preconditions of geographic compactness, political cohesiveness, and bloc voting on behalf of the majority group. Plaintiffs also failed to establish that, under a totality of the circumstances, the redistricting plan gives Democrats "less opportunity to participate in the political process and to elect candidates of their choice." Id. at 131, 106 S.Ct. at 2809. Thus, this final claim too must fail.

I. THE "PROCESS" CLAIMS

The Martinez plaintiffs alleged in count I of their second amended complaint that the process, practice, and procedure adopted and employed by Florida's legislature in conducting redistricting hearings and adopting the state's congressional and legislative plans violated the Equal Protection Clause of the Fourteenth Amendment and 42 U.S.C. § 1983. They also alleged a violation of Title VI of the Civil Rights Act of 1964, 42 U.S.C. § 2000d, based on the same conduct. We refer to these claims as the "process" claims.^[7] Because, as we explain below, these claims all require a showing of intentional discrimination, and because the question of discrimination is a question of fact, and not an issue of law, we briefly summarize the legal framework which governs the "process" claims.

A. LEGAL FRAMEWORK

The central purpose of the Fourteenth Amendment, as articulated by cases like Washington v. Davis, 426 U.S. 229, 239, 96 S.Ct. 2040, 2047, 48 L.Ed.2d 597 (1976), is "racial neutrality in governmental decision making." Clark v. Putnam County, 293 F.3d 1261, 1266 (11th Cir.2002). "[A]s a general rule, `[d]istinctions between citizens solely because of their ancestry are by their very nature odious to a free people whose institutions are founded upon a doctrine of equality.'" Oyama v. California, 332 U.S. 633, 646, 68 S.Ct. 269, 275, 92 L.Ed. 249 (1948) (quoting Hirabayashi v. United States, 320 U.S. 81, 100, 63 S.Ct. 1375, 1385, 87 L.Ed. 1774 (1943)).

To prevail on their equal protection claim relating to the redistricting hearing and adoption processes, the Martinez plaintiffs must show, by a preponderance of the evidence, that the Florida legislature acted with racially discriminatory intent. See Washington, 426 U.S. at 240, 96 S.Ct. at 2048 (noting that school desegregation cases have "adhered to the basic equal protection principle that the invidious quality of a law claimed to be racially discriminatory must ultimately be traced to a racially discriminatory purpose"); Village of Arlington Heights v. Metropolitan House. Dev. Corp., 429 U.S. 252, 265, 97 S.Ct. 555, 563, 50 L.Ed.2d 450 (1977) ("Proof of racially discriminatory intent or purpose is required to show a violation of the Equal Protection Clause."); cf. Crawford-El v. Britton, 523 U.S. 574, 592-94, 118 S.Ct. 1584, 1594, 140 L.Ed.2d 759 (1998) (rejecting argument that where state of mind is an element of a constitutional claim, mens rea must be shown by clear and convincing evidence). Although the Martinez plaintiffs need not establish that a racially discriminatory motive was the sole or primary purpose, they have to establish that it was a motivating factor of the challenged process. See Village of Arlington Heights, 429 U.S. at 265, 97 S.Ct. at 563 (noting that a plaintiff need not prove that a "challenged action rested solely on racially discriminatory purposes"); Garza v. County of Los Angeles, 918 F.2d 763, 771 (9th Cir.1990) (holding that the redistricting plan of county supervisor districts violated the Equal Protection Clause where the supervisors who drew the districts were motivated by both an intent to fragment the Hispanic voting population and an intent to preserve incumbencies, because "discrimination need not be the sole goal in order to be unlawful") (citing Arlington Heights, 429 U.S. at 265, 97 S.Ct. at 555). The fact that biased state decision makers may have had other motives that were not discriminatory does not make their conduct any less intentional. See, e.g., Barnett v. Daley, 32 F.3d 1196, 1199 (7th Cir.1994) (holding that "the fact that discrimination may have an ulterior motive that is not discriminatory does not make it any the less intentional" for equal protection purposes in a challenge to a redistricting plan of city alderman wards). Once the Martinez plaintiffs establish a prima facie case of discrimination, the defendants may only overcome the presumption of unconstitutional action by showing that permissibly neutral criteria and procedures produced the discriminatory result. See Washington, 426 U.S. at 241, 96 S.Ct. at 2048.

An invidious discriminatory racial purpose need not be express, and "may often be inferred from the totality of the relevant facts including the fact, if it is true, that the law bears more heavily on one race or another." Washington, 426 U.S. at 241-42, 96 S.Ct. at 2048-49. The factors that may be probative of a discriminatory purpose or intent include: (i) the impact of the official action; (ii) the historical background of the decision, particularly if it reveals a series of official actions taken for invidious purposes; (iii) a specific sequence of events leading up to the challenged decision; (iv) departures from normal procedural sequences; and (v) substantive departures from factors usually considered important by the decision maker and which strongly favor a decision contrary to the one reached. See Village of Arlington Heights, 429 U.S. at 266-67, 97 S.Ct. at 564-65.

Like the equal protection claim, the Title VI and section 1983 claims are dependent on a showing of intentional discrimination. First, Title VI provides no greater protection than the Equal Protection Clause. See United States v. Fordice, 505 U.S. 717, 732 n. 7, 112 S.Ct. 2727, 2737 n. 7, 120 L.Ed.2d 575 (1992) ("Our cases make clear and the parties do not disagree that the reach of Title VI's protection extends no further than the Fourteenth Amendment."); Elston v. Talladega County Bd. of Educ., 997 F.2d 1394, 1405-06 n. 11 (11th Cir.1993) ("Since Title VI itself provides no more protection than the Equal Protection Clause—both provisions bar only intentional discrimination—we will not engage in a separate discussion of the Title VI statutory analysis, as such an inquiry would duplicate our equal protection analysis."). Second, it is well settled that a section 1983 plaintiff must show a violation of a right secured by the Constitution or federal law in order to prevail. See, e.g., Collins v. City of Harker Heights, 503 U.S. 115, 120-22, 112 S.Ct. 1061, 1066-67, 117 L.Ed.2d 261 (1992); Baker v. McCollan, 443 U.S. 137, 140, 99 S.Ct. 2689, 2692, 61 L.Ed.2d 433 (1979); Monteiro v. Tempe Union High School Dist., 158 F.3d 1022, 1026 (9th Cir.1998). A "finding of intentional discrimination is a finding of fact" subject to clearly erroneous review. Anderson v. City of Bessemer City, 470 U.S. 564, 573, 105 S.Ct. 1504, 1511, 84 L.Ed.2d 518 (1985); see also, e.g., Easley v. Cromartie, 532 U.S. 234, 241-58, 121 S.Ct. 1452, 1458-66, 149 L.Ed.2d 430 (2001) (reviewing, under the clearly erroneous standard, finding of three-judge court that congressional redistricting plan was racially motivated). We therefore proceed to make our subsidiary findings of fact and our ultimate findings on the issue of intentional discrimination.

B. PRELIMINARY FINDINGS OF FACT FOR THE "PROCESS" CLAIMS

1. THE PARTIES

1. Raul L. Martinez is an Hispanic resident and the Mayor of Hialeah, Florida. Mr. Martinez is a voter registered as a Democrat and a resident of Congressional District ("CD") 21, Florida Senate District ("SD") 39, and Florida House District ("HD") 110.

2. George Maurer is a resident of Monroe County, Florida, and a resident of former CD 20 and new CD 18, HD 120, and SD 40. Mr. Maurer is a registered Democrat, is presently the Chair of the Lower Keys & Key West Democratic Club, the Vice-Chair of the 20th Congressional District Democratic Organization, a Monroe County Democratic State Committeeman, and a member of the Florida State Democratic Central & Executive Committees.

3. Victor T. Curry is a black resident of Miami-Dade County, Florida. Mr. Curry is a voter registered as a Democrat and a resident of CD 17 and HD 109.

4. Hattie D. Harden is a black resident of Broward County, Florida. Ms. Harden is a voter registered as a Democrat and resident of former and new CD 23.

5. The Southwest Voter Registration Education Project, Inc. ("SVREP") is a national, non-partisan organization headquartered in San Antonio, Texas, which supports voter registration throughout the country. SVREP has an office in Miami, Florida, but does not represent individuals or have individual members. Alvaro Fernandez presently serves as Executive Director of SVREP for the State of Florida, and has been involved with SVREP since September of 2000. Mr. Fernandez, a Cuban-American, has lived in Miami, Florida, for approximately the past 40 years. Prior to his work with SVREP, Mr. Fernandez was Special Assistant to Mayor Martinez of the City of Hialeah, has been a political candidate for the Florida House of Representatives, and has been a candidate for the position of county commissioner. SVREP maintains a database containing the names of many of those with whom it has worked over the years and regularly stays in contact with the persons listed therein through the regular circulation of letters, e-mails, and newsletters. The Florida Chapter of SVREP has been active in various communities throughout the State of Florida, including Miami-Dade County, Broward County, Palm Beach County, and several areas surrounding the City of Orlando. The organization has focused its attention on these areas of the State, in part, because of the surge in Hispanic residents therein. From the time of its inception to the present date, the Florida Chapter of SVREP has been successful in registering between 5,000 and 10,000 new voters in the State of Florida. On behalf of SVREP, Mr. Fernandez attended a number of the public redistricting hearings throughout the state (e.g., Tallahassee, Tampa, Orlando, Palm Beach, and Miami (two)).

6. Robert Poe is the Chairman of the Florida Democratic Party and a registered Democrat.

7. James O. Brown is a black resident of Duval County, Florida. Mr. Brown is a voter registered as a Democrat and a resident of former and new CD 3, former SD 2 (renumbered SD 1 in the new legislative redistricting plan), and former and new HD 15.

8. State Senator Betty Holzendorf is a black resident of Duval County, Florida. Senator Holzendorf is a voter registered as a Democrat and a resident of former and new CD 3 and SD 2 (renumbered SD 1 in the new legislative redistricting plan).

9. Carl Warren is a black resident of Hillsborough County, Florida. Mr. Warren is a voter and a resident of SD 21 (renumbered SD 18 in the new legislative redistricting plan).

10. State Representative E. Denise Lee has been a candidate for HD 15 and may be a candidate for SD 1.

11. State Representative Gary Siplin is a black resident of Orange County, Florida. Representative Siplin is a voter registered as a Democrat and a resident of former and new CD 3 and former and new HD 39.

12. State Representative Matthew Meadows is a black resident of Broward County, Florida. Representative Meadows is a voter registered as a Democrat and a resident of former and new HD 94.

13. State Representative Dorothy Bendross-Mindingall is a black resident of Miami-Dade County, Florida. Representative Bendross-Mindingall is a voter registered as a Democrat and a resident of former and new CD 17 and former and new HD 109.

14. Richard P. Dunn, III, is a black resident of Miami-Dade County, Florida. Mr. Dunn is a voter registered as a Democrat and a resident of former and new CD 3 and HD 109.

15. Enid Pinkney is a black resident of Miami-Dade County, Florida. Ms. Pinkney is a voter registered as a Democrat and a resident of former and new CD 3 and HD 109.

16. Jacqui Coyler is a black resident of Miami-Dade County, Florida. Ms. Coyler is a voter registered as a Democrat and a resident of former and new CD 3 and HD 109.

17. The Unrepresented People's Positive Action Council, Inc., ("UPPAC") is a not-for-profit Florida corporation headquartered in Miami-Dade County, Florida. UPPAC's goals include encouraging participation by members of the black community in the political process.

18. Alexander Kelly is a black resident of Miami-Dade County, Florida. Mr. Kelly is a voter registered as a Democrat and a resident of former and new CD 3 and HD 109. Mr. Kelly is a member of UPPAC.

19. Maria B. Reckley is a black resident of Miami-Dade County, Florida. Ms. Reckley is a voter registered as a Democrat and a resident of former and new CD 3 and HD 103. Ms. Reckley is a member of UPPAC.

20. Philippe Derose is a black resident of Miami-Dade County, Florida. Mr. Derose is a voter registered as a Democrat and a resident of former and new CD 3 and HD 108.

21. Nury Molina is an Hispanic resident of Miami-Dade County, Florida. Ms. Molina is a voter registered as a Democrat and a resident of CD 18 and HD 119.

22. State Representative Cindy Lerner is a resident of Miami-Dade County, Florida. Representative Lerner is a voter registered as a Democrat, a resident of former HD 119, and the representative in Florida's House of Representatives for former HD 119. Representative Lerner was elected to the Florida House of Representatives during the 2000 election year, beating her Republican opponent by a margin of about 700 votes.

23. Patrick Vilar is an Hispanic resident of Miami-Dade County, Florida. Mr. Vilar is a voter registered as a Democrat and a Democratic candidate for new HD 116.

24. Lorenzo S. Ruiz is an Hispanic resident of Miami-Dade County, Florida. Mr. Ruiz is a voter registered as a Democrat and a resident of CD 21.

25. Jose A. Paez is an Hispanic resident of Miami-Dade County, Florida. Mr. Paez is a voter registered as a Democrat and a resident of CD 18 and HD 107.

26. Lavita Holmes is a black resident of Hendry County, Florida. Ms. Holmes is a voter registered as a Democrat and a resident of former and new CD 23.

27. Congressman Peter R. Deutsch is a resident of former and new CD 20 and is a registered Democrat. Representative Deutsch is a member of the United States House of Representatives for former CD 20, and has served in the House since 1992. Immediately prior to his service in the House, Representative Deutsch was a 10-year member of the Florida House of Representatives. Throughout his years as a legislator, Representative Deutsch has had extensive experience in issues relating to redistricting and has even served as chairman of the Congressional Reapportionment Committee in the Florida House of Representatives.

28. Ida Scott is a black resident of Hallandale, Florida. Ms. Scott is a voter registered as a Democrat and a resident of former CD 23 and new CD 17, HD 101, and SD 29.

29. Josh Brown is a black resident of Hallandale, Florida. Mr. Brown is a voter registered as a Democrat and a resident of former CD 23 and new CD 17, HD 99, and SD 29.

30. Bobbi Grace is a black resident of Dania, Florida. Ms. Grace is a voter registered as a Democrat and a resident of former CD 23 and new CD 20, HD 99, and SD 29.

31. Wayne Johnson is a resident of Lauderhill, Florida. Mr. Johnson is a voter registered as a Democrat and a resident of former CD 19 and new CD 20, HD 97, and SD 32.

32. Isadore Nachimson is a resident of Century Village, Pembroke Pines, Florida. Ms. Nachimson is a voter registered as a Democrat and a resident of former CD 20 and new CD 23, HD 101, and SD 32.

33. Ruby Lobben is a black resident of Lauderhill, Florida. Ms. Lobben is a voter registered as a Democrat and a resident of former CD 19 and new CD 23, HD 101, and SD 32.

34. William Totin is a resident of Lauderhill, Florida. Mr. Totin is a voter registered as a Democrat and a resident of former CD 20 and new CD 23, HD 97, and SD 33.

35. Natalie Mayeri is a resident of Century Village, Pembroke Pines, Florida. Ms. Mayeri is a voter registered as a Democrat and a resident of former CD 20 and new CD 23, HD 101, and SD 32.

36. Jose I. Perez is a resident of Pembroke Pines, Florida. Mr. Perez is a voter registered as a Democrat and a resident of former CD 20 and new CD 21, HD 101, and SD 32.

37. Joseph F. Garcia is a resident of Surfside, Florida. Mr. Garcia is a voter registered as a Democrat and a resident of former CD 22 and new CD 18, HD 106, and SD 38.

38. Jacqueline Rubin is a resident of Surfside, Florida. Ms. Rubin is a voter registered as a Democrat and a resident of former CD 22 and new CD 18, HD 106, and SD 38.

39. Alex G. Fekete is the Mayor of Pembroke Pines, Florida. Mayor Fekete also serves on the Board of Directors for the Florida League of Cities and the National League of Cities and is a member of the U.S. Conference of Mayors.

40. The City of Pembroke Pines is located in Broward County, Florida, and has experienced very fast growth in the last decade, going from 65,000 residents in 1990 to over 140,000 today. Pembroke Pines was entirely within the former CD 20, but will now be in parts of four new CDs (new CDs 17, 20, 21, and 23).

41. Claudia Davenport is a voter registered as a Democrat. Ms. Davenport resides in Everglades, Collier County, Florida. Ms. Davenport resides in former CD 14, HD 102, and SD 29. Ms. Davenport attended and spoke at the public redistricting hearing at Naples High School on September 24, 2001, and communicated with the Redistricting Committee via e-mail.

42. Gerald R. Rizzo is a white, non-Hispanic resident of St. Petersburg, Pinellas County, Florida. Mr. Rizzo is a voter registered as a Democrat. Mr. Rizzo resides in former CD 10 and new CD 11.

43. John Ellis Bush is the Republican Governor of the State of Florida.

44. Tom Feeney is the Republican Speaker of the Florida House of Representatives.

45. John McKay is the Republican President of the Florida Senate.

46. Katherine Harris is the former Republican Secretary of State for the State of Florida.

47. Robert Butterworth is the former Democratic Attorney General for the State of Florida.

2. THE 2000 CENSUS

48. Florida's population grew by more than 3 million people during the 1990s. As a result of the 2000 census, Florida was apportioned two additional representatives in the United States House of Representatives, bringing the total to 25. The total population of the State of Florida in 1990 was 12,937,926. The 2000 population was 15,982,378, which represents an increase of 3,044,452 persons or 23.52% over the 10-year period.

49. The 2000 census documented that some areas of Florida had grown more rapidly than others.

3. THE LEGISLATURE'S REDISTRICTING PLANS

50. On March 22, 2002, the Florida legislature, which is controlled by Republicans in both the State House and State Senate, passed House Joint Resolution 1987. This state legislative plan established the lines for Florida's 120 House districts and 40 Senate districts.

51. The maps and statistics for the state legislative plan are set forth in Speaker's Exhibit 162, "Plan H062H001 Census and Election Statistics (House)"; Speaker's Exhibit 79, attached CDs L1 (House Districts) & L2 (Senate Districts); Speaker's Exhibit 135, "Notebook of District-by-District Analyses of House Districts 1-120"; and Martinez Exhibits 77-A, 77-B, & 77-C.

52. On March 27, 2002, Governor Bush signed into law House Bill 1993. This redistricting plan established the lines for Florida's 25 congressional districts.

53. The maps and statistics for the enacted congressional plan are set forth in Speaker's Exhibits 68 & 69, "Maps, Statistics, FREDS file of Plan S19C0017 and Individual Districts in Plan."

54. The districts in the congressional plan have populations of either 639,295 or 639,296 persons.

4. PRE-CLEARANCE UNDER SECTION 5 OF THE VOTING RIGHTS ACT

55. By letter dated April 29, 2002, and pursuant to section 5 of the Voting Rights Act, 42 U.S.C. § 1973c, Governor Bush, Speaker Feeney, and President McKay submitted the congressional plan and Speaker Feeney, and President McKay submitted the state legislative plan to the United States Department of Justice for pre-clearance.

56. On May 3, 2002, the Florida Supreme Court issued its opinion validating the state legislative districts under article III, section 16 of the Florida Constitution and following limited review of applicable principles arising under the United States Constitution. See In re Constitutionality of House Joint Resolution 1987, 817 So.2d 819 (Fla.2002). In its opinion, the Florida Supreme Court found that "the maximum percentage deviation between the largest and smallest number of people per [House district] (statistical overall range) is 2.79%." Id. at 825. In contrast, "the maximum percentage deviation between the largest and smallest number of people per [Senate district] (statistical overall range) is 0.93%." Id. at 826. The Florida Supreme Court thus held that "the Legislature has achieved a mathematical preciseness in the districts that complies with the equal protection requirements of both the Florida and United States Constitutions." Id. The Florida Supreme Court also said, among other things, that due to time constraints it was not able to rule on fact-intensive federal challenges to the state legislative districts, including claims of political gerrymandering under the Fourteenth Amendment's Equal Protection Clause and claims based on section 2 of the Voting Rights Act, 42 U.S.C. § 1973. See id. at 828-29. Such challenges, the Florida Supreme Court said, would have to be brought in a court of "competent jurisdiction." Id. at 829.^[8]

57. On May 14, 2002, Attorney General Butterworth filed suit in the United States District Court for the District of Columbia, initially asking for a declaration regarding the validity of the various redistricting plans under section 5 of the Voting Rights Act, and then amending his complaint to request a declaration of validity. See Florida v. United States, Case No. 1:02 CV 00941 (D.D.C.2002) (the "D.C. action"). A three-judge court was then impaneled in the District of Columbia pursuant to 28 U.S.C. § 2284.^[9]

58. On May 15, 2002, certainMartinez plaintiffs—Mr. Warren, Ms. Holmes, and SVREP—along with Representative Alcee Hastings and the UNO Federation for Community Services, Inc., moved to intervene as plaintiffs in the D.C. action. These Martinez plaintiffs alleged, in part, that the section 5 submissions by Governor Bush, Speaker Feeney, and President McKay were improper because they were not made by Attorney General Butterworth.^[10] On May 28, 2002, the three-judge court in the D.C. action granted the motion to intervene, but aligned the intervenors as defendants.

59. On May 24, 2002, the Deutsch intervenors —Representative Deutsch, Ms. Grace, Mr. Totin, Mr. Johnson, Ms. Lobben, Ms. Mayeri, Ms. Nachimson, and Mr. Rizzo—moved to intervene in the D.C. action as defendants.

60. On May 29, 2002, Governor Bush, Speaker Feeney, and President McKay moved to intervene in the D.C. action as plaintiffs.

61. In the D.C. action, the Department of Justice took the position that the pre-clearance submission had been properly made by Governor Bush, Speaker Feeney, and President McKay.

62. On June 4, 2002, the three-judge court granted the motions to intervene filed by Governor Bush, Speaker Feeney, and President McKay and by the Deutsch intervenors, aligning the Governor, the Speaker, and the President as plaintiffs and the Deutsch intervenors as defendants.

63. On June 7, 2002, the United States Department of Justice pre-cleared Florida's congressional redistricting plan, stating by letter that the Attorney General "d[id] not interpose any objection" to that plan. By directing the letter to Governor Bush, Speaker Feeney, and President McKay, the Department of Justice apparently concluded that the submission of the congressional plan by these Florida officials did not constitute a change in voting procedure that had to be pre-cleared pursuant to section 5.

64. On June 7, 2002, the Governor, the Speaker, and the President moved on an emergency basis to dismiss the D.C. action as moot.

65. On June 10, 2002, the United States also moved to dismiss the D.C. action as moot.

66. On June 13, 2002, the three-judge court in the D.C. action granted the motions to dismiss filed by the United States, Governor Bush, Speaker Feeney, and President McKay, and dismissed all pending claims in that action as moot.

67. By virtue of the dismissal, the three-judge court in the D.C. action rejected the claim raised by the intervening Martinez plaintiffs that the submission of the congressional plan by Governor Bush, Speaker Feeney, and President McKay constituted a change in voting procedure that had to be pre-cleared pursuant to section 5. This dismissal therefore includes the claim raised in count XV of the Martinez plaintiffs' second amended complaint, and constitutes an adjudication of the issue of whether the submission by Governor Bush, Speaker Feeney, and President McKay was proper under section 5 and a finding that their submission was proper.^[11]

68. On June 20, 2002, the Department of Justice pre-cleared Florida's State Senate redistricting plan.

69. On July 1, 2002, the Department of Justice, by letter, interposed an objection to Florida's State House plan, concluding that proposed HD 101 was retrogressive within the meaning of section 5 of the Voting Rights Act. Due to this objection, we held an emergency evidentiary hearing and issued an order adopting an interim State House plan that had been proposed by Speaker Feeney. See Order Granting Speaker's Motion for Emergency Relief and Adopting Interim Remedial State House Plan [D.E. 315]. We directed Secretary Harris to conduct the 2002 elections under this interim remedial plan.^[12]

5. THE REDISTRICTING PROCESS

70. Florida began the redistricting process in 2001. The legislature prepared the technological groundwork for redistricting, held a series of 24 public hearings throughout Florida, and then enacted redistricting legislation. During the 1991-1992 redistricting process, 33 public hearings and at least one statewide teleconference were held throughout Florida.

71. Both the House and Senate established redistricting websites. These websites made the following information available to the public:

• The names and districts of committee members for each of the redistricting committees.

• Every plan submitted by any member of the legislature.

• A committee and council meeting time-line, along with transcripts and meeting packets that were utilized in any of the meetings.

• A public hearings timeline, with information about the specific locations and times of each of the public hearings to be held.

• A public hearings archive including downloadable information following each of the public hearings.

• The constitutional requirements for redistricting Florida's legislative districts under the Florida Constitution.

• News and press releases regarding the redistricting process.

• A "Frequently Asked Questions" ("FAQ") page with information regarding the 2000 census data.

• Florida census data and a 54-page slide presentation providing an overview of Florida census data and trends.

• Information on the FREDS 2000^[13] software, including: an order form, installation instructions, a dataset with Florida redistricting information to be integrated for use with the software, a downloadable update to the software, and a page dedicated to providing tutorial and other information to assist users of the software.

• Information on how to contact the House and Senate Redistricting Committees and their staffs.

72. The times and dates of the public hearings were advertised in newspapers, as well as listed on the redistricting websites, though the timing of the announcements posted on the websites was not coordinated with the timing of the notices advertised in the media. The newspapers in which the public hearings were advertised are the Tallahassee Democrat, Florida Times-Union, Orlando Sentinel, Daytona News-Journal, Tampa Tribune, The Ledger, The Stuart News, Palm Beach Post, Sarasota Herald Tribune, Naples Daily News, Ft. Myers News Press, Miami Herald/El Nuevo Herald, Ft. Lauderdale Sun Sentinel, Pensacola News Journal, and the Gainesville Sun. Most of the newspaper advertisements were published only 2-3 days before the scheduled public hearing.

73. The websites provided hearing times and locations, as well as the protocol and procedure for the hearings. The websites encouraged persons unable to attend the meetings to submit written testimony and plans to the House or Senate Redistricting Committee, and also provided contact information so that members of the public could contact the committees either before or after the public hearings.

74. The legislative leadership did not conduct any public meetings, workshops, or committee meetings to discuss the 2001 public hearing process or the selection of the hearing sites before the schedule was presented to the public and Democratic members of the legislature.

75. The public hearings took place between July 12, 2001, and October 16, 2001, in Tallahassee (July 12 from 10:00 a.m. to 1:15 p.m.), Jacksonville (August 6 from 6:00 p.m. to 8:50 p.m.), Ocala (August 7 from 1:30 p.m. to 4:30 p.m.), Orlando (August 20 from 6:00 p.m. to 11:05 p.m.), Daytona Beach (August 21 from 10:00 a.m. to 12:45 p.m.), Tampa (August 27 from 9:30 a.m. to 1:35 p.m.), Clearwater (August 27 from 5:00 p.m. to 7:05 p.m.), Lakeland (August 28 from 1:30 p.m. to 3:00 p.m.), Vero Beach (September 5 from 9:00 a.m. to 11:20 a.m.), West Palm Beach (September 5 from 4:00 p.m. to 6:40 p.m.), Delray Beach (September 6 from 6:00 p.m. to 9:10 p.m.), Wachula (September 20 from 2:00 p.m. to 3:00 p.m.), Sarasota (September 21 from 9:00 a.m. to 11:00 a.m.), LaBelle (September 24 at 10:00 a.m.), Naples (September 24 from 6:00 p.m. to 8:45 p.m.), Ft. Myers (September 25 at 1:00 p.m.), Pembroke Pines (October 1 at 9:00 a.m.), Ft. Lauderdale (October 1 at 5:00 p.m.), Miami (October 2 at 4:00 p.m.), Miami (October 3 at 3:00 p.m.), Key Largo (October 3 at 10:00 a.m.), Pensacola (October 15 at 6:00 p.m.), Panama City (October 16 at 11:00 a.m.), and Gainesville (October 16 at 6:00 p.m.). Some citizens, including some minority citizens, were unaware of the hearings and/or were unable to attend the hearings in their respective areas because of the time of the hearings. There were no official interpreters at the hearings, and no announcements about interpreters were posted prior to the hearings, but some bilingual legislative staff members were available at the sign-in table at some of the hearings (e.g., the hearings in Miami) to translate in Spanish or Creole if someone requested assistance.

76. After several requests of State Senator Debbie Wasserman-Schultz—a Democratic member of the legislature—a second public hearing was conducted in Broward County. The Republican leadership, however, rejected the requests of several Democratic legislators that additional hearings be held in the southern portion of Miami-Dade County.

77. Only one public hearing was held in a predominantly black community in Miami-Dade County. This hearing was held at Miami Central High School on October 2, 2001, and began at 4:00 p.m.

78. Only one public hearing, convened on October 3, 2001, at 3:00 p.m. at Florida International University, was held in a predominantly Hispanic community in Miami-Dade County. Through Mayor Martinez, the City of Hialeah—which is predominantly Hispanic—offered to host a public hearing free of charge, but no hearing was held in the city.

79. Those public hearings that were held after 5:00 p.m. had the greatest participation of blacks and Hispanics.

80. Approximately 4,500 people attended the 24 public hearings in 2001, three times the number of people that attended the 33 redistricting hearings in 1991-1992. Over 1,000 people spoke at the 2001 public hearings, in contrast to the fewer than 400 people who spoke at the 1992 hearings. In the 1991-1992 cycle, approximately 17 legislators attended each meeting; in the 2001-2002 cycle, approximately 26 legislators attended.

81. The 2001 hearings consisted of an overview of the legal parameters of redistricting, a brief census overview outlining demographic facts in summary form, and public testimony, which included opportunities for legislators to ask clarifying questions. The public, however, was not permitted to pose questions to the legislators, and the legislators were not allowed to ask the public substantive questions.

82. The public hearing transcripts indicate that the appropriate protocol for the hearings was followed, and that such protocol was explained to the public prior to the hearings.

83 Following the public hearings, the redistricting committee websites made the following information available for downloading by the general public (and, of course, any legislators): (1) either an audio or video recording of the hearing; (2) a written transcript of the hearing; (3) the slide presentation made at the hearing; and (4) any public documents made available at the hearings and submissions (e.g., correspondence) from the public.

84. In addition to the public hearings, hundreds of letters and e-mail communications were received from the public.

6. THE FREDS REDISTRICTING SOFTWARE

85. As part of its redistricting efforts, the Florida legislature developed the Florida Redistricting System or "FREDS" software, starting the development process in late 1997 or the beginning of 1998. Development of FREDS was a joint House and Senate project coordinated by John Guthrie, Florida Senate Staff Director for Systems Operations and Training, and Todd Thomson, Florida House Staff Director of the House Redistricting Committee.

86. Mr. Guthrie is a career public servant who has worked for the Florida legislature for 14 years. For the past 13 years, Mr. Guthrie has worked for the Florida Senate, under both Democratic and Republican Senate leadership. Additionally, Mr. Guthrie worked with the Martinez plaintiffs' counsel, Norman Powell, Esq., on behalf of the Democratic majority in the 1991-1992 Florida redistricting, and worked with the Martinez plaintiffs' expert, Dr. Alan Lichtman, in defending Florida's redistricting plans in 1994-1996.

87. In 2001-2002, members of the legislature and the public had access to redistricting information and technology through FREDS. No other state in the country provided affordable redistricting software for personal computers to the general public.

88. The purpose of FREDS was to create a program for drawing and reviewing districting plans that would be easy to use by both the legislature and the public. Though FREDS was initially intended as a tool for the legislature in the redistricting process, it was recognized from the outset that FREDS could be made available to the public.

89. FREDS contained data from the 1990 and 2000 censuses and Florida election results. The census data included population counts broken down by race, including voting-age population, and registration counts broken down by party, gender, age, and race; the election data included actual election results for Florida elections in 1992, 1994, 1996, 1998, and 2000.

90. In 2000, based upon census guidelines, citizens could self identify with respect to race by choosing one or more of six options. To accommodate the 2000 census reporting options the United States Office of Management and Budget ("OMB") and the United States Department of Justice ("DOJ") issued guidelines with respect to allocation of minorities. Specifically, the DOJ guidelines set forth the allocation methodology appropriate for civil rights enforcement. FREDS incorporated the federal allocation guidelines. In July of 2000, the allocation method utilized by FREDS was discussed with Joe Rich, Chief of the Voting Rights Section of the DOJ and received a "positive" reaction. Due to the use of the DOJ and OMB guidelines, FREDS did not have a separate category for "Single-Race Black," though information for such a category was available in the 2000 census. Instead, FREDS had the following population categories: "Non-Hispanic Black;" "Hispanic Black;" "Hispanic;" "Single-Race White;" and "Other."

91. FREDS uses a disaggregation method (an estimation technique) for applying precinct data to census blocks. This disaggregation method used by the legislature—which seeks to take information from a larger geographic level and import the character of that data to a smaller geographic level—is consistent with past practice, and is an appropriate method for applying election results to census blocks in Florida.

92. A potential alternative method for applying voter registration counts to census blocks is known as geocoding. Geocoding involves the use of voter registration data, which includes address information, from the Florida central voter files, to match voters with their addresses and place them within census blocks. A state employee attempted to geocode the central voter file, but he did not know the particulars of which county or counties were selected and what methods were used. Mr. Guthrie asked Florida's Office of Demographic and Economic Research ("ODER") to attempt to geocode the election information to the census block level. The ODER reported, however, that it was unable to geocode "15 to 20 percent of the voters" to their census blocks. Mr. Guthrie therefore decided to use the disaggregation method.

93. In order to disaggregate election information from the precinct level to the census block level, the FREDS development team relied upon an assumption of homogenous distribution of population for every census category within a census block.

94. The FREDS development team did not develop or have an understanding of what the margin of error was for the assumption of homogenous census geography relied upon in developing FREDS.

95. The Martinez plaintiffs' second expert, Dr. Gerald Webster, testified in his deposition that it was not possible to geocode the data. There is no testimony by the Martinez plaintiffs' experts suggesting that disaggregation is not an appropriate methodology. In fact, some of the Martinez plaintiffs' experts used the district level estimates derived using the FREDS disaggregation method.

96. On July 12, 2001, a beta version of FREDS was released to members of the legislature, as well as to the public, for full scale testing. This initial release revealed that FREDS was incompatible with older versions of Microsoft Internet Explorer. FREDS, for example, crashed the City of Hialeah's computer. The necessary Internet Explorer update, however, was available at no cost on the Microsoft website long before the July 12, 2001, beta release of FREDS. In any event, the problem was diagnosed within hours of the release of FREDS, and the method of resolving it was publicized by posting the information on the redistricting websites and directly calling the 20 or so members of the public that had purchased FREDS at that time.

97. The FREDS software costs $20, in contrast to commercially available programs that could be used to draw district lines, such as Maptitude, which costs $3,500 and requires additional expenditures before it can be used for redistricting.

98. Despite its low cost, less than 300 people purchased the FREDS software.

99. The computer requirements for running FREDS are not substantial. At the time of trial, a computer with capabilities exceeding the requirements for running FREDS—a Windows 98 or higher operating system, a 300 megahertz or higher processor, and RAM of 64 megabytes or higher—could be purchased for less than $500.

100. In addition to selling the software for $20, FREDS was made available to the public through distribution to 508 public libraries across the state of Florida, 28 depository libraries, every public and private college library, and supervisors of elections in each county. FREDS was also available at no cost on the redistricting websites.

101. Training on how to use FREDS was made available in a variety of ways. After it was suggested that training be provided at the August 27, 2001, public hearing in Clearwater, assistance and training were available at each public hearing from that date forward for about an hour prior to the start of the hearing. These were not formal training sessions or classes, however, and consisted of Mr. Guthrie and others answering the questions of citizens who dropped by. The software package includes a 138-page help manual and a tutorial. This manual was available for viewing and downloading on the redistricting websites.

102. None of the traditional methods of contacting one's legislator were curtailed or affected by FREDS or the redistricting process used by the legislature. Accordingly, any Floridian with concerns about redistricting was able to visit, telephone, and write to his or her legislator, in addition to the additional methods provided through the distribution of FREDS, the public hearings, and the establishment of the House and Senate redistricting websites.

7. THE REDISTRICTING PLANS

103. The legislature reached out to specific voting rights and public interest groups. On December 21, 2001, Representative Johnnie B. Byrd, Jr., Chairman of the House Procedural and Redistricting Council, sent a letter to the Florida Democratic and Republican Parties, Common Cause, the Florida State Conference of NAACP Branches, the Florida League of Women Voters (both statewide and the Tallahassee chapter), and the American Federation of State, County and Municipal Employees. In the letter, Chairman Byrd invited these organizations to participate in the redistricting process:

In the continuing spirit of having the most fair and open redistricting in Florida's history I would like to personally invite you to our redistricting committee meetings on January 7, 2002 to present your view on the maps that have been proposed thus far and offer any redistricting plans or ideas that your organization has developed. These committee meetings will serve as yet another opportunity for the public to have their voice heard on redistricting matters.

104. On January 7, 2002, the Florida State Conference of NAACP Branches ("NAACP") responded to Chairman Byrd's letter with a memorandum to the House Redistricting Committee, noting that while the "NAACP usually includes alternative redistricting plans" when it presents analysis of redistricting plans being considered by state and local jurisdictions, it was "not necessary" to do so at that point in Florida's redistricting process, based on the plans already proposed.

105. The NAACP further observed that "Florida has done a better job than many states" in its endeavor to comply with section 2 of the Voting Rights Act. The letter also provided the NAACP's analysis of plans under consideration at that time.

106. On January 22, 2002, Representative Mario Diaz-Balart, Chairman of the House Congressional Redistricting Committee, sent another letter to the NAACP, thanking it for the memorandum and inviting NAACP participation in a committee meeting to workshop and discuss the memorandum.

107. There were four House committees that were responsible for drafting and approving the House legislative and congressional plans for full chambers consideration. These were the House Procedural and Redistricting Council ("Procedural"), House Congressional Redistricting Committee ("Congressional"), House Redistricting Committee ("House"), and Senate Redistricting Committee ("Senate") (collectively the "House Committees"). Speaker Feeney appointed each member of the House Committees and each of those committees had at least a majority of Republicans. The partisan break down for each House Committee was as follows: Procedural (13 Republicans and 6 Democrats); House (6 Republicans and 3 Democrats); and Senate (4 Republicans and 3 Democrats). No Democrat chaired or co-chaired a House Committee. No amendment to the House Committees' maps that was introduced by a Democratic committee member or member of the House was ever adopted. No Democratic member's amendment was ever adopted during the House floor debate on the committees' maps.

108. In total, the House Committees (e.g., House, Senate, and Congressional) held 16 meetings between October of 2001 and February of 2002 to discuss the concepts being presented by various plans and any amendments that had been filed.

109. During this time, the legislative leadership announced that the staff of the redistricting committees was available to assist any member of the legislature in understanding demographic, legal, or other relevant issues. All members were encouraged to engage in the process by filing plans and plan amendments, and to utilize the committee staff. Each of the plans filed was posted on the Internet and available for public review and comment.

110. There were few maps proposed by the public that were sponsored by legislative members.

111. Proposed congressional and state redistricting plans were filed for consideration by Representative William F. Andrews (H087H001 and H087H002); Representative Anna Holliday Benson (H003H001 and H003H002); Representative Phillip J. Brutus (H108H001 and H108H002); Representative Byrd, Jr. (H062H001, H062S001, and H062C002); Representative Larry Crow (H049H001 and H049H002); Representative Nancy Detert (H070H002 and H070H004); Representative Greg Evers (H001H001); Representative Frank Farkas (H052C001, H052H001 and H052H002); Representative Andy Gardiner (H040C001 and H040H001); Representative James Harper, Jr. (H084H001); Representative Mike Hogan (H013C001 and H013H003); Representative Lerner (H119H001); Representative Will Kendrick (H010H003 and H010H005); Representative Bev Kilmer (H007H001); Representative Jeffrey D. Kottkamp (H074H001); Representative Mitch Needelman (H031H003 and H031H005); Representative Joe Negron (H082H003); Representative Joe H. Pickens (H021H002); Representative Marco Rubio (H111H001, H111H003, and H111S001); Representative Timothy M. Ryan (H099H001, H099H002, and H099S001); Representative David Simmons (H037H002); Representative Joseph R. Spratt (H077H004); Representative Rob Wallace (H047H001 and H047S001); Senator Alex Diaz de la Portilla (S34S00017 and S34S00031); Senator Holzendorf (S02S00010, S02S00012, S02S00019, S02S00025, S02S00028, and S02S00035); Senator Bill Posey (S15S0002 and S15S00030); Senator Charlie Clary (S07S00033 and S07S00034); Senator Daniel Webster (S12S00027); Senator Wasserman-Schultz (S32S0008); Senator John F. Laurent (S17S0001, S17S0003, S17S0005, S17S0006, S17S0009, S17S00011, S17S00016, S17S00022, S17S00023, S17S00024, S17S00026, S17S00032, and S17S00036); Senator Steven A. Geller (S29S0007, S29S00013, S29S00014, S29S00015, S29S00018, S29S00020, S29S00021, and S29S00029); Representative Edward L. Jennings, Jr. (H023S001); Representative David D. Russell, Jr. (H044S001); Representative Bruce Kyle (H073S001); Representative Christopher L. Smith (H093C001 and H093S001); Senator Jack Latvala (S19C0004, S19C0005, S19C0010, S19C0011, S19C0015, S19C0016, S19C0017); Senator Daryl L. Jones (S40C0006 and S40C0012); Senator Burt L. Saunders (S25C0007); Senator Walter G. Campbell (S33C0009); Senator Ginny Brown-Waite (S10C0013); Senator Lesley Miller, Jr. (S21C0014); Representative J. Dudley Goodlette (H076C001); Representative Meadows (H094C001 and H094C002); Representative Diaz-Balart (H112C001 H112C002, H112C003, and H112C004).

112. Representative Lerner, a Democrat, drafted and participated in submitting an alternative house plan to the legislature that kept former HD 119 relatively unchanged. She withdrew the plan, however, because she was satisfied with the plan submitted by Representative Ryan.

113. The redistricting proceedings in the Senate mirrored the House proceedings, with one exception. The Senate did not draw a Senate legislative House plan. Instead, the Senate informed the House that it was not going to draw House districts, and the Senate expected that the House would not draw Senate districts. As noted above, however, the House drew a Senate map.

114. Three Senate committees were primarily responsible for drafting and approving the Senate legislative and congressional plans for full chambers consideration: Senate Committee on Reapportionment ("Senate Reapportionment"), Senate Reapportionment Subcommittee on Congressional Apportionment and Redistricting ("Senate Congressional"), and Senate Reapportionment Subcommittee on Legislative Apportionment and Redistricting ("Senate Legislative") (collectively "Senate Committees"). President McKay appointed each member of the Senate Committees. Republicans constituted at least a majority of the membership of each of the Senate Committees. As in the House, no amendment introduced by a Democratic committee member was adopted in committee or on the Senate floor.

115. Although the Senate Committees did not conduct simultaneous meetings, the Senate Legislative Committee cancelled three consecutive meetings, on February 28, 2002, March 7, 2002, and March 8, 2002. The Senate Legislative Committee did not post the final maps it would consider until after midnight on the morning of the Senate Legislative Committee meeting of March 12, 2002.

116. At the March 12, 2002 Senate Legislative Committee Meeting, the Chair of the Committee voted against the committee's map and voiced his concern that the public had not received adequate notice and an opportunity to consider the committee's proposed maps. Chairman Webster also voiced his concerns that the start time did not allow constituents enough time to contact the individual senators regarding any concerns they may have with the maps:

I think we told the public that we would have these things posted as soon as we could on the Internet and that people could look at them. And if they had questions, they could either call their members or call us or write us, or do something, get in touch with us in a way that would allow us to respond to whatever it is that they had a concern about. And that didn't happen this time, and so I guess for me, it is just a statement to say, I'm voting no because I feel like the public didn't have enough time on this particular issue. If you have a conscience that's bothering you, then I guess you could join me, but I'm not asking that. I'm just saying mine did, because I'm the one that made the statements, I'm the one that defended the Senate, I'm the one that said we would have an open and an inclusive and a fair process. And I felt like in this one step, we did not.

117. Despite Chairman Webster's concerns, the proposed Senate Legislative Committee bill was eventually approved.

118. Public testimony was also presented to the Senate Legislative Committee at the March 12, 2002 meeting that questioned the Committee's cancellation of its meetings and the adverse effect the cancellations had on the public's ability to participate. For example, the Republican Chairman of Marion County made the following statement:

It is a shame, however, that the public has to be put through what they have had to be put through to get to this point.

Many times, the meetings have been canceled and rescheduled and canceled and rescheduled. The public in many cases has given up on trying to even come and testify any longer. But it is sad process.

119. In total, the Senate Redistricting Committees held seven meetings between January 7, 2002 and March 12, 2002 to discuss the formulation of redistricting plans.

120. Disputes between Speaker Feeney and President McKay led to delays in the redistricting process.

121. When Speaker Feeney and President McKay reached a last-minute resolution of their dispute, the legislature did not refer the Senate's proposed Senate map to a House Committee or otherwise consider the map on the House floor. The Senate also did not refer the House proposed version of the legislative House map to a Senate committee or otherwise consider the House legislative districts.

122. Once the Senate adopted the Senate legislative map and substituted the same in the House proposed resolution (HJR 1987) for the Senate map, the House considered the revised resolution as a message from the Senate on a motion of Representative Goodlette, a Republican. Thereafter, Republican Representative Byrd moved that the House concur in Senate Amendment 3 under House Rule 12.6(a)(2). Prior to any debate, questions, or discussion of the Senate amendment, Representative Rubio, a Republican, moved the previous question on the motion to concur in the Senate amendment. The motion passed 74-43. All Democrats present voted no. Two Republicans also voted no. Under House Rule 11.8, a motion for the previous question is not debatable and a majority vote will end all debate, except that the sponsor and opponent of the bill are allowed three minutes of debate each. The proponent and opponent are also allowed to yield time to other members. In this case, Speaker Feeney selected two Republicans to debate the resolution, Representative Kyle and Representative Byrd. All debate on the resolution, therefore, was limited to two Republican members and no Democrat was afforded an opportunity to debate. A final vote was taken on the motion to concur in the Senate amendment, which passed 73-43. All Democrats voted no along with Representative Kyle.

123. The Republican-controlled legislature intended to maximize the number of Republican congressional and legislative seats through the redistricting process, and used its majority power to control the types of bills and maps that would be considered in the House and Senate.

124. One of the most complicated and significant issues the legislature faced was the result of the need to create an additional congressional district in south Florida —one of the two additional districts to which Florida was entitled as a result of the 2000 census—combined with the slow growth and under-population in several of the existing south Florida congressional districts (e.g., old CD 17). This problem was specifically addressed during the House floor debate on proposed congressional plans. As Representative Diaz-Balart put it: "There were significant changes in other south Florida districts largely due, Mr. Speaker, to the creation of the new district [CD 25], and also due to the slow growth in Districts 17, 18, 22, and 23 that required additional areas to be added to these districts."

C. ULTIMATE FINDINGS ON THE "PROCESS" CLAIMS

We find that the Martinez plaintiffs have failed to carry their burden of showing intentional discrimination against blacks, Hispanics, or any other minority groups on their "process" claims.

The Martinez plaintiffs contend that the Republican legislative leadership did not follow customary legislative protocol in devising Florida's public hearing schedule. In particular, they claim that no public meetings or legislative committee discussions were held concerning the proposed hearings or for the receipt of public input concerning the redistricting process. Furthermore, they point out, when the hearing schedule was presented to the Democratic members of the legislature and the public, the legislative leadership refused to answer questions regarding the selection of the hearing locations and start times. In short, the Martinez plaintiffs claim that all decisions regarding the redistricting hearing process were predetermined or made behind closed doors, and that the plans were therefore drawn devoid of public input and scrutiny.

The Martinez plaintiffs also assert that the public hearing process had a disproportionate impact on the state's blacks and other minorities. They maintain that these groups traditionally hold jobs that require daily attendance during working hours, and that 18 of the 24 hearings were held at some point during the day between normal business hours. In contrast, only four public hearings began after 6:00 p.m. They further argue that the public hearing schedule avoided high population cities where the vast majority of Florida's black and minority communities reside. Only one hearing was held in a predominately black community, and only one hearing conducted in a predominately Hispanic community. Moreover, the Martinez plaintiffs say that the notices that were provided to the public were inadequate and frequently provided too close in time to the meeting dates to provide the public with an opportunity to arrange for their participation.

At a general level, we share some of the concerns articulated by the Martinez plaintiffs in their "process" claims. First, although the Republican leadership of the legislature did not prevent anyone from presenting their views on redistricting, held numerous hearings throughout the state on redistricting which large numbers of people attended, and created a software program—FREDS—that was economical and available to the public, it appears that the leadership also wanted to limit the actual substantive debate on the merits of the various plans that were proposed, and did not much care about input from the Democratic minority or the public on the redistricting process or the plans themselves. This raw exercise of majority legislative power does not seem to be the best way of conducting a critical task like redistricting, but it does seem to be an unfortunate fact of political life around the country. Second, as the Martinez plaintiffs point out, it might have been preferable to provide more than two to five days' notice in newspapers to the public about upcoming hearings rather than rely on internet websites.

Other concerns voiced by the Martinez plaintiffs, however, are not well-founded. For example, the claim that choosing FREDS over other available redistricting software was somehow discriminatory is hollow. FREDS was much less expensive than Maptitude and therefore more accessible to the public. Similarly, the allocation of minority groups in FREDS was not motivated by any racial animus, as that allocation was based on a methodology that was consistent with DOJ and OMB guidelines. Finally, the use of a disaggregation method over geocoding was not discriminatory in any way.

In any event, our role is not to determine whether the procedural choices made by the legislature were the best among a range of options, whether the legislative action had to reflect the testimony received at the public hearings, or whether the legislature was required to consider that testimony or input during its redistricting deliberations.^[14] Our task, instead, is to determine whether the legislature's actions violated federal law, and our finding is they did not. The Martinez plaintiffs failed to present any evidence to show that the defendants were motivated by discrimination against blacks or Hispanics in deciding what redistricting software or allocation methodology to use, where to hold the public hearings, when to hold public hearings, what type of notice to provide, or whether to consider input from Florida citizens and Democratic legislators in drawing the redistricting plans.^[15]

II. VOTE DILUTION BASED ON RACE

After having addressed plaintiffs' process claims, we now turn to the allegations of racial vote dilution. For purposes of this litigation, the parties do not dispute that section 2 requires the creation of performing minority districts and that the number created by the Florida legislature is sufficient to satisfy that obligation. Rather, the plaintiffs claim that the reapportionment plan used by the Florida legislature results in the dilution of black voting power,^[16] and hence a violation of the Voting Rights Act, because the newly drawn performing districts are either not sufficiently likely to perform or are excessively packed so as to "waste" minority votes.

To evaluate these allegations concerning the composition of the performing minority districts—a necessarily fact-intensive inquiry —the bulk of this part is dedicated to setting forth our findings of fact with respect to racial voting behavior, the Florida redistricting process, and the makeup and voting history of each of the challenged districts. In subpart II.B. below, we conclude that the evidence does not support a finding that the strength of black voters will be negatively impacted by the reapportionment. The evidence suggests that black candidates of choice will prevail in a very high percentage of elections in the challenged districts and that none of the districts are impermissibly stacked.

A. FINDINGS OF FACT

1. GENERAL FINDINGS: VOTING BEHAVIOR BY RACE

125. There is a substantial degree of racially polarized voting in south Florida and northeast Florida—the areas of the state involved in plaintiffs' claims of racial vote dilution. A substantial majority of black voters in these areas sometimes favors candidates that most other voters do not favor. A substantial majority of non-black voters sometimes opposes black candidates of choice.

126. In any district in these areas in which blacks do not comprise a majority or near majority of actual voters, it is likely that the black candidate of choice (if different from the candidate of choice of non-black voters) will not often prevail; instead, the candidate of choice of non-black voters will prevail. Black candidates of choice often will prevail, however, in a district in which blacks comprise a near majority of actual voters, even when they do not comprise a majority, because a significant number of non-black voters support black candidates of choice.

127. Voting behavior by race for any district in these geographic areas cannot be projected with certainty. The most constant characteristic of the Florida population is that it constantly changes; voters in any district come and go. This is especially so in south Florida. By the time of the 2002 elections, the demographics of any district will differ from those reflected in the 2000 census. By 2008 or 2010, the differences will be substantial. Projections of 2002 election outcomes based on 2000 census data are uncertain. Projections of 2008 or 2010 election outcomes based on 2000 census data are even more uncertain.

128. Projections based on prior voting behavior also are uncertain. Although voting turnout and preference by race for the geographic area covered by a district can reasonably be estimated for prior elections, the movement of voters into and out of the area necessarily means the prior votes were cast by a different set of voters than will vote in any future election. And even those voters who stay in a district do not always vote as they did previously.

129. More fundamentally, neither voters nor candidates are fungible, whether within or without the same racial or ethnic groups. Turnout and voter preference vary not only with such factors as the type of election and race of the candidates, but also with overall political trends and with the particular issues and candidates involved in a particular election.

130. For the districts at issue in this case, it is impossible to project with certainty the voter turnout and preference by race that will occur during the coming decade.^[17] Nonetheless, reasonable projections can be made. In making such projections, all relevant circumstances and data properly are considered; no one type of analysis or set of data is sufficiently reliable to justify its use standing alone. Both demographics and past voting performance are relevant.

2. CONGRESSIONAL PLAN

a. BACKGROUND

131. During each recent decade, Florida's population has grown substantially, resulting in an increase in Florida's number of representatives in the United States House of Representatives.

132. As a result of the 1990 census, Florida's number of representatives increased from 19 to 23. The Florida legislature was unable to agree on an apportionment plan. Florida law provided no method for adopting a plan in the absence of action by the Florida legislature. Because of increases and shifts in population, the preexisting plan as adopted in 1982 was malapportioned, in violation of the one person, one vote principle. Litigation in the United States District Court for the Northern District of Florida led to the implementation of a plan drawn by a court-appointed expert based on the litigants' submissions.

133. The court-drawn plan included three districts (numbered 3, 17, and 23) that were projected to perform for a black candidate of choice and two districts (numbered 18 and 21) that were projected to perform for an Hispanic candidate of choice. Districts that are projected to perform for minority candidates of choice are sometimes referred to in this order as "minority districts."

134. Each of these five minority districts performed as anticipated. In 1992, voters elected Corrine Brown (in CD 3), Carrie Meek (in CD 17), and Alcee Hastings (in CD 23), each of whom is black, and Ileana Ros-Lehtinen (in CD 18) and Lincoln Diaz-Balart (in CD 21), each of whom is Hispanic. Each was the candidate of choice of the relevant minority in the district. Each was reelected in 1994, 1996,^[18] 1998, and 2000, and continues to serve.^[19]

135. As a result of the 2000 census, Florida's number of representatives in the United States House of Representatives increased from 23 to 25. The Florida legislature was controlled by a substantial Republican majority in each house. The legislature adopted a congressional reapportionment plan on March 22, 2002, with 25 single member districts. Governor Bush, also a Republican, signed the plan into law on March 27, 2002.

b. ULTIMATE FINDINGS^[20]

136. The legislature's overriding goal with respect to congressional reapportionment was to adopt a plan that would withstand legal challenges and would maximize the number of districts likely to perform for Republicans. The legislature had no purpose to discriminate against black or Hispanic candidates or voters. To the contrary, the legislature affirmatively intended not to discriminate against black or Hispanic candidates or voters. The legislature recognized that adopting districts that would not dilute black or Hispanic voting strength was fully consistent with— perhaps even essential to—the goal of adopting a plan that would withstand legal challenges.

137. The legislature's plan included three districts (still numbered 3, 17, and 23) intended to perform for black candidates of choice.

138. It is likely that CDs 3, 17, and 23 will in fact perform for black candidates of choice, and that black candidates of choice will be elected in these districts throughout the coming decade.

139. The legislature determined, correctly, that increases in the Hispanic population made it feasible to draw a third reasonably compact district that would perform for Hispanic candidates of choice. The legislature's plan included three districts (two still numbered 18 and 21 and a new district numbered 25) intended to perform for Hispanic candidates of choice.

140. It is likely that CDs 18, 21, and 25 will in fact perform for Hispanic candidates of choice, and that Hispanic candidates of choice will be elected in these districts throughout the coming decade.

141. Race was considered in the drawing of CDs 3, 17, and 23, and Hispanic status was considered in the drawing of CDs 18, 21, and 25, in an effort to enhance performance of the districts for black or Hispanic candidates of choice, to avoid dilution of the votes of black or Hispanic voters, and to comply with section 2 of the Voting Rights Act of 1965. Traditional districting principles also were considered and respected in the drawing of these districts. No effort was made to afford blacks or Hispanics substantially greater than proportional representation or to afford them representation inconsistent with traditional districting principles. Neither race nor Hispanic status was the controlling or predominant factor in the drawing of such districts.

142. CDs 3, 17, 23, 18, 21, and 25 are reasonably compact.

143. CDs 3, 17, 23, 18, 21, and 25 in fact comply to a reasonable extent with traditional districting principles.

144. No greater number of reasonably compact performing minority districts reasonably could have been drawn.

c. CENSUS DATA^[21]

145. The 1990 census included as a racial category "Black or Negro." The 2000 census included as a racial category "Black, African Am. or Negro." In this order, we use "black" as shorthand for this category. See supra note 2.

146. In the 2000 census, respondents were required to self identify as at least one race; they were free to self identify as more than one race. Various races (including "White," "Black, African Am., or Negro," "American Indian or Alaska Native," and various Asian nationalities) were listed, as was a category "Some other race—Print race." In the 1990 census, respondents were required to self identify as one and only one race; various races (substantively the same as for 2000) were listed, as was a category "Other race (Print race)." In both the 1990 and 2000 censuses, race was a wholly separate category from Hispanic status; regardless of a respondent's racial self-identification, the respondent was required to self identify as either Hispanic or not Hispanic.^[22]

147. It is likely that some individuals who identified themselves as both black and another race in 2000 identified themselves in the single category black in 1990. It also is likely that some individuals who identified themselves as both black and another race in 2000 did not identify themselves as black in 1990 but instead listed themselves as another race or as "Other race." So far as this record reflects or we are aware, there are no reliable studies of the effect of the change from the 1990 single race response requirement to the 2000 allowance of multiple race responses. It is unlikely that the change materially affects the analysis relevant to this case, because the number of persons counted as black under the 2000 methodology who would not be counted as black under the 1990 methodology probably is small. The discrimination of primary concern for present purposes is discrimination (that is, vote dilution) against blacks; it seems unlikely that such discrimination would be directed to single race blacks but not multi-race blacks, or vice versa.^[23]

d. CONGRESSIONAL DISTRICT 23

148. According to the 2000 census, the non-Hispanic black population of new CD 23 was 54.6%, and the Hispanic black population was 1%, for a total black population of 55.6%. The non-Hispanic black voting age population was 49.3%, and the Hispanic black voting age population was 0.8%, for a total black voting age population of 50.1%.

149. Voting age population, not total population, is relevant to assessing potential voting strength as of 2000, but some of those who were underage in 2000 already have come of age, and others of course will do so in ensuing years. Other things being equal, the 50.1% black voting age population will increase during the coming decade.^[24]

150. Even with only a 50.1% black voting age population, the district will afford black voters a reasonable opportunity to elect candidates of choice and probably will in fact perform for black candidates of choice.

151. This is so in part because, according to 2000 data, blacks constitute 59.2% of registered Democrats in new CD 23, and Democrats constitute 65.0% of registered voters. Republicans constitute only 19.5% of registered voters. Actual voting also is overwhelmingly Democratic; in the 2000 presidential election, voters in new CD 23 voted 78.5% for Democratic candidate Al Gore and 19.9% for Republican candidate George W. Bush. The black candidate of choice is likely to win a contested Democratic primary, and the Democratic nominee is likely to win the general election.^[25]

152. The history of prior CD 23— which had different boundaries but the same core and similar demographics as new CD 23—suggests that a 55.6% total black population and 50.1% black voting age population are sufficient to cause the district to perform for black candidates of choice. When first drawn in 1992,^[26] prior CD 23 had (according to the 1990 census) a non-Hispanic black population of 50.2%, Hispanic black population of 1.5%, and total black population of 51.7%, nearly four percentage points lower than the total black population of newly created CD 23 under the 2000 census. Similarly, when first drawn in 1992, prior CD 23 had (according to the 1990 census) a non-Hispanic black voting age population of 44.4%, Hispanic black voting age population of 1.4%, and total black voting age population of 45.8%, more than four points lower than new CD 23 has under the 2000 census.

153. The black percentage of total population in prior CD 23 increased throughout the decade of the 1990s, as did the black percentage of voting age population. By the time of the 2000 census, prior CD 23 had black total population of 58.7% and black voting age population of 53.9%— higher than new CD 23—but there is no reason to believe that percentages that high were necessary for the district to perform for black candidates of choice.

154. Prior CD 23 performed for the black candidate of choice in every election from 1992 through 2000. New CD 23 has demographics more favorable to black candidates of choice than did prior CD 23 when it was created. The demographics of new CD 23 are likely to become more favorable for black candidates of choice as time passes, just as occurred in prior CD 23. There is no reason to believe racial polarization has changed materially since 1992. There is no reason to believe that new CD 23 will perform less favorably for black candidates of choice in the coming decade than did old CD 23 in the prior decade.^[27]

155. Techniques of ecological regression and projection of future voting behavior that are reasonably reliable and accepted in the field of political science support the conclusion that CD 23 will perform for black candidates of choice. Such techniques require choosing past elections on which ecological regression is performed and making assumptions and choosing data sets for use in projecting the results of future elections. There is no single reliable or accepted approach for choosing past elections or choosing data sets for projecting future elections. The parties' experts have performed their analyses using various past elections and various assumptions and data sets for projecting future elections; they could have performed the analysis in many additional ways as well. No purpose would be served by addressing all the analyses that have been or could be done for CD 23. We address only so much of the analysis as necessary to explain why, even if plaintiffs' data are accepted, it is still clear that CD 23 can be expected to perform for black candidates of choice.

156. There have been two changes in Florida law in the last decade that affect the analysis. First, in elections prior to and during 1998, political party primaries always were "closed," that is, only registered members of a political party could vote in that party's primaries. Effective January 5, 1999, the Florida Constitution was amended to provide that, when the only candidates for an office are from the same party, so that that party's nominee will have no opposition in the general election, the party's primary is "open," that is, all voters may participate in the primary, regardless of party registration.^[28] The Florida Secretary of State has interpreted this provision to mean that a primary is "open" only if there are no other candidates who have qualified to have their names on the general election ballot and also no qualified write-in candidates.^[29] The interpretation of the Secretary of State has not, so far as we are aware, been challenged; it has been followed in all elections since the "open primary" provision was adopted. Nor, so far as we are aware, has the Florida "open primary" provision been challenged.^[30]

157. The second change in Florida law occurred in 2001. Prior to that time, there was a "majority vote requirement" in party primaries. If no candidate received a majority of the votes in a primary, there was a "runoff" or "second primary" between the two candidates who received the highest number of votes in the first primary. In 2001, the Florida legislature eliminated the majority vote requirement; there is now a single primary, with the candidate who receives the most votes, even if less than a majority, advancing to the general election. The new provision sunsets, however, on January 1, 2004. After that date, there will again be a majority vote requirement, unless the legislature takes action to the contrary.^[31]

158. It is clear that, in a closed Democratic primary in which only a single candidate has substantial support among black voters, CD 23 is likely to perform for that candidate, as against a single opposing candidate or indeed as against all comers. The parties' analyses consistently have projected a vote in excess of 50% for the black candidate of choice in such an election. Plaintiffs have presented no projections to the contrary. Further, the likelihood is overwhelming that the Democratic primary winner will prevail in any general election; plaintiffs concede as much.

159. The virtually uncontested findings set forth in paragraph 158, without more, make it likely that CD 23 will perform for black candidates of choice, because if ever the ability of CD 23 to elect a black candidate of choice is in jeopardy, it is likely that the Democratic primary will be closed (see infra paragraph 160), and it is likely that there will be a single candidate with substantial support among black voters (see infra paragraphs 161-162).

160. There have been 115 congressional elections in Florida since adoption of the 1992 congressional reapportionment plan—five in each of the 23 districts. Only two of those 115 elections would have had an "open" primary if the new Florida open primary provision had been in effect. Both were Republican primaries, one in 1992 and one in 1994. Thus, the conditions that would lead to an open Democratic primary have not occurred in at least a decade. The likelihood that any Democratic primary in any district will be "open" in the next decade is small. Moreover, it is likely that, in any election in which the ability of CD 23 to elect a black candidate of choice is in jeopardy because of the possibility that the primary will be open, a write-in candidate will qualify for the general election, thus closing the primary.^[32]

161. It is possible that in a given Democratic primary election in CD 23, there could be two or more candidates with substantial support among black voters. It is possible that, even though such candidates in combination would receive a majority of all votes cast in the primary, another candidate would win the primary, because of the elimination of the majority vote requirement. It is likely, however, that in any election in which the ability of CD 23 to elect a black candidate of choice would be in jeopardy for this reason, a single candidate favored by the black community would emerge, either as a result of withdrawal of competing candidates with support in the black community, or by coalescence of black voters behind the single candidate. Even if that did not occur, this would indicate not that black voters had not had a reasonable opportunity to elect a candidate of their choice, but only that, in a particular instance, that opportunity was not realized.

162. In any event, elimination of the majority vote requirement can make a difference only if there are three or more candidates in the primary. The possibility is more theoretical than real that there will be any such election in CD 23 prior to reinstatement of the majority vote requirement. Mr. Hastings, the incumbent, has drawn no primary opposition at all since 1992. The possibility that in 2002 there would be three candidates in the Democratic primary—including a candidate in addition to Mr. Hastings who has substantial support in the black community—was remote. The majority vote requirement will return in 2004 unless the legislature takes action to the contrary.

163. Even the analysis most favorable to plaintiffs fails to undermine the conclusion that CD 23 is likely to perform for black candidates of choice. Plaintiffs posit that if there is an open Democratic primary with Republicans and independents voting, and if the turnout of black voters as a percentage of overall voters is only the same as in the 1998 general election, and if black voters are only as cohesive as in a mix of low-visibility, sometimes nonpartisan elections, and if there is no incumbency effect, then a black candidate of choice would poll only 48.0% of the vote, and thus would lose. See Pl. Exh. 81-C at 5, Table 3. In fact, however, if the ability of CD 23 to elect a black candidate of choice is in jeopardy, there is not likely to be an open primary,^[33] black turnout relative to non-black turnout may well increase,^[34] and black cohesiveness may well be higher than plaintiffs suggest.^[35] Thus, even without an incumbent, the projection of a 48.0% vote for a black candidate of choice is unrealistically low. That the projected vote comes this close to a majority, even on these assumptions, is fully consistent with the conclusion that a black candidate of choice is likely to prevail, and that in all events blacks in CD 23 will have a reasonable opportunity to elect candidates of their choice.

164. Finally, the effects of incumbency provide further substantial support for the conclusion that CD 23 is likely to perform for a black candidate of choice. In each election from 1994 through 2000, Mr. Hastings was reelected; no election was close, and twice he had no opposition at all.^[36] Mr. Hastings now apparently is the candidate of choice of most of the district's black and non-black voters alike. New CD 23, while comprised of a different set of voters, has similar demographics and is likely to perform for the black candidate of choice, with or without an incumbent. Absent a significant change in circumstances, the district is nearly certain to perform in any election involving Mr. Hastings, the current incumbent black candidate of choice. Plaintiffs apparently concede this.^[37]

165. In sum, CD 23 is likely to perform for black candidates of choice. The votes of blacks in CD 23 have not been diluted.

e. CONGRESSIONAL DISTRICT 3

166. According to the 2000 census, the non-Hispanic black population of new CD 3 was 50.6%, and the Hispanic black population was 0.8%, for a total black population of 51.4%. The non-Hispanic black voting age population was 46.2%, and the Hispanic black voting age population was 0.7%, for a total black voting age population of 46.9%. Other things being equal, the 46.9% black voting age population will increase during the coming decade.^[38]

167. Even with only a 46.9% black voting age population, new CD 3 will afford black voters a reasonable opportunity to elect candidates of choice and probably will in fact perform for black candidates of choice.

168. This is so in part because, based on 2000 data, blacks constitute 61.3% of registered Democrats in new CD 3, and Democrats constitute 63.8% of registered voters. Republicans constitute only 22.7% of registered voters. Actual voting also is strongly Democratic; in the 2000 presidential election, voters in new CD 3 voted 63.7% for Mr. Gore and 34.2% for Mr. Bush. The black candidate of choice is likely to win a contested Democratic primary, and the Democratic nominee is likely to win the general election.

169. The history of prior CD 3—which had different boundaries but the same core and similar demographics as new CD 3—is consistent with the finding that a 51.4% total black population and 46.9% black voting age population are sufficient to cause the district to perform for a black candidate of choice. CD 3 as drawn in 1992 had a greater black population than as redrawn in 1996. See supra note 18. CD 3 as redrawn in 1996 covered an area that, according to the 1990 census, had a non-Hispanic black population of 46.5%, Hispanic black population of 0.5%, and total black population of 47.0%, more than four percentage points lower than the total black population of newly created CD 3 under the 2000 census. Similarly, as redrawn in 1996, prior CD 3 covered an area that, according to the 1990 census, had a non-Hispanic black voting age population of 42.3%, Hispanic black voting age population of 0.4%, and total black voting age population of 42.7%, again more than four points lower than new CD 3 under the 2000 census.

170. By the time of the 2000 census, prior CD 3 as redrawn in 1996 had a total black population of 50.8% and a total black voting age population of 46.7%—comparable to, but slightly lower than, new CD 3.

171. Prior CD 3 performed for the black candidate of choice in every election from 1992 through 2000. New CD 3 has demographics more favorable to black candidates of choice than did prior CD 3 when it was redrawn in 1996 and indeed more favorable to black candidates of choice than that version of CD 3 would have today. The demographics of new CD 3 are likely to become more favorable for black candidates of choice as time passes, just as occurred in prior CD 3. There is no reason to believe that new CD 3 will perform less favorably for black candidates of choice in the coming decade than did prior CD 3 in 1996, 1998, and 2000.

172. Techniques of ecological regression and projection of future voting behavior that are reasonably reliable and accepted in the field of political science support the conclusion that CD 3 will perform for black candidates of choice. It is clear that, in a closed Democratic primary in which only a single candidate has substantial support among black voters, CD 3 is likely to perform for that candidate. None of the parties' analyses have projected a vote below 50% for the black candidate of choice in such an election. Further, it is likely that the Democratic primary winner will prevail in any general election; plaintiffs concede as much.

173. The virtually uncontested findings set forth in paragraph 172, without more, make it likely that CD 3 will perform for black candidates of choice, because if ever the ability of CD 3 to elect a black candidate of choice is in jeopardy, it is likely that the Democratic primary will be closed (see supra paragraph 160), and it is likely that there will be a single black candidate of choice (see supra paragraphs 161-162).^[39]

174. Ms. Brown is a 10-year incumbent in CD 3. Her presence is not necessary to, but provides further support for, the conclusion that CD 3 is likely to perform for the black candidate of choice. See supra paragraph 164.

175. There is no evidence that any constitutionally-permissible alternative to CD 3 could be drawn that would meaningfully improve the projected performance of this district for black candidates of choice. Plaintiffs have virtually conceded this.

176. In sum, CD 3 is likely to perform for black candidates of choice. The votes of blacks in CD 3 have not been diluted.

f. CONGRESSIONAL DISTRICT 17

177. According to the 2000 census, the non-Hispanic black population of new CD 17 was 57.9%, and the Hispanic black population was 2.2%, for a total black population of 60.1%. The non-Hispanic black voting age population was 53.9%, and the Hispanic black voting age population was 2.1%, for a total black voting age population of 56.0%. Other things being equal, the 56.0% black voting age population will increase during the coming decade.^[40]

178. Even with a 56.0% black voting age population, new CD 17 will afford black voters a reasonable opportunity to elect candidates of choice and probably will in fact perform for black candidates of choice.

179. This is so in part because, according to 2000 data, blacks constitute 65.9% of registered Democrats in new CD 17, and Democrats constitute 71.5% of registered voters. Republicans constitute only 14.6% of registered voters. Actual voting also is overwhelmingly Democratic; in the 2000 presidential election, voters in new CD 17 voted 83.9% for Mr. Gore and 15.1% for Mr. Bush. A black candidate of choice is likely to win a contested Democratic primary, and the Democratic nominee is likely to win the general election.

180. The history of prior CD 17— which had different boundaries but the same core and similar demographics as the new CD 17—suggests that a 60.1% total black population and 56.0% black voting age population are sufficient to cause the district to perform for black candidates of choice. When first drawn in 1992,^[41] prior CD 17 had (according to the 1990 census) a non-Hispanic black population of 55.7%, Hispanic black population of 2.7%, and total black population of 58.4%, nearly two percentage points lower than the total black population of newly created CD 17 under the 2000 census. Similarly, when first drawn in 1992, prior CD 17 had (according to the 1990 census) a non-Hispanic black voting age population of 51.3%, Hispanic black voting age population of 2.7%, and total black voting age population of 54.0%, two points lower than new CD 17 under the 2000 census.

181. The percentage black voting age population and black total population in prior CD 17 increased during the decade of the 1990s. By the time of the 2000 census, prior CD 17 had a total black population of 63.4% and a total black voting age population of 60.0%—higher than new CD 17—but there is no reason to believe that percentages that high were necessary for the district to perform for black candidates of choice.

182. Prior CD 17 performed for the black candidate of choice in every election from 1992 through 2000. New CD 17 has demographics more favorable to black candidates of choice than did prior CD 17 when it was created. The demographics of new CD 17 are likely to become more favorable for black candidates of choice as time passes, just as occurred in prior CD 17. There is no reason to believe that new CD 17 will perform less favorably for black candidates of choice in the coming decade than did prior CD 17 in the preceding decade.

183. Techniques of ecological regression and projection of future voting behavior that are reasonably reliable and accepted in the field of political science support the conclusion that CD 17 will perform for black candidates of choice. It is clear that, in a closed Democratic primary in which only a single candidate has substantial support among black voters, CD 17 is likely to perform for that candidate. Indeed, even in an open Democratic primary, plaintiffs' own projections show a black candidate of choice prevailing with 54.1% of the vote. See Pl. Exh. 81-C at 15, Table 6. Further, it is likely that the Democratic primary winner will prevail in any general election; plaintiffs concede as much.

184. There is no evidence that any constitutionally-permissible alternative to CD 17 could be drawn that would meaningfully improve the projected performance of this district for a black candidate of choice. Plaintiffs have virtually conceded this.

185. In sum, CD 17 is likely to perform for black candidates of choice. The votes of blacks in CD 17 have not been diluted.

g. PROPORTIONALITY

186. According to the 2000 census, the voting age population of Florida's eight southernmost counties, which reasonably could be denominated south Florida, was 16.3% black. The population of those eight counties was sufficient for 8.9 congressional districts. Two districts constitute 22% of those 8.9 districts. CDs 17 and 23 are two of the districts in this eight-county area.^[42]

187. According to the 2000 census, the voting age population of 16 counties in northeast Florida was 14.7% black. The population of those 16 counties was sufficient for 5.8 congressional districts. One district constitutes 17% of those 5.8 districts. CD 3 is one of the districts in this 16-county area.^[43]

3. STATE LEGISLATURE

a. BACKGROUND

188. Florida's House of Representatives has 120 members and its Senate has 40.

189. In 1992, the Florida legislature adopted apportionment plans for the House and Senate based on the 1990 census. The United States Department of Justice pre-cleared the House plan but objected to the Senate plan as it impacted Hillsborough County.^[44] The Governor and legislative leaders chose not to call a special session to address the Department of Justice objections. The Florida Supreme Court redrew the Senate plan to address the objections. See In re Constitutionality of Senate Joint Resolution 2G, Special Apportionment Session 1992, 601 So.2d 543 (Fla.1992). The Department of Justice pre-cleared the redrawn plan.

190. The House and Senate plans were challenged before a three-judge panel of the United States District Court for the Northern District of Florida. See De Grandy v. Wetherell, 815 F.Supp. 1550 (N.D.Fla.1992).

191. The House plan was challenged on grounds of alleged dilution of black votes in Escambia County (the westernmost county in the Florida panhandle) and dilution of Hispanic votes in south Florida. The court upheld a settlement agreement addressing the Escambia claims, and the affected districts were redrawn. With respect to the south Florida claims, the court held that the plan impermissibly diluted Hispanic votes, and the court adopted an alternative plan. Id. The United States Supreme Court stayed the court's decision with respect to south Florida districts and ultimately reversed, see Johnson v. De Grandy, 512 U.S. 997, 114 S.Ct. 2647, 129 L.Ed.2d 775 (1994), thus leaving intact the House plan as adopted in 1992 by the Florida legislature (except for the changes in Escambia County).

192. The Senate plan was challenged on the grounds of alleged dilution of both black and Hispanic votes in south Florida. The district court concluded that the Senate plan diluted both black and Hispanic votes in south Florida but that any remedy favoring one of these minorities would further dilute votes of the other. The court left the Senate plan intact as an acceptable resolution of the competing interests of the two groups. The Supreme Court affirmed but on different grounds; it found no dilution. Id. The plan as drawn by the Florida Supreme Court thus was in effect for the 1992 and 1994 elections.

193. A separate lawsuit before a three-judge panel of the United States District Court for the Middle District of Florida asserted a claim of racial gerrymandering in Senate District 21 (in the Tampa area) in violation of Shaw v. Reno, 509 U.S. 630, 113 S.Ct. 2816, 125 L.Ed.2d 511 (1993). The court approved a settlement agreement under which SD 21 was redrawn. See Scott v. Department of Justice, 920 F.Supp. 1248 (M.D.Fla.1996).^[45] The Supreme Court affirmed. Lawyer v. Department of Justice, 521 U.S. 567, 117 S.Ct. 2186, 138 L.Ed.2d 669 (1997). The plan as redrawn was in effect for the 1996, 1998, and 2000 elections.

194. By 2000, the House and Senate plans as then in effect had become substantially malapportioned. On March 22, 2002, the Florida legislature adopted new House and Senate plans by joint resolution. Under the Florida Constitution, the Governor's approval was not required. See Fla. Const. Art. III § 16(a).

195. Under a procedure established by the Florida Constitution, see Fla. Const. Art. III § 16(c), the Florida Supreme Court reviewed the plans for facial compliance with the state and federal Constitutions and the Voting Rights Act of 1965. As noted earlier, the court held the plans valid. See In re Constitutionality of House Joint Resolution 1987, 817 So.2d 819 (Fla.2002). The court explicitly declined to address as-applied challenges of the type at issue in the case at bar. By its terms, the Florida Supreme Court's decision did not foreclose or express any opinion on claims such as those at issue here.

b. ULTIMATE FINDINGS

196. The legislature's overriding goal with respect to state legislative reapportionment in 2002 was to adopt plans that would withstand legal challenges and would maximize the number of districts likely to perform for Republicans. The legislature had no purpose to discriminate against black or Hispanic candidates or voters. To the contrary, the legislature affirmatively intended not to discriminate against black or Hispanic candidates or voters. The legislature recognized that adopting districts that would not dilute black or Hispanic voting strength was fully consistent with—perhaps even compelled by—the goal of adopting plans that would withstand legal challenges.

197. The Senate and House plans included districts ostensibly intended to perform for black or Hispanic candidates of choice. Plaintiffs do not challenge the overall number of ostensibly performing black or Hispanic districts. Plaintiffs do not challenge the compactness or configuration of any of the ostensibly performing black or Hispanic districts. Plaintiffs do not challenge the sufficiency of any of the Hispanic districts to perform for Hispanic candidates of choice. Plaintiffs do challenge the sufficiency of certain ostensibly performing black districts.

198. Among ostensibly performing black Senate districts, plaintiffs take issue only with SDs 1 and 18, which are successors to prior SDs 2 and 21. Plaintiffs concede, however, that there is no evidence that any constitutionally-permissible alternatives to SDs 1 or 18 could be drawn that would meaningfully improve the projected performance of these districts for black candidates of choice. Plaintiffs thus seek no relief with respect to the alleged dilution of black votes in these districts.

199. Among ostensibly performing black House districts, plaintiffs take issue only with HDs 103, 104, 108, 109, and 118 (all located in Miami-Dade County) and HDs 15, 39, and 94 (located outside Miami-Dade County). Plaintiffs concede, however, that there is no evidence that any constitutionally-permissible alternatives to HDs 15, 39, or 94 could be drawn that would meaningfully improve the projected performance of these districts for black candidates of choice. Plaintiffs thus seek no relief with respect to the alleged dilution of black votes in these districts.

200. Plaintiffs assert that three of the Miami-Dade County districts (HDs 104, 108, and 118) will not perform for black candidates of choice, and that the other two challenged Miami-Dade County districts (HDs 103 and 109) are unconstitutionally packed, thus wasting black votes. These five Miami-Dade County districts are addressed below in separate subsidiary findings. It is likely that each of these districts will in fact perform for black candidates of choice.

201. Race or Hispanic status was considered in the drawing of performing black or Hispanic State Senate and House districts in an effort to enhance performance of the districts for black or Hispanic candidates of choice, to avoid dilution of the votes of black or Hispanic voters, and to comply with section 2 of the Voting Rights Act. Traditional districting principles also were considered and respected in the drawing of these districts. No effort was made to afford blacks or Hispanics substantially greater than proportional representation or to afford them representation inconsistent with traditional districting principles. Neither race nor Hispanic status was the controlling or predominant factor in the drawing of such districts.

202. The minority districts are reasonably compact.

203. The minority districts in fact comply to a reasonable extent with traditional districting principles.

204. No greater number of reasonably compact performing minority districts reasonably could have been drawn.

c. STATE HOUSE DISTRICT 108

205. According to the 2000 census, the non-Hispanic black population of new HD 108 was 57.7%, and the Hispanic black population was 2.3%, for a total black population of 60.0%. The non-Hispanic black voting age population was 53.0%, and the Hispanic black voting age population was 2.2%, for a total black voting age population of 55.2%. Other things being equal, the 55.2% black voting age population will increase during the coming decade.^[46]

206. A substantial share of the blacks in HD 108 are Haitian.

207. Even with a 55.2% black voting age population, new HD 108 will afford black voters a reasonable opportunity to elect candidates of their choice and probably will in fact perform for a black candidate of choice.

208. This is so in part because, according to 2000 data, blacks constitute 56.4% of registered Democrats in new HD 108, and Democrats constitute 67.5% of registered voters. Republicans constitute only 16.7% of registered voters. Actual voting also is overwhelmingly Democratic; in the 2000 presidential election, voters in new HD 108 voted 79.8% for Mr. Gore and 18.9% for Mr. Bush. A black candidate of choice is likely to win a contested Democratic primary, and the Democratic nominee is likely to win the general election.

209. The history of prior HD 108— which had different boundaries but the same core and similar demographics as the new HD 108—is not inconsistent with the finding that a 60.0% total black population and 55.2% black voting age population are sufficient to cause the district to perform for a black candidate of choice. When first drawn in 1992, prior HD 108 had (according to the 1990 census) a non-Hispanic black population of 62.7%, Hispanic black population of 3.4%, and total black population of 66.1%. Similarly, when first drawn in 1992, prior HD 108 had (according to the 1990 census) a non-Hispanic black voting age population of 57.1%, Hispanic black voting age population of 3.3%, and total black voting age population of 60.4%. This is more than five percentage points higher than the comparable numbers for new HD 108 under the 2000 census.

210. The percentage black voting age population and black total population in prior HD 108 increased during the decade of the 1990s. By the time of the 2000 census, prior HD 108 had a total black population of 68.1% and a total black voting age population of 64.3%—more than eight points higher than new HD 108—but there is no reason to believe that percentages that high were necessary for the district to perform for a black candidate of choice.

211. Prior HD 108 performed for the black candidate of choice in every election from 1992 through 2000. In 1994 and 1996, the incumbent (a black first elected in 1992) was unopposed. In 1998, the only two candidates were black; the incumbent won the Democratic primary and was unopposed in the general election. In 2000, the seat was open. Philip J. Brutus, who is Haitian-American and black, won the Democratic primary with over 63% of the vote against a single opponent. Mr. Brutus won the general election with over 81% of the vote, thus becoming the first Haitian-American to serve in the Florida legislature.^[47]

212. The conclusion that HD 108 will perform for a black candidate of choice also is supported by the history of HD 118—a different district in the same county that has performed for black candidates of choice despite having a lower percentage of black voters. HD 118 performed for the black candidate of choice in each election from 1992 through 2000. According to the 1990 census, the black total population (34.6%) and black voting age population (31.6%) were substantially lower for prior HD 118 than the comparable numbers for new HD 108 under the 2000 census. So far as shown by this record, the likelihood that HD 108 will perform for black candidates of choice in the coming decade is higher than the likelihood would have been with respect to HD 118 based on information known in 1992. Prior HD 118 performed, and it is likely that new HD 108 will perform also.

213. Techniques of ecological regression and projection of future voting behavior that are reasonably reliable and accepted in the field of political science support the conclusion that HD 108 will perform for black candidates of choice. In a closed Democratic primary in which only a single candidate has substantial support among black voters, HD 108 is likely to perform for that candidate. None of the parties' analyses have projected a vote below 50% for the black candidate of choice in such an election. Defendants' projection of 52.0% is reasonable. See Speaker's Exh. 75-2, Table R6; Tr. 1334. Further, it is likely that the Democratic primary winner will prevail in any general election.

214. The virtually uncontested findings set forth in paragraph 213, without more, make it likely that HD 108 will perform for black candidates of choice, because if ever the ability of HD 108 to elect a black candidate of choice is in jeopardy, it is likely that the Democratic primary will be closed (see supra paragraph 160), and it is likely that there will be a single black candidate of choice (see supra paragraphs 161-162).

215. The likely margin by which HD 108 will perform for a black candidate of choice is not as high as that attending prior HD 108. For the two districts HD 108 and HD 118 combined, however, the likely performance for black candidates of choice has been enhanced; the likely performance of HD 118 for a black candidate of choice has been enhanced by more than the likely performance of HD 108 has been diminished.

216. In the unlikely event of an open primary in HD 108, a black candidate of choice likely would be competitive, but, other things being equal, would have less than an even chance of prevailing. Plaintiffs' projection that a black candidate of choice would draw 47.2% of the vote in such a primary is reasonable. See Pl. Exh. 81-C at 38, Table 13. Particular candidacies and particular issues could turn any such election either way.

217. In sum, HD 108 is likely to perform for black candidates of choice. The votes of blacks in HD 108 have not been diluted.

d. STATE HOUSE DISTRICT 118

218. According to the 2000 census, the non-Hispanic black population of new HD 118 was 43.1%, and the Hispanic black population was 2.0%, for a total black population of 45.1%. The non-Hispanic black voting age population was 39.9%, and the Hispanic black voting age population was 1.9%, for a total black voting age population of 41.8%.

219. Even with a 41.8% black voting age population, new HD 118 will afford black voters a reasonable opportunity to elect candidates of their choice and probably will in fact perform for black candidates of choice.

220. This is so in part because, according to 2000 data, blacks constitute 64.4% of registered Democrats in new HD 118, and Democrats constitute 61.8% of registered voters. Republicans constitute only 20.2% of registered voters. Actual voting also is overwhelmingly Democratic; in the 2000 presidential election, voters in new HD 118 voted 74.8% for Mr. Gore and 24.1% for Mr. Bush. A black candidate of choice is likely to win a contested Democratic primary, and the Democratic nominee is likely to win the general election.

221. The history of prior HD 118— which had different boundaries but the same core and similar demographics as the new HD 118—supports the finding that a 45.1% total black population and 41.8% black voting age population are sufficient to cause the district to perform for black candidates of choice. When first drawn in 1992, prior HD 118 had (according to the 1990 census) a non-Hispanic black population of 33.3%, Hispanic black population of 1.3%, and total black population of 34.6%. Similarly, when first drawn in 1992, prior HD 118 had (according to the 1990 census) a non-Hispanic black voting age population of 30.4%, Hispanic black voting age population of 1.2%, and total black voting age population of 31.6%. This is more than ten percentage points lower than the comparable numbers for new HD 118 under the 2000 census.

222. Prior HD 118 performed for the black candidate of choice in every election from 1992 through 2000. In 1992, a black candidate won a plurality in the initial Democratic primary, won the Democratic primary runoff with 63.2% of the vote, and won the general election with 62.7% of the vote.^[48] In 1994, the incumbent was reelected without opposition. In 1996, the incumbent was unopposed in the Democratic primary and won with 64.5% of the vote in the general election. In 1998, the incumbent was unopposed in the Democratic primary and won with 63.0% of the vote in the general election. In 2000, there was an open seat. A first and second Democratic primary cut six candidates to one, who was black;^[49] he in turn won the general election with only write-in opposition.^[50]

223. Techniques of ecological regression and projection of future voting behavior that are reasonably reliable and accepted in the field of political science support the conclusion that HD 118 will perform for black candidates of choice. In a closed Democratic primary in which only a single candidate has substantial support among black voters, HD 118 is likely to perform for that candidate. None of the parties' analyses have projected a vote below 50% for the black candidate of choice in such an election. Further, it is likely that the Democratic primary winner will prevail in any general election.

224. The virtually uncontested findings set forth in paragraph 223, without more, make it likely that HD 118 will perform for black candidates of choice, because if ever the ability of HD 118 to elect a black candidate of choice is in jeopardy, it is likely that the Democratic primary will be closed (see supra paragraph 160), and it is likely that there will be a single black candidate of choice (see supra paragraphs 161-162).

225. In the unlikely event of an open primary in HD 118, a black candidate of choice likely would prevail, but an opposing candidate could be competitive. Plaintiffs' projection that a black candidate of choice would draw 51.3% of the vote in such a primary is reasonable.^[51] Particular candidacies and particular issues could turn any such election either way.

226. There is no evidence that any constitutionally-permissible alternative to HD 118 could be drawn that would meaningfully improve the projected performance of this district for a black candidate of choice.

227. In sum, HD 118 is likely to perform for black candidates of choice. The votes of blacks in HD 118 have not been diluted.

e. STATE HOUSE DISTRICT 104

228. According to the 2000 census, the non-Hispanic black population of new HD 104 was 57.8%, and the Hispanic black population was 2.2%, for a total black population of 60.0%. The non-Hispanic black voting age population was 53.7%, and the Hispanic black voting age population was 2.2%, for a total black voting age population of 55.9%. Other things being equal, the 55.9% black voting age population will increase during the coming decade.^[52]

229. Even with a 55.9% black voting age population, new HD 104 will afford black voters a reasonable opportunity to elect candidates of their choice and probably will in fact perform for a black candidate of choice.

230. This is so in part because, according to 2000 data, blacks constitute 64.1% of registered Democrats in new HD 104, and Democrats constitute 73.3% of registered voters. Republicans constitute only 12.1% of registered voters. Actual voting also is overwhelmingly Democratic; in the 2000 presidential election, voters in new HD 104 voted 86.5% for Mr. Gore and 12.7% for Mr. Bush. A black candidate of choice is likely to win a contested Democratic primary, and the Democratic nominee is likely to win the general election.

231. The history of prior HD 104— which had different boundaries but the same core and similar demographics as the new HD 104—suggests that a 60.0% total black population and 55.9% black voting age population are sufficient to cause the district to perform for black candidates of choice. When first drawn in 1992, prior HD 104 had (according to the 1990 census) a non-Hispanic black population of 56.5%, Hispanic black population of 2.3%, and total black population of 58.8%, lower than the total black population of newly created HD 104 under the 2000 census. Similarly, when first drawn in 1992, prior HD 104 had (according to the 1990 census) a non-Hispanic black voting age population of 51.1%, Hispanic black voting age population of 2.2%, and total black voting age population of 53.3%, again lower than new HD 104 under the 2000 census.

232. The percentage black voting age population and black total population in prior HD 104 increased during the decade of the 1990s. By the time of the 2000 census, prior HD 104 had a total black population of 70.0% and a total black voting age population of 65.9%—ten points higher than new HD 104—but there is no reason to believe that percentages that high were necessary for the district to perform for a black candidate of choice.

233. Prior HD 104 performed for the black candidate of choice in every election from 1992 through 2000. New HD 104 has demographics more favorable to a black candidate of choice than did prior HD 104 when it was created. The demographics of new HD 104 are likely to become more favorable for black candidates of choice as time passes, just as occurred in prior HD 104. There is no reason to believe that new HD 104 will perform less favorably for black candidates of choice in the coming decade than did old HD 104 in the prior decade.

234. Techniques of ecological regression and projection of future voting behavior that are reasonably reliable and accepted in the field of political science support the conclusion that HD 104 will perform for black candidates of choice. In a closed Democratic primary in which only a single candidate has substantial support among black voters, HD 104 is likely to perform for that candidate. Indeed, even in an open Democratic primary, plaintiffs' own projections show a black candidate of choice prevailing with 53.7% of the vote. See Pl. Exh. 81-C at 38, Table 13. Further, it is likely that the Democratic primary winner will prevail in any general election.

235. In sum, HD 104 is likely to perform for a black candidate of choice. The votes of blacks in HD 104 have not been diluted.

f. STATE HOUSE DISTRICT 103

236. According to the 2000 census, the non-Hispanic black population of new HD 103 was 70.5%, and the Hispanic black population was 2.5%, for a total black population of 73.0%. The non-Hispanic black voting age population was 67.9%, and the Hispanic black voting age population was 2.5%, for a total black voting age population of 70.4%.

237. According to 2000 data, blacks constitute 82.2% of registered Democrats in new HD 103, and Democrats constitute 75.8% of registered voters. Republicans constitute only 11.5% of registered voters. Actual voting also is overwhelmingly Democratic; in the 2000 presidential election, voters in new HD 103 voted 88.0% for Mr. Gore and 11.4% for Mr. Bush. A black candidate of choice is likely to win a contested Democratic primary, and the Democratic nominee is likely to win the general election.

238. New HD 103 will afford black voters a reasonable opportunity to elect candidates of their choice and almost certainly will in fact perform for black candidates of choice. Plaintiffs concede this. Indeed, plaintiffs' claim is not that HD 103 will not perform; instead, their claim is that HD 103 is "packed," that is, that there are more blacks in HD 103 than needed to elect black candidates of choice, thus in effect "wasting" black votes.

239. New HD 103 could perform for black candidates of choice with far fewer blacks in the district.

240. HD 103 is contiguous to HDs 104, 105, 110, and 112.

241. HDs 110 and 112 are majority Hispanic districts. Moving black voters into HD 110 or HD 112 would not enhance the influence of those voters or of black voters generally.

242. HD 104 is already a performing black majority district. See supra paragraphs 228-235. Moving black voters into HD 104 would not enhance the influence of those voters or of black voters generally.

243. HD 105 has total population 45.0% white, 30.0% non-Hispanic black, 1.2% Hispanic black, 20.4% non-black Hispanic, and 3.4% other. The boundary between HD 103 and HD 105 is the county line separating Miami-Dade and Broward Counties. HD 103 is entirely in Miami-Dade County. HD 105 is entirely in Broward County. There was thus a reason unrelated to race for the drawing of the line separating HD 103 from HD 105. There has been no showing that blacks could be moved out of HD 103 into HD 105, consistent with traditional zoning principles and applicable law, to an extent that would materially increase black influence in HD 105. There has been no showing that moving black voters into HD 105 would enhance the influence of those voters or black voters generally.

244. Moving blacks out of HD 103 would not increase the number of districts expected to perform for black candidates of choice.

245. Moving blacks out of HD 103 would not result in any black voter having a greater opportunity to impact the electoral process than that voter has under the plan as adopted by the Florida legislature.

g. STATE HOUSE DISTRICT 109

246. According to the 2000 census, the non-Hispanic black population of new HD 109 was 61.1%, and the Hispanic black population was 3.3%, for a total black population of 64.4%. The non-Hispanic black voting age population was 56.8%, and the Hispanic black voting age population was 3.6%, for a total black voting age population of 60.4%.

247. According to 2000 data, blacks constitute 79.2% of registered Democrats in new HD 109, and Democrats constitute 77.5% of registered voters. Republicans constitute only 11.3% of registered voters. Actual voting also is overwhelmingly Democratic; in the 2000 presidential election, voters in new HD 109 voted 86.8% for Mr. Gore and 12.3% for Mr. Bush. A black candidate of choice is likely to win a contested Democratic primary, and the Democratic nominee is likely to win the general election.

248. New HD 109 will afford black voters a reasonable opportunity to elect candidates of their choice and almost certainly will in fact perform for black candidates of choice. Plaintiffs concede this. Indeed, plaintiffs' claim is not that HD 109 will not perform; instead, their claim is that HD 109 is "packed," that is, that there are more blacks in HD 109 than needed to elect black candidates of choice, thus in effect "wasting" black votes.^[53]

249. New HD 109 could perform for black candidates of choice with fewer blacks in the district, but the level of confidence in the district's performance would decrease as the black population decreased.

250. HD 109 is contiguous to HDs 104, 106, 107, 108, and 113. HDs 104 and 108 are already performing black majority districts. See supra paragraphs 228-235 and 205-217. HDs 107 and 113 are majority Hispanic districts. Moving black voters into HD 107 or HD 113 would not enhance the influence of those voters or of black voters generally. HD 106 has total population 52.0% white, 5.6% non-Hispanic black, 1.0% Hispanic black, 38.5% non-black Hispanic, and 2.9% other. There has been no showing that blacks could be moved out of HD 109 into HD 106 consistent with traditional zoning principles and applicable law, to an extent that would materially increase black influence in HD 106. There has been no showing that moving black voters into HD 106 would enhance the influence of those voters or black voters generally.

251. Moving blacks out of HD 109 would not increase the number of districts expected to perform for black candidates of choice.

252. Moving blacks out of HD 109 would not result in any black voter having a greater opportunity to impact the electoral process than that voter has under the plan as adopted by the Florida legislature.

h. PROPORTIONALITY

253. According to the 2000 census, the voting age population of Miami-Dade County was 19.2% black. The total population of Miami-Dade County was 2,253,362, enough for 16.9 members of the Florida House of Representatives. Under the apportionment plan adopted in 2002, there are 18 districts wholly or partly in Miami-Dade County. Five districts constitute 29.6% of 16.9 districts and 27.8% of 18 districts. CDs 103, 104, 108, 109, and 118 are performing black districts in Miami-Dade County.

B. LEGAL CONCLUSIONS

Plaintiff's claims of vote dilution based on race are governed by the Voting Rights Act of 1965, as amended. Section 2(a) of the Act prohibits any state or political subdivision from denying or abridging any citizen's right to vote on account of race or color, 42 U.S.C. § 1973(a), or membership in a language minority group, 42 U.S.C. § 1973b(f)(2). Section 2(b) provides that it is a violation of this standard if "the political processes leading to nomination or election ... are not equally open to participation by members of a class of citizens protected by" the statute "in that its members have less opportunity than other members of the electorate to participate in the political process and to elect representatives of their choice." 42 U.S.C. § 1973(b) (emphasis added).^[54] Section 2(b) also includes a proviso: "nothing in this section establishes a right to have members of a protected class elected in numbers equal to their proportion in the population." Id. This statutory language provides the starting point for our analysis.

To prevail on a claim under section 2, a plaintiff need not show that a challenged state action was undertaken with discriminatory intent. It is sufficient if a challenged action has a prohibited effect, irrespective of the intent with which it was adopted. See Thornburg v. Gingles, 478 U.S. 30, 35, 106 S.Ct. 2752, 2758, 92 L.Ed.2d 25 (1986).^[55]

Section 2 is violated when "a certain electoral law, practice, or structure interacts with social and historical conditions to cause an inequality in the opportunities enjoyed by black [or other covered minority voters] and white voters to elect their preferred representatives." Id. at 47, 106 S.Ct. at 2764. This requires that district lines be drawn in a manner that does not impermissibly dilute minority votes, that is, in a manner affording an equal opportunity for the election of minority candidates of choice, when a minority group is sufficiently large, compact, and cohesive to achieve that result. See id. at 49-51, 106 S.Ct. at 2765-66 (finding dilution in multi-member districts and announcing three factors necessary to a prima facie case); Growe v. Emison, 507 U.S. 25, 40, 113 S.Ct. 1075, 1084, 122 L.Ed.2d 388 (1993) (applying same analysis to single-member districts).

To prevail on a claim of vote dilution under section 2, a plaintiff must, at a minimum, establish the three Gingles factors: (1) that the protected group is "sufficiently large and geographically compact"; (2) that the protected group is "politically cohesive"; and (3) that other voters—ordinarily the white majority—vote sufficiently as a bloc that they "usually" defeat minority candidates of choice. Gingles, 478 U.S. at 50-51, 106 S.Ct. at 2766.^[56] All other circumstances affecting the protected group's ability to elect candidates of choice also are relevant; the three Gingles factors do not alone necessarily establish a section 2 violation. The essence of the required showing is this:

[P]laintiffs can prevail on a dilution claim only if they show that, under the totality of the circumstances, the State's apportionment scheme has the effect of diminishing or abridging the voting strength of the protected class.

Voinovich v. Quilter, 507 U.S. 146, 157, 113 S.Ct. 1149, 1157, 122 L.Ed.2d 500 (1993).^[57]

In the case at bar, it is undisputed that the Gingles factors are satisfied. In other words, there are groups of black voters in Florida that are sufficiently large and geographically compact to allow the creation of districts that would perform for black candidates of choice, black voters are politically cohesive, and other (non-black) voters vote sufficiently as a bloc that, in the absence of appropriately drawn districts, those voters would usually defeat black candidates of choice. It also is undisputed that, under the totality of the circumstances, any apportionment plan that included no districts likely to perform for black candidates of choice would have the effect of impermissibly diminishing or abridging black voting strength in violation of the Voting Rights Act. It is thus undisputed that section 2 clearly required the drawing of at least some number of districts likely to perform for black candidates of choice.

The parties have not explicitly set forth their respective positions on how many minority districts are required by section 2 under the circumstances of the case at bar, nor on how many of the districts drawn by the legislature should be counted as minority districts. That depends to some degree on semantics: how the term "minority district" is defined. Courts sometimes have suggested that districts or potential districts of consequence in section 2 dilution claims are only those with an actual majority, or even a super-majority, of minority population or voters. See, e.g., Perez v. Pasadena Independent School Dist., 165 F.3d 368 (5th Cir.1999) (holding that, under the first Gingles factor, no dilution occurs unless a district could be drawn in which the minority at issue would constitute an actual majority of citizens eligible to vote); McNeil v. Springfield Park Dist., 851 F.2d 937 (7th Cir.1988) (holding that, under the first Gingles factor, no dilution occurs unless a district could be drawn in which the minority at issue would constitute an actual majority of the voting age population); see also African American Voting Rights Legal Defense Fund, Inc. v. Villa, 54 F.3d 1345, 1348 n. 4 (8th Cir. 1995) (finding no section 2 violation because the percentage of safe black wards was roughly the same as the percentage black voting age population, and defining a safe ward as one in which the "black majority has a practical opportunity to elect the candidate of its choice," that is, a district with either 60% black voting age population or 65% black total population); Ketchum v. Byrne, 740 F.2d 1398, 1415 (7th Cir.1984) (disapproving district court's refusal to require super-majorities in districts adopted as a remedy for a section 2 violation and asserting "that a guideline of 65% of total population ... has achieved general acceptance in redistricting jurisprudence").

But the approach of focusing mechanically on the percentage of minority population (or voting age population or registered voters) in a particular district, without assessing the actual voting strength of the minority in combination with other voters, has been justly criticized. See, e.g., Bernard Grofman et al., Drawing Effective Minority Districts: A Conceptual Framework and Some Empirical Evidence, 79 N.C. L.Rev. 1383 (2001). Other courts have recognized the significance of districts —sometimes labeled "influence districts" —in which members of a minority group do not constitute a majority but nonetheless have the ability, acting together with other ("cross-over") voters, to elect minority candidates of choice. See, e.g., Voinovich, 507 U.S. at 157, 113 S.Ct. at 1157 (assuming without deciding that section 2 could be violated by a state's failure to create an "influence district" in which minority voters, although not a majority, could combine with white cross-over voters to elect minority candidates of choice); Uno v. City of Holyoke, 72 F.3d 973 (1st Cir.1995) (vacating district court's finding of no section 2 violation based on the impracticality of drawing any district in which the minority would constitute an actual majority and remanding for consideration of influence districts); Page v. Bartels, 144 F.Supp.2d 346 (D.N.J.2001) (three-judge court) (finding no violation of section 2 because minority voters, together with cross-over voters, could usually elect minority candidates of choice, thus negating the third Gingles factor); Rural West Tennessee African-American Affairs Council, Inc. v. McWherter, 877 F.Supp. 1096 (W.D.Tenn.1995) (three-judge court) (holding that any district in which a minority group composes 25% to 55% of the voting age population is an "influence district" that should be considered in evaluating the "totality of the circumstances" under Gingles).

We have no occasion to devise a formula for determining how many districts of one type or another are required by section 2. We use the term "performing minority district" to refer to a district that is likely to perform for—that is, usually to result in the election of—minority candidates of choice. Such a district may or may not have an actual majority (or super-majority) of minority population, voting age population, or registered voters.

There is no bright line test for determining whether a district is likely to perform for minority candidates of choice. All circumstances that are likely to affect voting behavior and election outcomes in the district are relevant. It ordinarily is sufficient if, under all the circumstances, the district is drawn in a manner likely to result in the election of minority candidates of choice in most elections. See, e.g., Gingles, 478 U.S. at 49, 106 S.Ct. at 2765-66 (stating that in order for there to be impermissible dilution, "a bloc voting majority must usually be able to defeat candidates supported by a politically cohesive, geographically insular minority group") (emphasis in original); Johnson v. De Grandy, 512 U.S. 997, 1014 n. 11, 114 S.Ct. 2647, 2658 n. 11, 129 L.Ed.2d 775 (1994) ("the ultimate right of § 2 is equality of opportunity, not a guarantee of electoral success"); Johnson v. Hamrick, 296 F.3d 1065 (11th Cir.2002) (rejecting section 2 racial vote dilution claim on ground that plaintiffs failed to show that white bloc voting regularly defeated the candidate preferred by black voters); Clarke v. City of Cincinnati, 40 F.3d 807 (6th Cir.1994) (same).

It is uncontested in the case at bar that section 2 imposes no requirement greater than the creation of the number of performing minority districts that ostensibly are included in the plans adopted by the Florida legislature. It also is uncontested that nothing in section 2 or the Constitution prohibits the creation of this number of performing minority districts. The dispute in the case at bar is not whether this number of performing minority districts is required, nor whether this number of performing minority districts is sufficient. Instead, the parties implicitly accept, for purposes of this litigation, that this number of performing districts is both required and sufficient; the parties disagree only on whether the ostensibly performing districts drawn by the legislature are in fact likely to perform.

In addressing this issue, the parties have appropriately focused on actual voting strength, as have we. Under Gingles, our task

is to consider the "totality of the circumstances" and to determine, based "upon a searching practical evaluation of the `past and present reality,'" S. Rep. [No. 97-417, 97th Cong. 2nd Sess. 30 (1982)], U.S.Code Cong. & Admin. News 1982, p. 208 (footnote omitted), whether the political process is equally open to minority voters. "`This determination is peculiarly dependent upon the facts of each case,'" Rogers [v. Lodge, 458 U.S. 613, 621, 102 S.Ct. 3272, 73 L.Ed.2d 1012 (1982)], quoting Nevett v. Sides, 571 F.2d 209, 224 (C.A.5 1978), and requires "an intensely local appraisal of the design and impact" of the contested electoral mechanisms. 458 U.S. at 622, 102 S.Ct. 3272.

Gingles, 478 U.S. at 79, 106 S.Ct. at 2781.

A searching practical evaluation of past and present reality, based on an intensely local appraisal of the totality of the circumstances, focuses on the likely actual effects of the apportionment plans at issue.

This is particularly appropriate under the specific circumstances of this case. This case does not involve an attempt to create minority districts within a jurisdiction for the first time. Instead, this case involves districts with an unbroken 10-year history of demonstrated performance for minority candidates of choice. By agreeing, at least implicitly, that the numbers of ostensibly performing districts included in the current plans are the appropriate numbers, the parties have in effect agreed that if the ostensibly performing black districts will in fact perform for black voters—as we have found they will— then under the totality of the circumstances, black voters will have a reasonable opportunity to elect candidates of their choice, and black votes will not be diluted.

As set forth in detail in our findings of fact, in the districts that have been challenged in the case at bar, black candidates of choice will prevail in most (if not all) elections. This is sufficient to defeat plaintiffs' racial vote dilution claims. We thus address no further the percentage of total population, voting age population, or registered voters that might be required, in other circumstances, to cause a district to perform for minority candidates of choice.

In sum, under the congressional and state legislative apportionment plans adopted by the Florida legislature, black voters will not have "less opportunity than other members of the electorate to participate in the political process and to elect representatives of their choice." 42 U.S.C. § 1973(b). The apportionment plans will not have "the effect of diminishing or abridging the voting strength" of black voters. Voinovich, 507 U.S. at 157, 113 S.Ct. at 1157. No violation of section 2 has been established.

III. VOTE DILUTION BASED ON POLITICAL PARTY

The plaintiffs and intervenors also allege that the Florida legislature politically gerrymandered the congressional districts^[58] to favor Republicans, thereby violating the rights of Democratic voters guaranteed by the Equal Protection Clause of the Fourteenth Amendment.^[59] Citing Davis v. Bandemer, 478 U.S. 109, 106 S.Ct. 2797, 92 L.Ed.2d 85 (1986), these parties contend that the redistricting plan was intentionally designed to have the impermissible effect of diluting the voting strength of Florida Democrats. Specifically, they argue that, in the state of Florida, where the Republican/Democrat divide is near 50/50, the redistricting plan is unconstitutional because eighteen of the proposed districts will likely perform for a Republican candidate, and only seven districts will do so for a Democratic candidate. Although the plaintiffs and intervenors acknowledge that the Equal Protection Clause does not guarantee proportional representation, they claim that this 18/7 disparity simply goes "too far" and requires a federal court to come to their rescue.

"Too far," however, is not a workable standard with which to analyze properly the redistricting plan and its effects on Florida's Democratic voters. Therefore, before discussing the relevant facts in this case, our first task is to discern the correct rule of law. Our starting point is Bandemer, where a majority of the Court found justiciable a claim by Indiana Democrats that a reapportionment plan designed by the Republican-controlled legislature was an unconstitutional political gerrymander. Bandemer, 478 U.S. at 125, 106 S.Ct. at 2806. Only a plurality, however, agreed on how exactly the Democrats could succeed on their claim. Writing for that plurality, Justice White indicated that the "plaintiffs were required to prove both intentional discrimination against an identifiable political group and an actual discriminatory effect on that group." Id. at 127, 106 S.Ct. at 2808 (citing City of Mobile v. Bolden, 446 U.S. 55, 67-68, 100 S.Ct. 1490, 1499-1500, 64 L.Ed.2d 47 (1980)).^[60]

Although Bandemer's two-part test for equal protection claims had been well-established in the racial vote dilution context, the plurality did not explain how that test should apply in political gerrymandering cases.^[61] All that is clear from the plurality's opinion is that the test for political gerrymandering claims is strongly linked to the line of cases involving racial vote dilution. Indeed, in enunciating and then discussing the proper analysis of political gerrymandering claims, the plurality repeatedly drew on that long history, beginning with Whitcomb v. Chavis, 403 U.S. 124, 91 S.Ct. 1858, 29 L.Ed.2d 363 (1971), and White v. Regester, 412 U.S. 755, 93 S.Ct. 2332, 37 L.Ed.2d 314 (1973), and evolving through City of Mobile v. Bolden, 446 U.S. 55, 100 S.Ct. 1490, 64 L.Ed.2d 47 (1980), and Rogers v. Lodge, 458 U.S. 613, 102 S.Ct. 3272, 73 L.Ed.2d 1012 (1982).^[62]

Therefore, in our attempt to articulate the rule of law in this case, we rely heavily on those well-worn cases that analyzed vote dilution claims brought by racial and ethnic groups under the Equal Protection Clause.^[63] In subpart A, we trace the analytical evolution of those cases and emphasize a point the Bandemer plurality did not—that the minority group must show, among other factors, that the challenged scheme is actually the cause of their political impotence. In Rogers, for example, the Court articulated a group of evidentiary factors that a court should consider to determine whether unconstitutional vote dilution has been established. Rogers, 458 U.S. at 620 n. 8, 102 S.Ct. at 3277 n. 8. Prominent among those was racially polarized bloc voting, without which, the Court explained, the minority group could not establish that the electoral scheme was the actual cause of its inability to participate politically. Id. at 623, 102 S.Ct. at 3279 ("[W]ithout [proof of] bloc voting[,] the minority candidates would not lose elections solely because of their race."). This requirement of proving that an electoral scheme caused the minority group's inability to participate politically was more fully explained in Thornburg v. Gingles, 478 U.S. 30, 106 S.Ct. 2752, 92 L.Ed.2d 25 (1986), which analyzed a vote dilution claim under the Voting Rights Act, 42 U.S.C. § 1973.

In subpart B, we examine the Gingles Court's analysis, which required that a minority group establish three "necessary preconditions": (1) that it is geographically compact; (2) that it is politically cohesive; and (3) that the majority group votes as a bloc and usually defeats the minority group's candidates. Id. at 48-51, 106 S.Ct. at 2765-67. We maintain that, even though Gingles did not involve an equal protection claim, the three factors were derived by the Court from the principles set forth in the vote dilution cases brought under the Equal Protection Clause. We therefore conclude that the three preconditions have always been and remain elements of constitutional vote dilution claims.

In subpart C, we hold that the three Gingles preconditions, along with other factors, establish a prima facie case of "intentional discrimination against an identifiable political group and an actual discriminatory effect on that group." Bandemer, 478 U.S. at 127, 106 S.Ct. at 2808. First, we explain why the three Gingles preconditions apply in the political gerrymandering context just as they do in racial vote dilution cases. Second, we discuss the two-prong test set forth in Bandemer-legislative intent and discriminatory effects —in light of the three Gingles preconditions.

Finally, in subpart D, we analyze the facts in the instant case and determine whether the plaintiffs and intervenors have properly established an Equal Protection Clause violation. We conclude that they have not, primarily because they have not shown that Democrats are politically cohesive and geographically compact, and that Republicans bloc vote in a way that will "usually defeat" the Democrats' candidates of choice.

A. EQUAL PROTECTION VOTE DILUTION CASES

The Supreme Court first decided, in Baker v. Carr, 369 U.S. 186, 82 S.Ct. 691, 7 L.Ed.2d 663 (1962), that a claim asserted under the Equal Protection Clause challenging the constitutionality of a state's legislative apportionment presented a justiciable controversy. Two years later, the Court struck down two legislatively-proposed apportionment plans that provided election districts of disparate sizes. Reynolds v. Sims, 377 U.S. 533, 84 S.Ct. 1362, 12 L.Ed.2d 506 (1964). The Court in Reynolds was not concerned that a person was denied his right to vote or that his vote was not counted, but was instead worried that the vote of a person living in one district was more heavily weighted than the vote of one living in another, and therefore violated the one person, one vote clause. Id. at 563-64, 84 S.Ct. at 1382; see also Bandemer, 478 U.S. at 123, 106 S.Ct. at 2805-06. The Reynolds Court recognized that "each and every citizen has an inalienable right to full and effective participation in the political processes of his State's legislative bodies." Reynolds, 377 U.S. at 565, 84 S.Ct. at 1383. According to the Court, this "opportunity for equal participation of all voters" is guaranteed by the Equal Protection Clause, and, therefore, "[d]iluting the weight of votes because of place of residence impairs basic constitutional rights under the Fourteenth Amendment." Id. at 566, 84 S.Ct. at 1384.

Apportionment schemes that satisfy the one person, one vote requirement of Reynolds are still vulnerable, the Court subsequently held, if they "operate to dilute or cancel the voting strength of racial or political elements." Whitcomb, 403 U.S. at 144, 91 S.Ct. at 1869. See also Fortson v. Dorsey, 379 U.S. 433, 439, 85 S.Ct. 498, 501, 13 L.Ed.2d 401 (1965); Burns v. Richardson, 384 U.S. 73, 89, 86 S.Ct. 1286, 1295, 16 L.Ed.2d 376 (1966). In Whitcomb, the residents of a "ghetto," an area within a multi-member state legislative district with a concentrated minority population, brought an equal protection claim against Indiana's legislative apportionment scheme. 403 U.S. at 128-29, 91 S.Ct. at 1861. The plaintiffs argued that, under the scheme, the ghetto residents were unable to elect their own candidates. Id. at 128-29, 91 S.Ct. at 1861-62.

The Supreme Court denied the residents' claim, however, pointing out that they had not shown that the multi-member district had been "conceived or operated as [a] purposeful device[] to further racial or economic discrimination." Id. at 149, 91 S.Ct. at 1872. Moreover, "that the number of ghetto residents who were legislators was not in proportion to ghetto population [does not] satisfactorily prove invidious discrimination absent evidence and findings that ghetto residents had less opportunity than did other [non-ghetto] residents [who also lived in the multi-member district] to participate in the political process and to elect legislators of their choice." Id. Rather, in order for the scheme to rise to the level of an equal protection violation, it was the ghetto residents' burden to prove that they were "not allowed to register or vote, to choose the political party they desired to support, to participate in its affairs or to be equally represented on those occasions when legislative candidates were chosen." Id.

Two years later, the plaintiffs in White v. Regester met that burden. In White, two different minority groups, blacks and Mexican-Americans, brought equal protection challenges to Texas's legislative apportionment plan, which placed each group in a separate multi-member district in which it constituted a minority. 412 U.S. at 758-59, 93 S.Ct. at 2336. In ruling on those claims, the Court reaffirmed the proposition that "it is not enough that the [minority] group allegedly discriminated against has not had legislative seats in proportion to its voting potential." Id. at 765-66, 93 S.Ct. at 2339. Rather, it is the plaintiffs' burden to show "that the political processes leading to nomination and election were not equally open to participation by the group in question—that its members had less opportunity ... to participate in the political processes and to elect [candidates] of their choice." Id. at 766, 93 S.Ct. at 2339.

For the black plaintiffs, that burden was met with evidence that only two blacks had ever even been slated by the local Democratic party, a white-dominated organization that "did not need the support of the Negro community"; that "racial campaign tactics" had been used in white precincts to defeat black candidates of choice; that Texas had a history of official racial discrimination; and that Texas employed certain electoral devices, like the majority-vote requirement and the "place rule," which resulted in elections from multi-member districts being reduced to head-to-head contests for each position. Id. at 766-67, 93 S.Ct. at 2339-40. The Mexican-American plaintiffs met their burden with evidence that they typically "suffer[] a cultural and language barrier that makes ... participation in community processes extremely difficult"; that "cultural incompatibility conjoined with the poll tax and the most restrictive voter registration procedures in the nation have operated to effectively deny Mexican-Americans access to the political processes"; that there was a history of discrimination against Mexican-Americans, reflected in the fact that voting registration is very low; and that the local delegation was "insufficiently responsive to Mexican-American interests." Id. at 768-69, 93 S.Ct. at 2340-41 (internal footnotes and punctuation omitted). Weighing all of this together, the Court held, leads to the conclusion that the multimember district was designed and operated invidiously to exclude these groups from participation in political life. Id. at 767, 769-70, 93 S.Ct. at 2340-41. The proper remedy the district court had fashioned was to divide the multi-member districts into single-member districts in which the minority groups could constitute majorities. Id. at 765, 93 S.Ct. at 2339.

From the Court's opinion in White, the Fifth Circuit identified in Zimmer v. McKeithen, "a panoply of factors, any number of which may contribute to the existence of [unconstitutional vote] dilution." 485 F.2d 1297, 1305 (5th Cir.1973). Even if racial motivation cannot be proven, plaintiffs can succeed on their vote dilution claim by demonstrating "a lack of access to the process of slating candidates, the unresponsiveness of legislators to their particularized interests, a tenuous state policy underlying preference for multi-member or at-large districting, or that the existence of past discrimination in general precludes the effective participation in the election system." Id. Their claim is enhanced by showing that certain electoral devices, like majority vote requirements or anti-single shot voting provisions,^[64] work to the minority's disadvantage. Id.^[65]

After Zimmer was issued, the Supreme Court decided Washington v. Davis, 426 U.S. 229, 96 S.Ct. 2040, 48 L.Ed.2d 597 (1976), and one year later, Village of Arlington Heights v. Metropolitan Housing Development, 429 U.S. 252, 97 S.Ct. 555, 50 L.Ed.2d 450 (1977). Both cases involved claims by racial minorities asserted under the Equal Protection Clause, although neither case arose in the vote dilution context. The Court held first in Davis and then in Arlington Heights that, in order to succeed on a claim brought under the Equal Protection Clause of the Fourteenth Amendment, plaintiffs must present proof of racially discriminatory intent or purpose. Davis, 426 U.S. at 240, 96 S.Ct. at 2048; Arlington Heights, 429 U.S. at 265, 97 S.Ct. at 563. The Court recognized that "[d]etermining whether invidious discriminatory purpose was a motivating factor demands a sensitive inquiry into such circumstantial and direct evidence of intent as may be available." Arlington Heights, 429 U.S. at 266, 97 S.Ct. at 564.

The Fifth Circuit subsequently revised its vote dilution analysis in light of these cases. In Nevett v. Sides, 571 F.2d 209 (5th Cir.1978), cert. denied, 446 U.S. 951, 100 S.Ct. 2916, 64 L.Ed.2d 807 (1980), the court held that "a showing of racially motivated discrimination is a necessary element in an equal protection voting dilution claim," id. at 219, and that this requirement "may be satisfied by direct or circumstantial evidence," id. at 221. The panoply of factors outlined in Zimmer, the court held, is exactly the kind of circumstantial evidence courts should consider in determining legislative intent, particularly when there is no direct evidence. Id. at 222. Indeed, "a finding of racially discriminatory dilution under the Zimmer criteria raises an inference of intent and, therefore, ... a finding under the criteria satisfies the intent requirement." Id. at 217.

More significantly, the Nevett court streamlined the myriad Zimmer factors into two: racially polarized bloc voting and lack of government responsiveness to minority needs. Id. at 223. "When bloc voting has been demonstrated, a showing under Zimmer that the governing body is unresponsive to minority needs is strongly corroborative of an intentional exploitation of the electorate's bias." Id. Bloc voting, according to the court, "allows representatives to ignore minority interests without fear of reprisal at the polls." Id. The court further explained that bloc voting may be indicated by a showing under Zimmer of "the existence of past discrimination in general ..., large districts, majority vote requirements, anti-single shot voting provisions and the lack of provision for at-large candidates running from particular geographical subdistricts." Id. at 223 n. 18 (quoting Zimmer, 485 F.2d at 1305).^[66] "The establishment under Zimmer that blacks have been denied access to slating, registration, or other aspects of political participation" is an explanation of why representatives have ignored black interests—"because blacks cannot achieve the input to which they are entitled." Id. at 223.

On the same day Nevett was issued, the Fifth Circuit released three other vote dilution cases brought under the Equal Protection Clause: Bolden v. City of Mobile, 571 F.2d 238 (5th Cir.1978); Blacks United for Lasting Leadership v. City of Shreveport, 571 F.2d 248 (5th Cir.1978); Thomasville Branch of the NAAACP v. Thomas County, 571 F.2d 257 (5th Cir. 1978). In all three cases, the court analyzed the vote dilution claims under the newly-announced standard from Nevett. In Bolden, for example, the court affirmed the district court's finding that the city's at-large method of electing commissioners diluted the voting strength of blacks because the district court "gave careful consideration to each of the primary Zimmer criteria." Bolden, 571 F.2d at 243. Specifically, the court found evidence of racially polarized voting "of an acute nature," preventing any black from ever being elected to the city commission, general unresponsiveness on the part of the city commissioners to the needs of blacks, a significant history of past discrimination against blacks, and the use of various electoral devices, like the majority vote requirement.^[67]Id. at 243-44. "[T]hese findings ... compel the inference that the system has been maintained with the purpose of diluting the black vote, thus supplying the [intent requirement from Davis and Arlington Heights]." Id. at 245.

A plurality of the Supreme Court rejected the Fifth Circuit's effects-only test under the Equal Protection Clause in Bolden v. City of Mobile. Justice Stewart, writing for the plurality,^[68] first held that a showing of discriminatory purpose was required to prove vote dilution under the Fourteenth and Fifteenth Amendments and even under section 2 of the Voting Rights Act of 1965, 42 U.S.C. § 1973.^[69]Bolden, 446 U.S. at 65-66, 70, 100 S.Ct. at 1498-99, 1501. The plurality then noted that Zimmer was "decided upon the misunderstanding that it is not necessary to show a discriminatory purpose in order to prove a violation of the Equal Protection Clause—that proof of discriminatory effect is sufficient." Bolden, 446 U.S. at 71, 100 S.Ct. at 1501-02. The lower courts' reliance on Zimmer, according to the plurality, was therefore improper because the Zimmer criteria alone were "insufficient to prove an unconstitutionally discriminatory purpose in the present case." Id. at 73, 100 S.Ct. at 1503.

Four other Justices, however, believed that the evidence offered below was sufficient to satisfy the purposeful discrimination standard.^[70] Justice White, one of the four who found the evidence sufficient, wrote a separate dissenting opinion expressing his belief that a consideration of the factors articulated by the Fifth Circuit in Zimmer was the proper approach in determining whether vote dilution was caused by purposeful discrimination. Id. at 101-03, 100 S.Ct. at 1517-18 (White, J., dissenting). Justice White pointed out that Zimmer "articulated the very factors deemed relevant by White v. Regester and Whitcomb v. Chavis ... and that both [lower courts] considered these factors with the recognition that they are relevant only with respect to the question whether purposeful discrimination can be inferred." Id. at 101, 100 S.Ct. at 1517.^[71]

Justice White persuaded a majority of the Court of his position two years later in Rogers v. Lodge.^[72] In writing for the majority, Justice White acknowledged that, "[b]ecause the District Court in the present case employed the evidentiary factors outlined in Zimmer, it is urged that [the lower court's finding of unconstitutional vote dilution] is infirm for the same reasons that led to the reversal in [Bolden]." Id. at 620-21, 102 S.Ct. at 3277. The Court, however, disagreed and pointed out that the lower court had relied heavily on Nevett v. Sides, in which the Fifth Circuit had required a showing of discriminatory purpose and had held that, "although the evidentiary factors outlined in Zimmer were important considerations [in proving intent], the plaintiff is not limited to those factors." Id. at 621, 102 S.Ct. at 3277. This, the Court held, was the "proper legal standard." Id. at 622, 102 S.Ct. at 3278.

In reviewing the application of that standard to the facts, the Rogers Court concluded that the district court properly found unconstitutional vote dilution. First and foremost, the Court pointed out, "[t]here was ... overwhelming evidence of bloc voting along racial lines," which, along with evidence that "no black had ever been elected to the [county] [c]ommission[,] ... bear[s] heavily on the issue of purposeful discrimination." Rogers, 458 U.S. at 623, 102 S.Ct. at 3279. Like the Fifth Circuit in Nevett, the Court explained that "[v]oting along racial lines allows those elected to ignore black interests without fear of political consequences, and without bloc voting the minority candidates would not lose elections solely because of their race." Id. Because, however, "such facts are insufficient in themselves to prove purposeful discrimination," id. at 624, 102 S.Ct. at 3279, the Court took into consideration other findings: the impact of past discrimination on the ability of blacks to participate politically, the lack of responsiveness on the part of elected officials to the needs of the black community, and the use of certain electoral devices that worked to dilute the voting strength of blacks. Id. at 624-28, 102 S.Ct. at 3279-81.

B. VOTING RIGHTS ACT CASES

One month before the Supreme Court decided Rogers v. Lodge, Congress amended section 2 of the Voting Rights Act of 1965, 42 U.S.C. § 1973.^[73] "The amendment was largely a response to [the Supreme] Court's plurality opinion in [Bolden], which had declared that, in order to establish a violation either of section 2 or of the Fourteenth or Fifteenth Amendments, minority voters must prove that a contested electoral mechanism was intentionally adopted or maintained by state officials for a discriminatory purpose." Gingles, 478 U.S. at 35, 106 S.Ct. at 2758 (internal citations omitted). According to the Supreme Court's interpretation of the amendment, "Congress substantially revised [section] 2 to make clear that a violation could be proved by showing discriminatory effect alone and to establish as the relevant legal standard the `results test,' applied by this Court in White v. Regester, and by other federal courts before Bolden." Id. (internal citations omitted).

Indeed, the Senate Judiciary Committee Majority Report that accompanied the bill amending section 2 listed "typical factors" helpful in proving a discriminatory effect; the report explicitly mentioned that the factors were derived from White v. Regester, as refined and developed by the Fifth Circuit in Zimmer v. McKeithen.^[74]See S. REP. 97-417, at 28 n. 113 (1982), reprinted in 1982 U.S.C.C.A.N. 177, 206; see also Gingles, 478 U.S. at 36 n. 4, 106 S.Ct. at 2759 n. 4. Importantly, among the factors listed are those on which the Fifth Circuit placed primary reliance in Nevett, namely "the extent to which voting ... is racially polarized" and "whether there is a significant lack of responsiveness on the part of elected officials to the [minority group's needs]." S. REP. 97-417, at 28-29 (1982), reprinted in 1982 U.S.C.C.A.N. 177, 206-07. Finally, the amendment incorporated the proposition developing in the case law with regard to equal protection claims that "nothing in this section establishes a right to have members of a protected class elected in numbers equal to their proportion in the population." 42 U.S.C. § 1973(b).

Gingles prompted the Supreme Court's first interpretation of the amendment to section 2. The Gingles plaintiffs brought a section 2 challenge against the inclusion of several multimember state legislative districts in North Carolina's redistricting plan. Gingles, 478 U.S. at 35, 106 S.Ct. at 2758.^[75] Upon analyzing the statute, the Court confirmed that Congress indeed shifted plaintiffs' burden from proving discriminatory intent to showing discriminatory effects: "The `right' question ... is `whether as a result of the challenged practice or structure plaintiffs do not have an equal opportunity to participate in the political processes and to elect candidates of their choice.'" Gingles, 478 U.S. at 44, 106 S.Ct. at 2763 (quoting S. REP. 97-417, at 28 (1982), reprinted in 1982 U.S.C.C.A.N. 177, 206).^[76] Proving the denial of the ability to participate in the political process and to elect candidates of choice, the Court held, required more than a mere showing of the factors enunciated in Zimmer v. McKeithen and listed in the Senate Report. Id. at 48, 106 S.Ct. at 2765. "While many or all of [those] factors ... may be relevant ... unless there is a conjunction of the following circumstances, the use of multi-member districts generally will not impede the ability of minority voters to elect representatives of their choice": (1) the minority group "is sufficiently large and geographically compact to constitute a majority in a single-member district"; (2) the minority group is "politically cohesive"; and (3) the majority group "votes sufficiently as a bloc to enable it ... usually to defeat the minority's preferred candidate." Id. at 48-51, 106 S.Ct. at 2765-67.

According to the Court, these three circumstances are "necessary preconditions" to a vote dilution claim. Id. at 50, 106 S.Ct. at 2766. If a minority group cannot demonstrate its geographic compactness, for example, then "the multi-member form of the district cannot be responsible for minority voters' inability to elect its [sic] candidates." Id. (emphasis in original). That is to say, "if the minority group is spread evenly throughout the multimember district, or if, although geographically compact, the minority group is so small in relation to the surrounding white population that it could not constitute a majority in a single-member district, these minority voters cannot maintain that they would have been able to elect representatives of their choice in the absence of the multi-member electoral structure." Id. at 50 n. 17, 106 S.Ct. at 2766 n. 17. The purpose of inquiring into the second and third preconditions, which together demonstrate racially polarized voting, is similarly used to establish that it is the electoral structure itself which actually "thwarts distinctive minority group interests" and "impedes its ability to elect its chosen representatives." Id. at 51, 106 S.Ct. at 2766-67. Thus, the confluence of the three Gingles preconditions establishes a causal link between the challenged electoral scheme and the vote dilution injury plaintiffs allege.

In a subsequent case, Johnson v. De Grandy, the Court explained that the three Gingles preconditions, while necessary to prove a section 2 claim, are not sufficient. 512 U.S. 997, 1011, 114 S.Ct. 2647, 2657, 129 L.Ed.2d 775 (1994). This was true in Gingles "because the ultimate conclusions about equality or inequality of [a minority group's political] opportunity were intended by Congress to be judgments resting on comprehensive, not limited, canvassing of relevant facts." Id. Courts, therefore, after determining that the three Gingles preconditions exist, should then "examine other evidence in the totality of the circumstances, including the extent of the opportunities minority voters enjoy to participate in the political processes." Id. at 1011-12, 114 S.Ct. at 2657; see also Gingles, 478 U.S. at 80, 106 S.Ct. at 2781 (citing the district court's findings under the totality of the circumstances test, specifically, "the legacy of official discrimination in voting matters, education, housing, employment, and health services; and the persistence of campaign appeals to racial prejudice"); Negron v. City of Miami Beach, 113 F.3d 1563, 1566 (11th Cir. 1997) (holding that the Gingles preconditions are "not the end of the story").

C. POLITICAL GERRYMANDERING

Applying the Gingles factors to political gerrymandering claims provides the most intelligible way of determining which claims present an Equal Protection Clause violation. The basic test for these claims, as set forth by the Court in Bandemer, requires plaintiffs to establish a legislature's discriminatory intent in redistricting and a history (past or predicted) of actual adverse effects resulting from that intent. Bandemer, 478 U.S. at 127, 139, 106 S.Ct. at 2808, 2814. This test is therefore merely a restatement of the test established in the racial vote dilution cases explored above. As in the racial cases, Bandemer instructs that the focus of our inquiry should be an examination of the effects—specifically, "whether a particular group has been unconstitutionally denied its chance to effectively influence the political process." Id. at 132-33, 106 S.Ct. at 2810. As we explain below, the three Gingles factors are necessary preconditions to any showing of this effect by a political group. We therefore hold that in order to succeed on a claim of political gerrymandering brought under the Equal Protection Clause, a political group must first set forth evidence of (1) geographical compactness, (2) political cohesiveness of that group, and (3) bloc voting on the part of the majority group.^[77]Gingles, 478 U.S. at 50-51, 106 S.Ct. at 2766. Furthermore, we hold that once these three factors are satisfied, the plaintiffs must then establish (4) that the totality of the circumstances indicates vote dilution. De Grandy, 512 U.S. at 1011-12, 114 S.Ct. at 2657.

1. RATIONALE

We adopt the standards of Gingles and De Grandy to our analysis of the instant equal protection claim for four reasons. First, without an initial showing of the three factors discussed in Gingles, a plaintiff cannot demonstrate that any adverse effect actually results from the challenged legislation, no matter what other circumstantial factors of vote dilution may be alleged. See Gingles, 478 U.S. at 48-50, 106 S.Ct. at 2765-66. It is, of course, axiomatic that a plaintiff claiming a violation of his equal protection rights must prove that his injury is caused by the challenged government action. The confluence of the three Gingles factors demonstrates that the minority group has the potential, absent discriminatory interference from the majority group, to participate in the political process and that it is the discriminatory interference which prevents the minority group from doing so and thereby causes its injury. See id. at 50 n. 17, 106 S.Ct. at 2767 n. 17.

The requirement of causation as an element of a Fourteenth Amendment vote dilution claim was expressly recognized by the Eleventh Circuit in Johnson v. DeSoto County Board of Commissioners, 204 F.3d 1335, 1345-46 (11th Cir.2000). In DeSoto, the court held that a constitutional vote dilution claim is made out upon a showing that "(1) the county's black population lacks an equal opportunity to participate in the political process and elect candidates of its choice; (2) this inequality of opportunity results from the county's at-large voting scheme; and (3) a racially discriminatory purpose underlies [that] scheme." Id. at 1345. The court, recognizing that this test may appear to depart from precedent requiring merely a showing of discriminatory effects plus intent, cautioned that no such change was in fact intended. Id. at 1345 n. 20. Rather than adding a causation element to the test, the court held that "the `discriminatory effects' requirement encompasses both inequality of opportunity and a causation element." Id. The separation of these subcomponents into separate elements was merely done to "promote[] clarity." Id. Similarly, here, we do not seek to change the well-developed standards governing constitutional vote dilution claims. We merely recognize the burden placed on plaintiffs of establishing a causal link between their injury and the challenged legislation and adopt a standard by which a court can determine whether the plaintiffs have borne this burden.^[78]

Second, although political cohesiveness, bloc voting, and geographic compactness were specified as preconditions to claims brought under section 2 of the Voting Rights Act in Gingles, they have always figured prominently in the analysis of vote dilution claims brought under the Equal Protection Clause. See, e.g., Rogers, 458 U.S. at 623, 102 S.Ct. at 3279 (recognizing the importance of proof of bloc voting— "without bloc voting the minority candidates would not lose elections solely because of their race."); id. at 616, 102 S.Ct. at 3275 (recognizing dilution occurs where a "distinct minority" would otherwise be able to elect a representative in a single-member district but is instead submerged in a multimember district characterized by bloc voting); Nevett, 571 F.2d at 223 n. 16 ("In the absence of polarized voting, black candidates could not be denied office because they were black, and a case of unconstitutional dilution could not be made."). In fact, the congressional amendment to section 2 of the Voting Rights Act, which precipitated the Court's holding in Gingles, merely codified the standards that had been developed for testing racial vote dilution claims brought under the Equal Protection Clause in Whitcomb and White. See S. REP. 97-417, at 28 (1982), reprinted in 1982 U.S.C.C.A.N. 177, 206 (adopting the results test of Whitcomb and White and rejecting the intent test of Bolden). The result is that the same "general standard, governing proof of injury, [controls] both section 2 and constitutional vote dilution cases." DeSoto, 204 F.3d at 1344. Just as the Gingles factors were imported into the section 2 context from the Court's constitutional vote dilution jurisprudence, so, too, is it now reasonable to bring them back to our equal protection analysis.

Third, it is only logical to conclude that if the Gingles factors are required of racial groups seeking to succeed on statutory vote dilution claims, so, too, must they be required of political groups seeking to succeed on similar constitutional claims. To hold otherwise would be to render it easier for a political group than for a racial minority to obtain protection from electoral discrimination. In other words, such a rule would make political groups more equal than racial groups. The Equal Protection Clause simply makes no room for such a holding. See Zobel v. Williams, 457 U.S. 55, 71, 102 S.Ct. 2309, 2319, 72 L.Ed.2d 672 (1982) (Brennan, J., concurring).

Fourth and finally, the language of Bandemer itself necessitates proof of the three Gingles factors. The plurality requires proof of (1) "intentional discrimination against an identifiable political group" and (2) "an actual discriminatory effect on that group." Bandemer, 478 U.S. at 127, 106 S.Ct. at 2808 (emphasis added). The reality of American politics is that "voters sometimes prefer Democratic candidates, and sometimes Republican," without regard to the party affiliation noted on their voter registration cards. Id. at 135, 106 S.Ct. at 2812. To identify the members of any one political group, plaintiffs must therefore provide proof of more than just party registration. As Justice Stevens recognized in his concurrence to Bolden, "[i]t is only when [a group's] common interests are strong enough to be manifested in political action that the need [for protection against gerrymandering] arises. For the political strength of a group ... is a function of numbers—specifically the number of persons who will vote in the same way." 446 U.S. at 88, 100 S.Ct. at 1511 (Stevens, J., concurring in judgment). In the political gerrymandering context, the best evidence of political identity will thus necessarily consist of proof of bloc voting and political cohesiveness.^[79] These factors are the "glue" that binds together voters and makes them an identifiable political group.

As with claims of racially polarized voting, establishing politically polarized voting requires both a showing of political cohesiveness among members of the minority political group and a showing of bloc voting by the majority group. Gingles, 478 U.S. at 56, 106 S.Ct. at 2769. These elements, in turn, may most easily be established by a showing that the members of each group tend to vote for the same candidate and that the candidate supported by each group is different from that supported by the other. Id. This examination requires evidence of cohesiveness and bloc voting over some significant period of time^[80] and may not be satisfied merely with proof of the minority party's actual or predicted loss in a single election. Id. at 56, 106 S.Ct. at 2769. Likewise, the success of the minority party in one election will not defeat a claim of polarized voting. Id. at 57, 106 S.Ct. at 2770.

Of course, once the identifiable political group has demonstrated, through evidence of these three preconditions, that they have the potential to participate as a group in the political process and elect legislators of their choice, they must also establish the intent and actual effects elements expressly required under Bandemer and all vote dilution claims brought under the Equal Protection Clause.

2. INTENT & EFFECT

The Bandemer plurality's analysis of the first prong of its two-step test, the intent requirement, sets a low bar for plaintiffs asserting political gerrymandering claims. Recognizing the reality that "whenever a legislature redistricts, those responsible for the legislation will know the likely political composition of the new districts and will have a prediction as to whether a particular district is a safe one for a Democratic or Republican candidate or is a competitive district that either candidate might win," Bandemer, 478 U.S. at 128, 106 S.Ct. at 2808, the Court concluded that "it should not be very difficult to prove that the likely political consequences of the reapportionment were intended," id. at 129, 106 S.Ct. at 2809. However, the Court was careful to point out that just because it should not be difficult to prove politically discriminatory intent in challenges to legislative redistricting "does not, of course, mean that it need not be proved at all to succeed on such a claim." Id. at 129 n. 11, 106 S.Ct. at 2809 n. 11.

The Bandemer plaintiffs established this element to the satisfaction of the district court, which emphasized the testimony of the Speaker of the Indiana House in its decision finding in favor of the plaintiffs on the political gerrymandering claim. Bandemer v. Davis, 603 F.Supp. 1479, 1484 (S.D.Ind.1984). The Republican Speaker provided direct evidence of this intent, admitting that the reasons for structuring certain districts were "[p]olitical" and that the Republicans who drew those districts "wanted to save as many incumbent Republicans as possible." Id. The Supreme Court upheld the district court's determination of discriminatory intent as not clearly erroneous. Bandemer, 478 U.S. at 127, 106 S.Ct. at 2808.

The Court then proceeded to examine the adverse effects imposed upon the political minority by the redistricting plan. As in the racial vote dilution cases, the focus of this examination was on determining whether or not the challenged legislation operated "to cancel out or minimize the voting strength" of Indiana Democrats. White, 412 U.S. at 765, 93 S.Ct. at 2339. Thus, as in the racial cases, the Court was looking for proof that Democrats had been fenced out of the political process; in other words, that they "had less opportunity than did other residents in the district to participate in the political processes and to elect legislators of their choice." Id. at 766, 93 S.Ct. at 2339 (citing Whitcomb, 403 U.S. at 149-50, 91 S.Ct. at 1872). Since Bandemer does not explicitly establish the proof required to show a fencing out of one political party by another in the voting context, we look to the plurality's opinion and the cases it relies on for the necessary elaboration.

First, the Court was explicit in squarely rejecting any possibility that "a reapportionment law would violate the Equal Protection Clause because the voters in the losing party do not have representation in the legislature in proportion to the statewide vote received by their party candidates." Bandemer, 478 U.S. at 130, 106 S.Ct. at 2809. That more is required than a mere showing of disproportionate election results to establish an equal protection violation is well established in the line of racial vote dilution cases, see, e.g., Whitcomb, 403 U.S. at 160, 91 S.Ct. at 1878, and was also held to apply to claims brought under section 2 of the Voting Rights Act, see Gingles, 478 U.S. at 57, 106 S.Ct. at 2769. This remains true even where disproportionate results are gross or persist year after year; the Court has "not yet deemed it a denial of equal protection to deny legislative seats to losing candidates, even in those so-called `safe' districts where the same party wins year after year." Whitcomb, 403 U.S. at 153, 91 S.Ct. at 1874. As a result, where groups cannot present evidence of "fencing out" beyond mere disproportion, we can only conclude that "the failure of the [minority] to have legislative seats in proportion to its populations [is more] a function of losing elections than of built-in bias against [that party]" and no equal protection claim may lie. Id.

That is not to say that a history—"actual or projected"—of disproportionate results is wholly irrelevant to the ultimate inquiry. On the contrary, the Bandemer Court explicitly recognized its relevance; however, it is only where such history is accompanied by "strong indicia of lack of political power and the denial of fair representation" that a claim of political vote dilution may succeed. 478 U.S. at 139, 106 S.Ct. at 2814. In other words, election results are but one inconclusive part of the ultimate showing—whether the minority party has been "denied its chance to effectively influence the political process." Id. at 132-33, 106 S.Ct. at 2810.

In addition to past or projected election results, therefore, plaintiffs must come forward with other evidence that their minority party had experienced a "less[er] opportunity to participate in the political processes and to elect candidates of their choice." Id. at 131, 106 S.Ct. at 2809 (quoting Rogers, 458 U.S. at 624, 102 S.Ct. at 3279). The Court arrived at this standard by distilling its precedent in racial vote dilution cases involving individual multimember districts. See id. This indication of vote dilution, the Court held, "would apply equally to claims by political groups in individual districts," id. at 131 n. 12, 106 S.Ct. at 2810 n. 12, and to claims of statewide political gerrymandering, see id. at 132, 106 S.Ct. at 2810.

The same race-based cases from which the Court derived the standard of vote dilution—the diminishment of a group's opportunity to participate in the political process and elect candidates of its choice— also provide us with a framework for evaluating proof of that effect. See, e.g., Rogers, 458 U.S. at 624, 102 S.Ct. at 3279; United Jewish Org's v. Carey, 430 U.S. 144, 167-68, 97 S.Ct. 996, 1010-11, 51 L.Ed.2d 229 (1977); White, 412 U.S. at 765-66, 93 S.Ct. at 2339; Whitcomb, 403 U.S. at 149-50, 91 S.Ct. at 1872. Those cases instruct us to look beyond evidence of numerosity and geographic compactness and bloc voting to other evidence that, in the totality of the circumstances, supports an ultimate determination that legislation causes unconstitutional vote dilution. See, e.g., Rogers, 458 U.S. at 624, 102 S.Ct. at 3279. In our examination of the totality of the circumstances, the Court directs us to the factors developed by the Fifth Circuit in Zimmer and Nevett.^[81]See id.

Interestingly, the Court in Bandemer stresses only one of the Zimmer factors— "the lack of responsiveness by those elected to the concerns of the relevant groups." Bandemer, 478 U.S. at 131, 106 S.Ct. at 2810. While the Court acknowledged, at least in passing, the other factors traditionally considered in vote dilution cases, the Court recognized that political minorities would be less able to establish many of the types of circumstantial evidence of vote dilution often alleged by racial minorities. Id. at 131 n. 12, 106 S.Ct. at 2810 n. 12. Specifically, the Court recognized that political minorities would likely not be able to point to a past pattern of exclusion from the voting booth and political office to support their dilution claims. Id. Not only will many of the other Zimmer factors be difficult for political parties to establish, the most likely cause for concern among members of the minority party will be that the elected representative is or will be unresponsive to their concerns.

Although evidence of lack of responsiveness may be asserted more easily than other factors, it is unclear how that would be shown in the political gerrymandering context. The Bandemer plurality clearly instructs that we may not presume that Republicans have been or will be deaf to the concerns of the Democrats. On the contrary, "a group of individuals who votes for a losing candidate is usually deemed to be adequately represented by the winning candidate and to have as much opportunity to influence that candidate as other voters in the district." Id. at 132, 106 S.Ct. at 2810.^[82] It has likewise long been accepted that "[t]he mere fact that one interest group or another ... has found itself out-voted and without legislative seats of its own provides no basis for invoking constitutional remedies where ... there is no indication that this segment of the population is being denied access to the political system." Whitcomb, 403 U.S. at 154-55, 91 S.Ct. at 1875.

The bar to establishing lack of responsiveness will only be heightened when the challenged redistricting is of congressional districts. Members of Congress, unlike state and local representatives, are not engaged in the business of providing constituents with government services, such as fire and police protection, schools, utilities, and road improvements. Where a state or local body refuses to provide one group with such services, the evidence of lack of responsiveness will be apparent. See Rogers, 458 U.S. at 626, 102 S.Ct. at 3280. To allege that a member of Congress is being unresponsive to one political group or the other, on the other hand, is to claim that he is failing to promote that group's political views. As with any minority in our democratic system founded on majority rule, the rejection of its views could not raise constitutional concerns. To hold otherwise would be tantamount to holding that "it is invidiously discriminatory for a [political subdivision] to elect its delegation by a majority vote based on party or candidate platforms and so to some extent predetermine legislative votes on particular issues." Whitcomb, 403 U.S. at 155, 91 S.Ct. at 1875. That the majority may reject the views of the minority is "inherent in government by elected representatives." Id.

Finally, even if one political party could establish a history of disproportionate results in conjunction with some other evidence in the totality of the circumstances of unconstitutional vote dilution, no group will be able to prove a lack of ability to participate in the political process. Any such claim will necessarily be refuted by the logical inference to be drawn from a disparity between voting age population and registered voters or between registered voters and voter turnout. The way political groups participate in the process is by registering voters and persuading them to support their policies and candidates. A failure to do these things will be the actual cause of any inability on the group's part to elect candidates of its choice and have those candidates represent its interests.

D. RULE APPLICATION

Turning now to the case at bar, we analyze the evidence offered by the plaintiffs and intervenors to determine whether it rises to the level of proof sufficient to make out an equal protection claim—that is, whether the evidence demonstrates "both intentional discrimination against an identifiable political group and an actual discriminatory effect on that group." Bandemer, 478 U.S. at 127, 106 S.Ct. at 2808.

1. INTENT

Like the Court in Bandemer, we have no trouble at all finding that the intent of the Florida legislature, comprised of a majority of Republicans, was to draw the congressional districts in a way that advantages Republican incumbents and potential candidates. Indeed, at trial, the defendants stipulated as much. Moreover, the Bandemer Court made clear that, "[a]s long as redistricting is done by a legislature, it should not be very difficult to prove that the likely political consequences of the reapportionment were intended." Id. at 129, 106 S.Ct. at 2809.

2. EFFECTS

We do not accept, however, the plaintiffs' and intervenors' arguments that the effect of the redistricting scheme is "sufficiently serious to require intervention by the federal courts." Id. at 134, 106 S.Ct. at 2811. In particular, these parties have failed to prove that Florida Democrats will have "less opportunity to participate in the political processes and to elect candidates of their choice" in congressional districts as a result of the enacted redistricting plan. Id. at 131, 106 S.Ct. at 2809 (quoting Rogers, 458 U.S. at 624, 102 S.Ct. at 3279).

a. GINGLES PRECONDITIONS

As we have explained, we must begin our analysis with a consideration of the three Gingles preconditions—geographic compactness, political cohesiveness, and bloc voting on behalf of the majority group. To properly consider these factors, however, we must first understand the political group to whom they apply. In their complaints, the plaintiffs and intervenors identify the relevant group in several ways: "registered Democrats," "members of the Florida Democratic Party," and "Democratic voters."^[83] At trial, the parties submitted evidence of the number of registered Democrats in each congressional district as well as information about various elections, indicating how many voters actually voted Democratic. It is unclear to us which group the complaining parties would have us consider—registered Democrats or those voters who consistently vote Democratic. Thus, we take each group in turn because, as we demonstrate, registered Democrats and true Democratic voters are not necessarily the same in the State of Florida.

In treating registered Democrats as the identifiable political group, the plaintiffs and intervenors offer only that Democrats make up a slight majority of registered voters statewide but a minority of registered voters in fifteen—well over half—of the proposed congressional districts.^[84] Thus, they claim, the gerrymandering underlying the enacted plan unconstitutionally dilutes the voting strength of registered Democrats. This assumption is not trustworthy, however, without proof that registered Democrats are politically cohesive. In other words, simply showing that registered Republicans outnumber registered Democrats—even in a great majority of the districts—does not prove that Democrats will always lose, much less prove that Democrats can no longer influence the political process.

Indeed, the evidence demonstrates that registered Democrats are not a politically cohesive group because they often vote Republican. Over the last decade, for example, a plurality of the registered voters in CDs 1, 6, 10, 12, and 22, was Democratic, but Republican candidates were elected.^[85] The starkest example is CD 12 (as drawn in 1996), in which registered Democrats have outnumbered Republicans at least since 1996 by an average of 49.2% Democrat to 38.5% Republican. In the three most recent general elections for Congress, however, Republican candidates have won—Charles Canady winning by a wide margin in 1996 with almost 66.6% of the vote and again in 1998 with 100% of the vote, and Adam Putnam winning safely in 2000 with 57% of the vote. Statewide, registered Democrats are likewise unpredictable, often voting for Republicans. For example, in the race for State Treasurer, a low-profile race between two non-incumbents, the Republican candidate won by a margin of 18%. In another low-profile, state-wide 1992 race, the Republican candidate for State Comptroller, Bob Milligan, defeated the six-term, Democratic incumbent by a slight margin. Four years later, Milligan beat the Democratic candidate by 21.2%.^[86] Without proof that registered Democrats are politically cohesive, we simply cannot assume otherwise.

It is much more likely, therefore, that the plaintiffs' and intervenors' claim was asserted on behalf of true Democrats— those voters who consistently vote Democratic.^[87] In attempting to identify this group, the plaintiffs and intervenors submitted evidence of a "party voting index," computed from twenty-three state-wide general election contests from 1992 through 2000.^[88] For each election the Republican and Democratic share of the two-party vote within a given district was computed; these vote percentages were then averaged over the twenty-three elections to yield a broadly based index of the voting tendency of any given district. Overall, the mean voting index for the entire state of Florida is 49.0% Republican and 51.0% Democratic. Nevertheless, under the newly enacted plan, eighteen of the districts have an average Republican vote share over 50%, and only seven have a Democratic vote share exceeding 50%. This disparity, argue the plaintiffs and intervenors, is unconstitutional.

The voting index, however, does not prove that there is actually a group of true Democrats who consistently vote Democratic.^[89] Indeed, the index does not identify a group of voters at all; it merely reflects a result without explaining which voters produced that result. It is not necessarily true, for example, that the voters who voted for Democrats (or Republicans) in a given election were the same people voting for Democrats (or Republicans) in other elections. It is quite possible that the group of voters on one side of the political fence in one election will be on the opposite side in the next election. Thus, asserting that in District 22 the average Democratic vote share was 56.3% does not necessarily mean that 56.3% of voters in that district consistently vote Democratic.

Even if the voting index could be tied to an identifiable group of voters, it is unreliable in proving political cohesiveness because it does not include congressional races—the very races that are most relevant in this case. For example, the vote index shows that, in Districts 10 and 22, the average Democratic vote share exceeds 50%, which, according to the plaintiffs and intervenors, means that a Democratic candidate will win.^[90] Over the last decade, however, in each of those districts, a Republican candidate has won the congressional general election.^[91] Conversely, in Districts 1 and 15, two districts that, according to the vote index, should perform for a Republican candidate, Democratic congressional representatives have been elected.^[92] If the voting index demonstrates what the plaintiffs and intervenors claim—that each voting average represents a group of voters that consistently votes for that party—then the candidates in the four aforementioned districts could not possibly have been elected. That they were elected proves that the voting index is unreliable: voters often cross-over and vote for candidates of other parties.

The plaintiffs and intervenors simply have not shown the political cohesiveness of an identifiable political group. At times, they identified a group—registered Democrats —but failed to establish the group's political cohesiveness. At other times, they submitted election results, typical proof of political cohesiveness, but failed to identify the group that produced those results. Without proof of cohesiveness, however, the parties' claim must fail because "it cannot be said that the [redistricting scheme] thwarts distinctive minority group interests." Gingles, 478 U.S. at 51, 106 S.Ct. at 2766.

Even if it is assumed that such evidence demonstrates a politically-cohesive group of true Democrats, we nonetheless find that the parties' claim fails primarily because there is so little evidence of geographic compactness. That is to say, the parties have not shown us that the true Democrats are situated in compact areas so that we may redraw the districts in their favor. Without proof of compactness, there is no way for us to fashion a remedy, even if the parties prevail on the other two Gingles preconditions. See id. at 50 n. 17, 106 S.Ct. at 2766 n. 17. The only evidence submitted on this point was through the testimony of Congressman Peter Deutsch, who testified that there is one area near Tampa Bay with a high percentage of true Democrats. According to Congressman Deutsch, the legislature plotted the lines in such a way that 69,909 residents (of whom almost all were black Democrats) were put into District 11, thereby removing a sufficient number of Democratic voters from District 10 to restore a Republican margin in the 10th District and packing Democrats into already-Democratic District 11.

Congressman Deutsch's testimony, however, highlights only one small area in the entire state where Democrats are geographically compact.^[93] There was no similar showing with regard to the other five or so districts the plaintiffs and intervenors claim are manipulated in the Republican Party's favor. Moreover, the plaintiffs and intervenors submitted no evidence indicating where true Republicans reside nor how lines could be drawn in a way to pack those Republicans and disperse Democrats, thereby giving Democrats more districts. Just as courts have long held in the racial vote dilution cases, we find that, without more demonstrations like the one Congressman Deutsch made, there is no way for us to cure the very gerrymandering about which the parties complain.^[94]

Finally, we note that the plaintiffs' and intervenors' claim must fail because there was no proof that the majority group, presumably Republicans, consistently vote as a bloc to defeat Democratic candidates. Such a showing is imperative in demonstrating that "[the redistricting scheme] impedes [true Democrats'] ability to elect [their] chosen representatives." Gingles, 478 U.S. at 51, 106 S.Ct. at 2767. The only evidence submitted was information about registered Republicans and the voting index for "Republicans." For all the reasons previously articulated, however, such evidence is insufficient proof that Republicans —whether registered Republicans or "true" Republicans—is an identifiable political group that consistently votes as a bloc. Without such evidence, we cannot assume that, just because true Democrats may be a minority group, they will most often lose. It may very well be that the majority group is politically fractured, with voters sometimes voting Republican, Democrat, or Independent.

b. TOTALITY OF THE CIRCUMSTANCES

Even if the plaintiffs and intervenors had successfully established that true Democrats are geographically compact and politically cohesive and that Republicans vote as a bloc to defeat their candidates, their claim fails because they have not shown that, under the totality of the circumstances, the redistricting scheme gives true Democrats "less opportunity to participate in the political processes and to elect candidates of their choice." Bandemer, 478 U.S. at 131, 106 S.Ct. at 2809 (quoting Rogers, 458 U.S. at 624, 102 S.Ct. at 3279).

In particular, there has been no evidence regarding true Democrats' inaccessibility to political processes (such as the slating of candidates), the weight of the state policy behind the redistricting scheme,^[95] nor the effect of past discrimination on true Democrats' ability to participate in the election system. Finally, as we predicted in subpart III.C.2, the plaintiffs and intervenors submitted no evidence to prove that Republican representatives are unresponsive to the "particularized interests" of true Democrats.^[96]

On the contrary, a review of the totality of the circumstances reveals a proposition quite opposite the one argued by the plaintiffs and intervenors: true Florida Democrats possess—and will continue to possess—just as much influence over the political process as Florida Republicans. Indeed, in the state of Florida, there is a total voting age population of roughly 12,338,395, of whom only 8,880,396^[97] are registered to vote. The plaintiffs and intervenors have offered nothing to explain why they cannot influence these 3,457,999 potential voters to register as Democrats and consistently vote Democratic. Of the eighteen districts that the plaintiffs and intervenors contend will perform for a Republican candidate, all eighteen districts have a gap between voting age population and registered voters significant enough to make a difference for the Democratic Party. If the plaintiffs and intervenors are as motivated to go out in the hustings and garner more voters as they are to litigate this case, then the outcome of future congressional elections is by no means certain.

Even those voters who have registered with one party and consistently vote for that party are convertible. Political affiliation, unlike race, is not static. In this democracy of ours, where the right to speak and associate is heavily protected, groups are able to communicate their views and persuade others to join them. Indeed, the Florida Republican Party appears to have done just that in the last decade. In 1992, for example, the Florida legislature was dominated by Democrats, with a 61.6% majority in the Florida House and a 52.5% majority in the Florida Senate. Ten years later, however, that political make-up flip-flopped, as voters began to elect more Republicans. Today, the Florida House is 64.1% Republican and the Florida Senate is 62.5% Republican.^[98]

3. PLAINTIFF'S "TOO FAR" ARGUMENT

The plaintiffs and intervenors, however, submit that the redistricting scheme here is unconstitutional because "[i]n [Bandemer], the partisan bias was not as extreme and the projected duration not as long as they are in this case."^[99] Essentially, they contend that the enacted redistricting plan violates the Equal Protection Clause because its discriminatory effects are greater than the reapportionment plan at issue in Bandemer. The primary support for this assertion is the disparity between Republican and Democratic districts. In the State of Florida, where the partisan division is near 50/50, they argue, the redistricting plan is unconstitutional because 18 of the proposed districts have an average Republican vote share over 50%; whereas, the average Democratic vote share exceeds 50% in only seven of the proposed districts. At oral argument, counsel for plaintiff-intervenor Congressman Deutsch urged us to establish the "belt line," a point at which disproportionate representation goes "too far" in this case and becomes unconstitutional. In the instant case, counsel presumably contends, the 18/7 ratio exceeds that line.

We fail to see how this argument is any different from the one explicitly rejected by the Court in Bandemer and in just about every other racial vote dilution case the Court has heard. "Our cases ... clearly foreclose any claim that the Constitution requires proportional representation or that legislatures in reapportioning must draw district lines to come as near as possible to allocating seats to the contending parties in proportion to what their anticipated statewide vote will be." Bandemer, 478 U.S. at 130, 106 S.Ct. at 2809 (citing Whitcomb, 403 U.S. at 153, 156, 160, 91 S.Ct. at 1874, 1876, 1876, and White, 412 U.S. at 765-66, 93 S.Ct. at 2339-40). Notwithstanding the plaintiffs' and intervenors' contention, disproportionate representation—no matter how severe —cannot by itself rise to the level of an equal protection claim. "Only where there is evidence that excluded groups have `less opportunity to participate in the political processes and to elect candidates of their choice'" may we refuse to approve the redistricting scheme. Id. at 131, 106 S.Ct. at 2809 (quoting Rogers, 458 U.S. at 624, 102 S.Ct. at 3279). "[W]e have found equal protection violations only where a history of disproportionate results appeared in conjunction with strong indicia of lack of political power and the denial of fair representation." Id. at 139, 106 S.Ct. at 2814.

For the foregoing reasons, we deny the plaintiffs' and intervenors' claims of political gerrymandering.

Conclusion

As explained above, we conclude that the plaintiffs and intervenors have failed to establish their claims. Final judgments will be issued separately.

HINKLE, District Judge, concurring.

I concur in the court's opinion and judgment and offer some additional views on political gerrymandering. For purposes of this concurring opinion, I define "political gerrymandering" as the drawing of legislative district lines in a manner intended to benefit candidates of one party and thereby provide that party greater representation than would be achieved through the drawing of lines without regard to partisan political consequences.

The court's opinion is a thorough and correct analysis of the political gerrymandering claims at issue under principles originally developed in the context of racial vote dilution. I write separately in order (1) to note why I believe this is the correct ultimate standard under existing Supreme Court decisions, (2) to note what I believe is the most critical step in the application of this standard to the facts of this case, and (3) to respond to a defendant's assertion that the degree of political gerrymandering exhibited in these plans should be a source of pride.

I

As an original matter, courts could take three possible approaches to political gerrymandering claims.

First, courts could reject such claims in all circumstances, either on grounds that the Constitution does not proscribe political gerrymandering, no matter how severe, or on grounds of nonjusticiability. Such an approach, however, would not comport with accepted constitutional principles. The Constitution does and should protect not just the right to cast a ballot, but also the right to a meaningful opportunity to have that ballot affect the political process. The cases collected in the court's opinion, primarily in the racial vote dilution context, are an application of this principle. Thus a state legislature cannot constitutionally deny a voter or group of voters all opportunity to affect the political process, and for this purpose it matters not whether the method of denial is political gerrymandering or some other device. Further, while articulating and enforcing standards for protection of this constitutional right is no easy task, it is not beyond the competence of the federal courts, as the Supreme Court recognized in Davis v. Bandemer, 478 U.S. 109, 106 S.Ct. 2797, 92 L.Ed.2d 85 (1986).

Second, courts could hold political gerrymandering unconstitutional, period, and provide remedies requiring the practice to stop. Such an approach would not be without at least superficial support in our jurisprudence; thus, for example, partisan political considerations cannot be the basis for hiring or firing most rank and file government workers, and no one would doubt the unconstitutionality of a state's efforts to parcel out most public benefits— medicaid payments come to mind—based on an applicant's party affiliation. It thus may not be immediately apparent why it is permissible for a state legislature to draw district lines that afford some candidates an advantage, and saddle others with a disadvantage, based solely on the candidates' party affiliation. But partisan political considerations long have affected the drawing of district lines. It is unlikely the Framers intended to make this practice unconstitutional. Nor should the redistricting process be managed by federal courts to the extent that would result from any effort to enforce an absolute ban on political gerrymandering. Perhaps for these reasons, the Supreme Court has made clear, sometimes without even discussing the issue, that political gerrymandering is not, without more, unconstitutional. See, e.g., Easley v. Cromartie, 532 U.S. 234, 121 S.Ct. 1452, 149 L.Ed.2d 430 (2001) (holding that an irregularly shaped district that would have been unconstitutional if drawn based on voters' race was constitutional because drawn based on voters' party loyalty).^[1]

Third, courts could take a middle approach, holding political gerrymandering permissible to a point, but concluding that at some point there is a constitutional limit. Both sides seem to embrace this approach in this case, perhaps because, as set forth above, the Supreme Court's decisions appear to foreclose the alternatives. The issue is how much is too much. The answer corresponds with the right that is at stake: the right to a meaningful opportunity to have one's ballot affect the political process. This is the same right at issue in the racial vote dilution cases. See, e.g., Thornburg v. Gingles, 478 U.S. 30, 44, 106 S.Ct. 2752, 2763, 92 L.Ed.2d 25 (1986). The court's opinion in the case at bar thoroughly analyzes and correctly applies the standards derived from those cases.^[2]

II

The Florida Legislature sought to draw district lines as favorably for Republican candidates as possible (as favorably, that is, as could be done consistently with other constraints imposed by state and federal law). From all indications, the legislature did an outstanding job; in a state with a notoriously close division of Democratic and Republican voters statewide, 18 of the 25 congressional districts have been drawn to cover areas in which voters have exhibited a clear voting preference for Republicans.^[3] A similar pattern is present for the State Senate and House. In the short run, there will undoubtedly be instances in which the drawing of the lines will affect the outcome. Republicans will win some seats that Democrats would have won had the lines been drawn without regard to politics. Were these results written in stone, I would have grave doubts about the constitutionality of these reapportionment plans.

But the results are not written in stone. The Democrats controlled the process in 1992 and undoubtedly attempted to draw district lines favorably for Democratic candidates. The Democrats may have cut the Republicans more slack, though this record gives no hint of that, or the Democrats may not have been quite as good at their craft. Any differences in this respect between 1992 and 2002, however, were surely minor. The result during the decade when the 1992 plans were in effect, far from what the Democrats hoped for or intended, was a dramatic shift from Democratic to Republican control. This happened not because of where the district lines were drawn, but because of a quantum shift in the partisan voting behavior of Floridians.

Voters move into and out of Florida daily. They move within the state. Immigrants become citizens and register to vote. Other unregistered but eligible voters choose to register for the first time. Children reach voting age, and other voters die. Voters change their party loyalties. And at least as importantly, they vote not always for the candidate that shares their political party affiliation, but for the candidate they find most acceptable given the issues and personalities involved in a given race. If this is not the nation's most dynamic state politically, it is surely close. If Republicans are still in control in 2012, it will be because, in the majority of races, their candidate beat the Democratic candidate on the merits, not because the 2002 Florida Legislature accurately predicted demographic changes and voting behavior ten years in advance.

In short, Democratic voters have not been denied a meaningful opportunity to elect candidates of their choice. They will elect many candidates of their choice; others will lose. Who will win more often, Democrats or Republicans, will depend on the myriad factors that drive electoral politics.

III

That we have correctly decided this case, in light of the existing Supreme Court precedents, seems clear. This does not mean, however, that the degree of political gerrymandering exhibited in this case is to be applauded. One of the defendants asserted in closing argument that Floridians should be justly proud of the apportionment plans adopted in this case. Floridians should be proud, perhaps, of the innovative FREDS software and the technical proficiency with which it has been applied, but not of the degree of political gerrymandering that drove these plans. Floridians should perhaps be proud, instead, when an apportionment plan is adopted in an effort to be fair to all, rather than in an effort to optimize the outcome for one party or the other.

Political gerrymandering is as old as the republican form of government. The Framers assigned the responsibility for congressional apportionment to the state legislatures, undoubtedly knowing that partisan politics would play a role. Still, the issue has recently taken on a different dimension; the "sea change" of advancing technology, as one witness described it, has substantially increased the extent of successful political gerrymandering that is achievable, as a few minutes online with FREDS will confirm. Some states have turned to non-partisan commissions in an effort to take raw partisan politics out of the equation. Whether that works, and whether the results are better or more fair, are subjects for a different branch of government, or for academic study. The Constitution does not commit such decisions to the judiciary. The plans adopted by the Florida Legislature are constitutional.

JORDAN, District Judge, concurring.

I join the court's opinion and judgment, and write separately only to suggest that it is time for the Supreme Court to again address political gerrymandering.

In the 16 years since its decision in Davis v. Bandemer, 478 U.S. 109, 106 S.Ct. 2797, 92 L.Ed.2d 85 (1986), the Supreme Court has not revisited the issue of political gerrymandering. The result is that lower courts continue to struggle in an attempt to interpret and apply the "discriminatory effect" prong of the standard articulated by the Bandemer plurality. As one three-judge court recently (and correctly) put it, "the recondite standard enunciated in Bandemer offers little concrete guidance." Vieth v. Pennsylvania, 188 F.Supp.2d 532, 544 (M.D.Pa.2002).

Although some political gerrymandering claims have withstood motions to dismiss, see, e.g., Republican Party of North Carolina v. Martin, 980 F.2d 943, 956-59 (4th Cir.1992), rehearing en banc denied, 991 F.2d 1202 (1993) (three judges dissenting), apparently no such claims have ultimately succeeded (at least not in published opinions), thereby leading some to say that Bandemer "announced a liability standard for partisan gerrymanders that was essentially impossible to satisfy," Pamela Karlan, The Fire Next Time: Reapportionment After the 2000 Census, 50 Stan. L.Rev. 731, 737 (1998), and others to call for the elimination of the "discriminatory effect" prong under a theory of representational harm, see Megan Creek Frient, Similar Harm Means Similar Claims: Doing Away With Davis v. Bandemer's Discriminatory Effect Requirement in Political Gerrymandering Cases, 48 Case W. Res. L.Rev. 617, 644-48 (1998). The waters have become even more muddied in light of post-Bandemer cases explaining that political gerrymandering can be—for lack of a better term—a defense to a charge of racial gerrymandering. See, e.g., Hunt v. Cromartie, 526 U.S. 541, 551, 119 S.Ct. 1545, 1551, 143 L.Ed.2d 731 (1999) ("Our prior decisions have made clear that 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."); John Hart Ely, Gerrymanders: The Good, the Bad, and the Ugly, 50 Stan. L.Rev. 607, 621 (1998) ("It is true that by its impossibly high proof requirements the Court in Bandemer essentially eliminated political gerrymandering as a meaningful cause of action, but only after it had essentially declared the practice unconstitutional. And yet, a scant decade later, the dissenters in the cases under discussion stand ready to invoke it as an `innocent' alternative explanation of what appears to be racially motivated gerrymandering. Effectively making a practice nonjusticiable doesn't make it constitutional, especially when the Court has indicated that it isn't. Especially when the Court was right.").

The lower federal courts need guidance on this recurring issue. I urge the Supreme Court to note probable jurisdiction in this case or one of the other political gerrymandering cases arising from this electoral cycle and hear oral argument.

NOTES

[*] Honorable Gerald Bard Tjoflat, U.S. Circuit Judge for the Eleventh Circuit, sitting by designation; Honorable Robert L. Hinkle, U.S. District Judge for the Northern District of Florida, sitting by designation; and Honorable Adalberto Jordan, U.S. District Judge for the Southern District of Florida.

[1] Plaintiffs brought their claims, the case was tried, and this opinion was drafted in substantial part prior to the November 2002 elections. Some of the references in this opinion are to circumstances as they existed prior to those elections. To our knowledge, none of our findings as drafted prior to the November 2002 elections are inconsistent with anything that occurred in those elections.

[2] In this order we use "black" as shorthand for the census category in which respondents identify themselves as "Black, African Am. or Negro." We use "black" rather than "African American" to avoid any confusion regarding whether blacks who immigrated (or whose ancestors immigrated) to the United States from places other than Africa—for example, Haiti—are included; they are.

[3] Initially, plaintiffs also claimed vote dilution based on national origin; plaintiffs claimed the legislature intentionally favored Hispanics of Cuban descent against other Hispanics, diluting the votes of non-Cuban Hispanics. In fact, the software used in the reapportionment process did not even track data on whether Hispanics in any area were Cuban or non-Cuban. If, as plaintiffs assert, a higher percentage of Cuban than non-Cuban Hispanics were included in performing Hispanic districts, the cause was geography or politics, not discrimination. Cuban Hispanics in south Florida are more concentrated geographically (and thus more likely to be included in performing Hispanic districts), while non-Cuban Hispanics are more widely dispersed. In addition, a higher percentage of Cuban than non-Cuban Hispanics vote Republican, so that efforts to draw Hispanic districts-in a manner favoring Republicans might well result in the inclusion of more Cuban than non-Cuban Hispanics in such districts —because of politics, not national origin. In any event, plaintiffs now acknowledge that there is no evidence that any constitutionally-permissible alternative could be drawn that would meaningfully improve the relative treatment of Cuban and non-Cuban Hispanics. Plaintiffs no longer seek relief on this basis and thus we address it no further.

[4] The cases tried before the three-judge court were not the only actions relating to the legislature's redistricting plans. For example, different plaintiffs tried to challenge the congressional redistricting plan in both state and federal court.

On January 24, 2002, Representative Corrine Brown, Representative Alcee Hastings, Representative Carrie Meek—all members of the U.S. House of Representatives—and Sallie Stephens filed an action against the State of Florida, Governor Bush, Attorney General Robert Butterworth, and Secretary Katherine Harris in the 17th Judicial Circuit of Broward County challenging the congressional redistricting plan under federal law. See Complaint [D.E. 1 in Brown v. State of Florida, No. 02-60267-Civ-Jordan]. On March 27, 2002, after the action was removed to federal court, Speaker Tom Feeney and President John McKay were granted leave to intervene. See Order [D.E. 35 in Brown v. State of Florida, No. 02-60267-Civ-Jordan]. The plaintiffs voluntarily dismissed their federal claims on April 15, 2002. See Notice of Voluntary Dismissal [D.E. 61 in Brown v. State of Florida, No. 02-60267-Civ-Jordan].

Subsequently, these plaintiffs filed a new action in the 17th Judicial Circuit of Broward County challenging the congressional redistricting plan under state law. See Complaint [D.E. 1 in Brown v. State of Florida, No. 02-60459-Civ-Jordan]. This action was removed by Speaker Feeney to federal court. Upon motions filed by the plaintiffs and Attorney General Butterworth, however, the matter was remanded to state court on May 2, 2002 [D.E. 40 in Brown v. State of Florida, No. 02-60459-Civ-Jordan].

Following the remand, Secretary Harris again removed the action to federal court under 28 U.S.C. § 1443(2). See Notice of Removal [D.E. 1 in Brown v. State of Florida, No. 02-60689-Civ-Jordan]. The plaintiffs and Attorney General Butterworth moved to remand, and their motion was granted on June 4, 2002. See Brown v. State of Florida, 208 F.Supp.2d 1344 (S.D.Fla.2002). On June 17, 2002, the state court dismissed the action for lack of subject-matter jurisdiction. See Order of Dismissal in Brown v. State of Florida, No. 02-6106(09). [D.E. 282 in Martinez v. Bush, No. 02-20244-Civ-Jordan].

[5] On July 2, 2002, we entered an order concluding that the plaintiffs and intervenors had not met their respective burdens with regard to any of their claims concerning the congressional districts and the state senate districts. We explained that a detailed opinion containing findings of fact and conclusions of law would be issued as soon as possible. This is that opinion.

[6] We use "black candidate of choice" to mean the candidate who would be elected if the only persons eligible to vote in the election at issue were black. The term says nothing about the race of the candidate.

[7] Count I also alleged a violation of section 2 of the Voting Rights Act of 1964, 42 U.S.C. § 1973. We previously granted the defendants' motion for judgment on the pleadings under Federal Rule of Civil Procedure 12(b)(6) on the section 2 claim in count I.

[8] As previously noted, the cases before us here were not the only actions filed concerning redistricting in Florida. See supra note 4.

[9] Most, if not all, of the critical submissions and orders in the D.C. action have been filed in the Martinez action at bar (e.g., D.E. 186, 246, 269, 272, 289, 307).

[10] This was in essence the same claim found in count XV of the Martinez second amended complaint.

[11] In view of the three-judge court's inquiry, Speaker Feeney supplemented the record, by filing the pre-clearance letters from the Department of Justice establishing that the open primary amendment to the Florida Constitution was pre-cleared by letter dated February 26, 1999, and that the suspension of the runoffs was pre-cleared on August 17, 2001. See Notice of Filing [D.E. 280].

[12] We stated in that order that we would provide a more detailed rationale when we issued our final findings of fact and conclusions of law. Having reviewed that order in light of the record, however, we believe that no additional explanation is necessary.

[13] See infra section II.B.6.

[14] See, e.g., Chapman v. Meier, 420 U.S. 1, 27, 95 S.Ct. 751, 766, 42 L.Ed.2d 766 (1975).

[15] We note that neither the United States Constitution nor Florida law requires a specific process for public participation prior to the legislative enactment of redistricting plans. Any state-instituted process for redistricting is purely voluntary, as a state is not required to hold public hearings at all, let alone at a particular place or during certain hours. Of course, if a state creates such a process, it cannot be instituted or applied in a racially discriminatory way. As explained in the text, there was no such illegal conduct here.

[16] As previously noted, similar claims of vote dilution against certain Hispanic groups were raised by the plaintiffs but have since been dropped. See supra note 3.

[17] One of plaintiffs' experts accurately "retrodicted" —that is, predicted after the fact based on hindsight—the results of 113 of the 115 congressional general elections in Florida between 1992 and 2000, inclusive. The expert's "retrodiction" of the relative performance of the Democratic and Republican candidates in any given congressional election was the average vote cast for the Democratic and Republican candidates, respectively, in the particular congressional district, in the 23 contested statewide (non-congressional) elections during the same 1992-2000 time period, with nine percentage points added for incumbency, also calculated based on hindsight. If we knew the relative votes that would be cast for Democrats and Republicans in statewide elections between 2002 and 2010 and the proper incumbency advantage that would be calculated for that period after the fact, the same methodology could be used to predict outcomes in congressional elections in the coming decade. But those figures of course cannot be known in advance. The expert did not compare data from the 1980s to congressional election outcomes for the 1990s, but it is likely that the predictive value of the 1980s data would have proven very poor; Democrats handily won most statewide elections in the 1980s but lost most congressional elections in the 1990s.

[18] In 1995, the Northern District of Florida held that CD 3 constituted an unconstitutional racial gerrymander in violation of Shaw v. Reno, 509 U.S. 630, 113 S.Ct. 2816, 125 L.Ed.2d 511 (1993). See Johnson v. Mortham, 926 F.Supp. 1460 (N.D.Fla.1996). The legislature redrew CD 3 to make it more compact and regular in shape. The court upheld the redrawn CD 3. See Johnson v. Mortham, 1996 WL 297280 (N.D.Fla. May 31, 1996). The change took effect for the 1996 election. The change did not affect the other minority districts.

[19] Four of the five were expected to stand for reelection in 2002. After conclusion of the trial in the case at bar, Ms. Meek publicly announced her retirement, and her son, Kendrick Meek, who at the time was a member of the Florida Senate, announced his candidacy for her seat.

[20] Paragraphs 136, 137, 138, and 143 constitute our ultimate findings on the critical disputed issues of fact. Paragraphs 139 through 142 and 144 set forth ultimate findings that are supported by this record, have not been challenged in this action by any party, and are accepted by us on that basis.

[21] Information regarding the format of census self-reporting is based on census forms that are publicly available on the Census Bureau website. See United States Census Bureau, http://www.census.gov/dmd/www/2000quest.html (last modified March 27, 2002). We take judicial notice of this information, which is "capable of accurate and ready determination by resort to sources whose accuracy cannot reasonably be questioned." Fed. R. Ev. 201(b)(2). By motion to alter or amend, any party may request an opportunity to be heard as to the propriety of taking judicial notice or the tenor of the matter noticed. See Fed. R. Ev. 201(e).

[22] Hispanic respondents also were required to list their country of origin. As previously noted, unless otherwise indicated, "black" is used in this memorandum opinion to include all persons who, for purposes of the census, self identify as black (regardless of whether they also self identify as members of another race and regardless of whether they also self identify as Hispanic).

[23] If, as is likely, some persons were counted as black under the 2000 methodology who were not counted as black under the 1990 methodology, then the likely non-black vote for black candidates of choice calculated based on 1990 data will be overstated when applied to 2000 data. One thus would expect that, other things being equal, the percentage of non-blacks voting for black candidates of choice would be slightly lower than one would project using data based on the 1990 methodology. It is likely that the discrepancy would be very small, too small to affect our conclusions regarding any district. No party has asserted that this difference in census methodology makes a difference in the overall analysis, nor has any party proposed a method for adjusting the projections to account for the change in census methodology.

[24] According to the 2000 census, there were 124,663 underage blacks in CD 23 out of 179,980 total underage residents. See Maurer Exh. 1 at 79. Thus 69.3% of the underage residents were black. As they come of age, the percentage black voting age population in the district will increase, unless blacks leave the district faster than non-blacks. No evidence suggests there will be such a migration pattern, and, at least for some areas with high percentages of black residents in some cities, the experience has been exactly the opposite.

[25] Of course, these percentages do not guarantee that the democratic candidate—or the minority candidate of choice—will win a given election. As discussed in part III.D, infra, minority political parties are not always cohesive and majority parties do not always vote as a bloc to defeat minority parties. Neither the Equal Protection Clause nor the Voting Rights Act require such a guarantee.

[26] CD 23 was not affected by the 1996 amendments to the 1992 plan.

[27] The data set forth in paragraphs 148 through 154 primarily address total population and voting age population, not registered or actual voters. The percentage of black registered voters in both old and new CD 23 was in each instance slightly lower than the percentage of voting age population, but the percentage of blacks who actually voted was slightly higher. Thus, for example, in 1998 (the last non-presidential election year), non-Hispanic blacks constituted 59% of the registered Democrats in what has become new CD 23, but 64% of the actual voters in the Democratic primary, according to estimates based on ecological regression. See Pl. Exh. 81-C at 4, Table 1.

[28] Fla. Const. Article VI, Section 5(b) now provides: "If all candidates for an office have the same party affiliation and the winner will have no opposition in the general election, all qualified electors, regardless of party affiliation, may vote in the primary elections for that office."

[29] See Fla. Dep't of State, Div. of Elecs., Op. 00-06 (2000). Under Florida law, write-in candidates must qualify in advance. Fla. Stat. § 99.061(3) (2001). There is no fee for doing so. § 99.092(1), Fla. Stat. (2001). If no write-in candidate qualifies for an election, the ballot provides no block for writing in a candidate. See Fla. Stat. § 99.061(3) (2001). If a write-in candidate does qualify, the ballot allows a voter to write in a name, but the name is not listed on the ballot. Id. If an election is uncontested—that is, if there is only one candidate whose name would appear on the ballot and no write-in candidates have qualified—the race does not even appear on the ballot; the single candidate is deemed to have prevailed. See Fla. Stat. § 101.151(7) (2001).

[30] In at least some circumstances, state laws opening political primaries to voters who are not members of the party violate the First Amendment. See California Democratic Party v. Jones, 530 U.S. 567, 120 S.Ct. 2402, 147 L.Ed.2d 502 (2000) (invalidating California's "blanket primary" provision). There are substantial differences between the Florida and California systems. Whether those differences would save Florida's system from constitutional challenge is an issue not now before us.

[31] See 2001 Laws of Florida ch.2001-40 § 46.

[32] This conclusion does not rely on the assumption that someone associated with a black candidate of choice or even with the Democratic party would recruit a write-in candidate. It relies instead on the assumption that one person out of the hundreds of thousands who might have an incentive to do so would notice that the ability of CD 23 to perform for a black candidate of choice was in jeopardy and would choose to invest the 15 minutes (and no money) necessary to close the primary by signing up as a write-in candidate. Some would call this unseemly. But unseemly or not, the willingness of someone to qualify as a write-in candidate and thus close the primary seems virtually certain in any election exhibiting the kind of racial polarization and impact on a black candidate of choice that plaintiffs have posited. In fact, the number of write-in candidates for Florida congressional seats increased sharply after adoption of the open primary provision, from three in 1998 to 30 in 2000.

[33] See supra paragraph 160.

[34] The 1998 Democratic primary in Miami-Dade County had no races between blacks and whites. The top of the ticket was an uncompetitive attorney general election. An election generating greater turnout of course ordinarily would generate greater turnout of both blacks and non-blacks, so that the relative turnout of blacks and non-blacks might be the same. But it also might not, especially when compared to the situation of interest here: an election in which the ability of CD 23 to perform for a black candidate of choice is in jeopardy.

[35] Plaintiffs' analysis accepts racial polarization figures computed by defense experts based on data from a set of low-visibility elections. A significant portion of the data are from non-partisan judicial elections, even though the process of electing judges in Florida is "fundamentally different from the methods used to choose other elected state officials." Nipper v. Smith, 39 F.3d 1494, 1500 (11th Cir.1994). Using these data, the defense experts project—and plaintiffs' analysis accepts—that 73% of black voters will vote for the black candidate of choice, and 32% of non-black voters will vote for the black candidate of choice. Using such data to project congressional outcomes, especially in highvisibility, racially divisive elections, is hazardous. In Democratic primaries in CDs 3 and 23 involving both black and white candidates in 1992, the percentage of black voters who voted for black candidates was substantially higher, and the percentage of non-black voters who voted for black candidates was substantially lower. The net effect of these differences was better performance for black candidates than one would have projected based on the polarization data developed by the defense experts. The defense polarization data deal with all voters, not just Democrats, but they are nonetheless inconsistent with the voting behavior exhibited in the 1992 Democratic primaries in CDs 3 and 23.

[36] These election results are taken in part from the Florida Department of State website. See Florida Dep't of State, Div. of Elec., Elec. Reporting Sys., http://election.dos.state.fl. us/elections/resultsarchive/Index.asp (last visited November 8, 2002). We take judicial notice of the election results, which are "capable of accurate and ready determination by resort to sources whose accuracy cannot reasonably be questioned." Fed. R. Ev. 201(b)(2). By motion to alter or amend, any party may request an opportunity to be heard as to the propriety of taking judicial notice or the tenor of the matter noticed. See Fed. R. Ev. 201(e).

[37] That a district will perform for a particular black candidate of choice of course does not immunize the district from a dilution claim. But neither is that performance irrelevant. Dilution claims involve practical judgments about the real world. Our finding is that CD 23 would perform for a black candidate of choice in an election to fill an open seat, but our level of confidence in our ruling is increased by the likelihood that for some (and possibly all) of the elections at issue, there will be an incumbent black candidate of choice for whom the district's likely performance has not been questioned.

[38] The basis for this finding parallels that applicable to CD 23. See supra paragraph 149.

[39] We take judicial notice, see supra note 36, that Ms. Brown, the incumbent, has drawn no primary opposition at all since 1994. The possibility that in 2002 there would be three candidates in the Democratic primary—including a candidate in addition to Ms. Brown who has substantial support in the black community—was remote.

[40] The basis for this finding parallels that applicable to CD 23. See supra paragraph 149.

[41] CD 17 was not affected by the 1996 amendments to the 1992 plan.

[42] The eight counties included in this analysis are Broward, Collier, Hendry, Martin, Miami-Dade, Monroe, Palm Beach, and St. Lucie. Inclusion of eight counties rather than seven or nine or six or twelve is somewhat arbitrary but provides a reasonable approximation of the overall south Florida area within which black districts could be drawn.

[43] The 16 counties are Alachua, Baker, Bradford, Clay, Columbia, Duval, Flagler, Lake, Marion, Nassau, Orange, Putnam, St. Johns, Seminole, Union, and Volusia. Inclusion of 16 counties is somewhat arbitrary but provides a reasonable approximation of the overall northeast Florida area within which a black district could be drawn.

[44] Five counties in Florida are covered jurisdictions under section 5 of the Voting Rights Act of 1965, as amended in 1982. Hillsborough County is one of the five. Voting changes impacting these counties, including reapportionment plans, cannot take effect unless pre-cleared under section 5, either by the Department of Justice or by a three-judge United States District Court for the District of Columbia.

[45] After court approval, the Department of Justice pre-cleared the changes.

[46] The basis for this finding parallels that applicable to CD 23. See supra paragraph 149. In addition, there is a substantial population of black Haitian immigrants in HD 108. Many are not citizens and thus not eligible to register to vote. There is a substantial difference between black voting age population (55.2%) and black registered voters (44.8%). Whether this gap will close as recent immigrants become eligible to apply for citizenship is unclear.

[47] These election results are taken in part from the Florida Department of State website. See supra note 36.

[48] That the candidate who won was black does not necessarily mean he was the black candidate of choice. On this record, there is no reason to believe he was not.

[49] There is no evidence that the winner was not the black candidate of choice. See supra note 48.

[50] These election results are taken in part from the Florida Department of State website. See supra note 36.

[51] See Pl. Exh. 81-C at 38, Table 13. Using the same methodology, plaintiffs assert a black candidate of choice would draw only 45.5% of the vote in such a primary in prior HD 118. That the district elected a black candidate from among six Democrats in the 2000 primary, with only write-in opposition in the general election, supports the conclusion that this district is likely to perform for a black candidate of choice, and that the strength of this district in an open primary is unlikely to make a difference.

[52] The basis for this finding parallels that applicable to CD 23. See supra paragraph 149. In addition, the black total population in what is now new HD 104 was 45.3% according to the 1990 census. That grew to 60.0% under the 2000 census. The relative black population has trended rapidly upward.

[53] The black voting age population of HD 109, at 60.4%, is only 4.5 percentage points higher than the 55.9% black voting age population of HD 104. Plaintiffs assert HD 104 dilutes black votes but that HD 109 is packed. There is no requirement under the Constitution or under the Voting Rights Act that a state legislature create minority districts with the precision demanded by plaintiffs: less than a 4.5 point spread between impermissibly diluted and impermissibly packed.

[54] Section 2 of the Voting Rights Act, as amended, reads in its entirety:

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

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

42 U.S.C. § 1973 (emphasis in original).

[55] The Equal Protection Clause of the United States Constitution also prohibits dilution of minority votes, but, at least in respects relevant to the case at bar, is not more restrictive than the Voting Rights Act. See infra part III.C. Unlike section 2, the Equal Protection Clause is violated only when there is discriminatory intent. See City of Mobile v. Bolden, 446 U.S. 55, 100 S.Ct. 1490, 64 L.Ed.2d 47 (1980).

[56] In Gingles, the Supreme Court articulated the first factor in the section 2 analysis as whether the protected group is sufficiently large and geographically compact "to constitute a majority in a single-member district." Gingles, 478 U.S. at 50, 106 S.Ct. at 2766 (emphasis added). As discussed infra, we doubt this was intended as a literal, mathematical requirement. More recently, the Court has assumed, but not decided, that less than a majority will suffice. See, e.g., Voinovich v. Quilter, 507 U.S. 146, 158, 113 S.Ct. 1149, 1157-58, 122 L.Ed.2d 500 (1993). In making this assumption, the Court noted that the Gingles factors "cannot be applied mechanically and without regard to the nature of the claim." Id. Because the parties agree that racial minorities, and districts likely to perform for minority candidates, fall within the protections of the Voting Rights Act, and because we trace the development of these Gingles factors—as well as the totality of the circumstances test—in some detail as part of our discussion of political gerrymandering and vote dilution based on political party, we leave a more detailed discussion for that section. See infra part III.B.

[57] For further discussion on the totality of the circumstances required showing and its relationship to the Gingles factors, see infra part III.B.

[58] In their second amended complaint, the Martinez plaintiffs included political gerrymandering claims with regard to the state house and senate plans as well. They conceded in their second amended proposed findings of fact and conclusions of law, however, that no evidence had been offered to prove political gerrymandering of the state senate plan. Regarding their claim that the state house plan was gerrymandered, we reject their claim because they failed to satisfy the test we announce infra for political gerrymandering claims under the Equal Protection Clause of the Fourteenth Amendment.

[59] The Fourteenth Amendment provides in pertinent part, "No State shall ... deny to any person within its jurisdiction the equal protection of the laws." U.S. CONST. amend. XIV, § 1.

[60] Agreeing with Justice White were Justices Brennan, Marshall, and Blackmun. Although only representing a plurality, the test articulated by Justice White is binding precedent. See Marks v. United States, 430 U.S. 188, 193, 97 S.Ct. 990, 993, 51 L.Ed.2d 260 (1977) ("When 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.") (internal citation and punctuation omitted).

[61] The Bandemer plurality itself recognized that its test "may be difficult of application." Bandemer, 478 U.S. at 142, 106 S.Ct. at 2815 (White, J., plurality opinion). See also O'Lear v. Miller, 222 F.Supp.2d 850, 855-56 (E.D.Mich.2002) ("Bandemer's standard for demonstrating an `actual discriminatory effect' is somewhat murky....[and the] plurality never offered a checklist of features which together would comprise a systemic frustration of majority will."); Vieth v. Pennsylvania, 188 F.Supp.2d 532, 544 (M.D.Pa.2002) ("We begin our analysis [of the effects prong] by stating that the recondite standard enunciated in Bandemer offers little concrete guidance.").

[62] See, e.g., Bandemer, 478 U.S. at 127, 132, 106 S.Ct. at 2808, 2810 (citing City of Mobile v. Bolden, 446 U.S. 55, 100 S.Ct. 1490, 64 L.Ed.2d 47 (1980)); id. at 124, 128, 130, 131, 136-37, 106 S.Ct. at 2806, 2808, 2809-10, 2812 (citing Whitcomb v. Chavis, 403 U.S. 124, 91 S.Ct. 1858, 29 L.Ed.2d 363 (1971); id. at 124, 130, 131, 134, 137, 140 n. 17, 106 S.Ct. at 2806, 2809-10, 2811, 2812-13, 2814 n. 17) (citing White v. Regester, 412 U.S. 755, 93 S.Ct. 2332, 37 L.Ed.2d 314 (1973)); id. at 131, 106 S.Ct. at 2809-10 (citing Rogers v. Lodge, 458 U.S. 613, 102 S.Ct. 3272, 73 L.Ed.2d 1012 (1982)).

[63] Although the plaintiffs' and intervenors' equal protection claim is one of political gerrymandering, not vote dilution, the distinction is insignificant. The Fifth Circuit has explained:

We see no distinction that would call for different constitutional requisites in a racial gerrymander case than in a voting dilution case ... [because] [t]he right allegedly infringed is the same in both contexts: the right to effective participation in the electoral process. An unconstitutional gerrymander violates this right by compartmentalizing or fencing out a group or by slicing up a compact minority. An invidious at-large scheme [typical of voting dilution cases] merely achieves the same end, denial of effective participation, by submerging an interest group in a constituency large enough and polarized enough to place that group in the minority consistently.

Nevett v. Sides, 571 F.2d 209, 218-19 (5th Cir.1978) (internal citations omitted); see also Bonner v. City of Prichard, 661 F.2d 1206, 1209 (11th Cir.1981) (en banc) (adopting as binding precedent all decisions of the former Fifth Circuit handed down prior to October 1, 1981). Our reliance in subpart III.A, infra, on cases involving racial vote dilution is therefore proper.

[64] Anti-single shot voting provisions require that each elector cast votes for as many candidates as there are positions. See, e.g., Nevett, 571 F.2d at 217 n. 10. The rule applies "only in the context of an electoral scheme that selects winners by ranking all candidates in the order of the number of votes they receive," also known as the "single ballot-plurality voting system." Id. "An anti-single shot rule invalidates all ballots that do not show votes for as many candidates as there are positions," which can disadvantage minority voters by "forc[ing] them to vote for non-minority candidates, thus depreciating the relative position of minority candidates." Id.

[65] The Fifth Circuit later divided this list of factors into two categories: "primary" and "enhancing." Nevett v. Sides, 571 F.2d 209, 217 (5th Cir.1978). Those factors considered "primary" include the group's accessibility to the political process, the responsiveness of representatives to the needs of the minority group, the state policy behind the districting scheme, and the effect of past discrimination upon the group's participation in the electoral system. Id. The "enhancing" factors include "the size of the district; the portion of the vote necessary for election (majority or plurality); where the positions are not contested for individually, the number of candidates for which an elector must vote; and whether candidates must reside in subdistricts." Id. (internal footnotes omitted).

[66] The court also noted that bloc voting may, of course, be shown by "more direct means, such as statistical analyses or the consistent lack of success of qualified candidates." Nevett, 571 F.2d at 223 n. 18 (internal citations omitted).

[67] The court reiterated its statement from Nevett, 571 F.2d at 224, that the finding of one or even several of the Zimmer criteria does not necessarily establish the existence of intentional discrimination. See Bolden, 571 F.2d at 243. "Although the failure of black candidates because of polarized voting is not sufficient to invalidate a plan, it is an indication of lack of access to the political processes. It is one piece of the circumstantial evidence puzzle, whose successful completion supports the illation of dilution." Id. (internal citations omitted).

[68] The Justices concurring in Justice Stewart's opinion were the Chief Justice, Justice Powell, and Justice Rehnquist. In holding that the "so-called Zimmer criteria ... were most assuredly insufficient to prove an unconstitutionally discriminatory purpose," the plurality pointed out that the evidence also showed that "Negroes register and vote in Mobile `without hindrance,' and that there are no official obstacles in the way of Negroes who wish to become candidates for election to the Commission." Bolden, 446 U.S. at 73, 100 S.Ct. at 1503. Moreover, proof of a history of racial discrimination "cannot, in the manner of original sin, condemn governmental action that is not itself unlawful." Id. at 74, 100 S.Ct. at 1503.

[69] The full text of section 2 is set out, supra note 54.

After examining section 2 of the Voting Rights Act, the plurality concluded that, "[i]n view of the section's language and its sparse but clear legislative history, it is evident that this statutory provision adds nothing to the appellees' Fifteenth Amendment claim." Bolden, 446 U.S. at 61, 100 S.Ct. at 1496. The plurality then analyzed the Fifteenth Amendment and found that "that Amendment prohibits only purposefully discriminatory denial or abridgment by government of the freedom to vote `on account of race, color, or previous condition of servitude.'" Id. at 65, 100 S.Ct. at 1498.

[70] The four Justices were Justices White, Blackmun, Brennan, and Marshall. Both Justices White and Blackmun agreed that proof of purposeful discrimination should be an element in racial vote dilution cases and that the Zimmer criteria were relevant in that determination, though Justice Blackmun concurred in the Stewart plurality's result based on his belief that "the relief afforded appellees by the District Court was not commensurate with the sound exercise of judicial discretion." Bolden, 446 U.S. at 80, 100 S.Ct. at 1507 (Blackmun, J., concurring). Justice Marshall wrote a separate dissenting opinion and maintained that proof of discriminatory intent is not required in racial voting dilution claims, id. at 104-105, 100 S.Ct. at 1490 (Marshall, J., dissenting), but, if such proof were necessary, the evidence in the instant case should satisfy that requirement, id. at 137, 100 S.Ct. at 1519. Justice Brennan wrote separately to voice his agreement with Justice Marshall. Id. at 103, 100 S.Ct. at 1518 (Brennan, J., dissenting). The ninth and final Justice, Justice Stevens, altogether rejected the intentional discrimination standard enunciated by the Stewart plurality, id. at 90-91, 100 S.Ct. at 1512 (Stevens, J., concurring), but concurred in its result because the proof offered in the court below failed to satisfy the legal standard he believed was applicable, id. at 92-94, 100 S.Ct. at 1513-14.

[71] In the face of the plurality's holding that there was no vote dilution because blacks can register to vote and slate candidates without hindrance, Justice White noted that "[t]he absence of official obstacles to registration, voting, and running for office heretofore has never been deemed to insulate an electoral system from attack under the Fourteenth and Fifteenth Amendments." Bolden, 446 U.S. at 102, 100 S.Ct. at 1518 (White, J., dissenting).

[72] While the Court did not explicitly overrule the Bolden plurality, Justice Powell, in his dissenting opinion in Rogers, stated, "Whatever the wisdom of [Bolden], the Court's opinion [today] cannot be reconciled persuasively with that case." Rogers, 458 U.S. at 629, 102 S.Ct. at 3281 (Powell, J., dissenting).

[73] The full text of section 2 is set out, supra note 54.

Section 2 had, for the most part, lain dormant since its inception. See Bolden, 446 U.S. at 60-61, 100 S.Ct. at 1496 (stating that appellees' section 2 claim "adds nothing to [their] complaint" and holding that section 2's legislative history "makes clear that it was intended to have an effect no different from that of the Fifteenth Amendment itself"). Most vote dilution plaintiffs, therefore, brought their claims under the Fourteenth and Fifteenth Amendments, and these constitutional provisions provided the basis on which most courts rested their holdings. See, e.g., id. The Congressional amendment to section 2, however, intended to change that practice. See Gingles, 478 U.S. at 42-46, 106 S.Ct. at 2762-64 (discussing the changes the amendment to section 2 was intended to accomplish).

[74] Factors a court may consider in evaluating a racial vote dilution claim were set forth by the Senate in its report accompanying the 1982 amendments to section 2 of the Voting Rights Act. Those factors include:

1. the extent of any history of official discrimination in the state or political subdivision that touched the right of the members of the minority group to register, to vote, or otherwise to participate in the democratic process;

2. the extent to which voting in the elections of the state or political subdivision is racially polarized;

3. the extent to which the state or political subdivision has used unusually large election districts, majority vote requirements, anti-single shot provisions, or other voting practices or procedures that may enhance the opportunity for discrimination against the minority group;

4. if there is a candidate slating process, whether the members of the minority group have been denied access to that process;

5. the extent to which members of the minority group in the state or political subdivision bear the effects of discrimination in such areas as education, employment and health, which hinder their ability to participate effectively in the political process;

6. whether political campaigns have been characterized by overt or subtle racial appeals; and

7. the extent to which members of the minority group have been elected to public office in the jurisdiction.

Additional factors that [may be probative of vote dilution in some cases] are:

[8.] whether there is a significant lack of responsiveness on the part of elected officials to the particularized needs of the members of the minority group; and

[9.] whether the policy underlying the state or political subdivision's use of such voting qualification, prerequisite to voting, or standard, practice or procedure is tenuous.

S. REP. 97-417, at 28-29 (1982), reprinted in 1982 U.S.C.C.A.N. 177, 206-07 (internal footnotes omitted).

[75] Subsequently, in Growe v. Emison, the Court held that the standards developed in Gingles for testing section 2 challenges to multimember districts also controlled vote dilution challenges to single-member districts. 507 U.S. 25, 40-41, 113 S.Ct. 1075, 1084, 122 L.Ed.2d 388 (1993) (quoting Gingles, 478 U.S. at 50-51, 106 S.Ct. at 2766-67).

[76] Justice Brennan authored the majority opinion, in which Justices White, Marshall, Blackmun, and Stevens joined. Justice White, however, did not join the portion of that opinion which, according to Justice White, "states ... that the crucial factor in identifying polarized voting is the race of the voter and that the race of the candidate is irrelevant." Gingles, 478 U.S. at 83, 106 S.Ct. at 2783 (White, J., concurring). Justice White had doubts that "this is what Congress had in mind in amending [section] 2 as it did [and stated that] it seems quite at odds with the discussion in Whitcomb v. Chavis, 403 U.S. 124, 149-60, 91 S.Ct. 1858, 1872-78, 29 L.Ed.2d 363 (1971)." Gingles, 478 U.S. at 83, 106 S.Ct. at 2783 (White, J., concurring).

[77] When considered together, factors two and three constitute politically polarized voting.

[78] In DeSoto, the Eleventh Circuit stressed the importance of the causation element in its vote dilution analysis. Even assuming the plaintiffs had established the remaining two elements (intent and a denial of the opportunity to participate in the political process), the court held that plaintiffs claim ultimately failed because they had not proven causation. DeSoto, 204 F.3d at 1345. Specifically, plaintiffs had failed to show that any alternative scheme could be drawn wherein blacks could elect their candidate of choice. Id. at 1346. In other words, plaintiffs had failed to establish that they were "sufficiently large and geographically compact to constitute a majority in a single-member district." Gingles, 478 U.S. at 50, 106 S.Ct. at 2766.

[79] The district court in Bandemer recognized the need to clarify the subgroup of Democrats who comprised an identifiable group, although it disagreed on how to calculate its numbers. See Bandemer, 478 U.S. at 135 n. 5, 106 S.Ct. at 2811 n. 5. The majority would have counted as Democrats, for example, all those voters—however registered—who had voted for a Democratic Assembly candidate in the most recent election. Id. The dissenting judge, however, would only have counted those Democrats who consistently voted a straight Democratic ticket. Id. The Supreme Court declined to resolve this issue in Bandemer. Id. We suggest that the Court in Bandemer did not have to resolve this issue because the plaintiffs there had alleged no more discriminatory effects than disproportionate election results—effects which the Court had held time and time again to be insufficient to sustain an equal protection challenge.

[80] As in racial vote dilution cases, the number of elections required to establish polarized voting must be made on a case-by-case basis.

[81] See list, supra note 74.

[82] In recognizing that political parties may have to rely on proof of lack of responsiveness alone to establish a diminished opportunity to participate in the political process, the Court in Bandemer implicitly recognized that political parties would face greater obstacles to prevailing on vote dilution claims than would racial minorities.

[83] In their proposed findings of fact and conclusions of law, the Deutsch plaintiff-intervenors also argue that the redistricting scheme violates the equal protection rights of two other subgroups of voters: residents of Pembroke Pines and Collier County. The intervenors fail to explain how these groups are "identifiable political groups" at all; there is no evidence of their voting registration or patterns, nor any evidence of whether or why Republican legislators would want to discriminate against them.

[84] The evidence submitted by the Deutsch intervenors indicates that there are roughly 3,872,000 registered Democrats and 3,501,000 registered Republicans in the state of Florida. Based on voter registration in 2000, the distribution of registered voters throughout the newly-enacted plan shows the following:

(1) District 1: 174,676 Republicans; 165,967 Democrats.

(2) District 2: 104,874 Republicans; 245,969 Democrats.

(3) District 3: 67,915 Republicans; 191,061 Democrats.

(4) District 4: 139,205 Republicans; 171,634 Democrats.

(5) District 5: 171,302 Republicans; 167,760 Democrats.

(6) District 6: 150,506 Republicans; 158,585 Democrats.

(7) District 7:165,532 Republicans; 151,186 Democrats.

(8) District 8: 142,505 Republicans; 121,405 Democrats.

(9) District 9: 172,378 Republicans; 143,481 Democrats.

(10) District 10: 173,951 Republicans; 145,258 Democrats.

(11) District 11: 85,331 Republicans; 159,525 Democrats.

(12) District 12: 133,064 Republicans; 153,202 Democrats.

(13) District 13: 205,010 Republicans; 139,948 Democrats.

(14) District 14: 191,914 Republicans; 107,767 Democrats.

(15) District 15: 170,222 Republicans; 138, 858 Democrats.

(16) District 16: 182,295 Republicans; 148,688 Democrats.

(17) District 17: 38,713 Republicans; 189,576 Democrats.

(18) District 18: 126,814 Republicans; 96,928 Democrats.

(19) District 19: 111,432 Republicans; 202,878 Democrats.

(20) District 20: 107,640 Republicans; 182,305 Democrats.

(21) District 21: 119,450 Republicans; 85,769 Democrats.

(22) District 22: 180,057 Republicans; 141,277 Democrats.

(23) District 23: 54,894 Republicans; 183,043 Democrats.

(24) District 24: 156,292 Republicans; 126,976 Democrats.

(25) District 25: 100,847 Republicans; 81,809 Democrats.

[85] The official results from the Florida Department of State, Division of Elections, see supra note 36, demonstrate the following:

(1) CD 1 (as drawn in the congressional map effective in 1996): in 1996, 46.5% of the registered voters were Democrat, and 43.6% were Republican. In the congressional general election that year, however, Republican candidate Joe Scarborough won 72.5% of the vote; whereas, the Democratic candidate, Kevin Beck, garnered only 27.4%. By 1998, Republicans had gained a slight plurality of registered voters (Republican: 44.2%; Democrat: 43.9%), but in the general election for Congress, the Republican candidate won by an overwhelming majority—99.5%. Indeed, there was not even a Democratic candidate. An almost identical scenario occurred in the 2000 general election.

(2) CD 6 (as drawn in the congressional map effective in 1996): in 1996, 45.1% of voters registered Democrat; 43.9% registered Republican. The Republican candidate for Congress that year, however, won 67.2% of the vote in the general election held in November; whereas, the Democratic representative won only 32.8%. Although registered Democrats had a slight plurality in 1998 (by 0.1%), the Republican candidate for Congress that year ran unopposed. By the year 2000, Republican registered voters had edged out Democrats by 3.5% (Republican: 44%; Democrat: 41.5%), but in the congressional election that year, the Democratic Party again did not even field a candidate. The Republican candidate won 99.9% of the vote.

(3) CD 10 (as drawn in the congressional map effective in 1996): in the years 1996 and 1998, Republican registered voters outnumbered Democrats by a small margin (1996: by 1.4%; 1998: by 0.1%); yet, in both years, the Republican candidate for Congress enjoyed a safe win, winning 66.6% of the vote in 1996 and 100% in 1998. By the year 2000, Democrats had gained a lead over Republicans in registered voters; 40.3% were registered Democrat and 40.0% were registered Republican. That same year, however, the Republican candidate for congress won the general election with 75.7% of the vote. The Democrats once again did not even field a candidate.

(4) CD 12 (as drawn in the congressional map effective in 1996): in 1996, a majority of the registered voters were Democrat (51.5%); only 38.6% were Republican. In the general election for Congress that year, however, the Republican candidate won 61.6% of the vote, and the Democratic candidate won only 38.4%. In 1998, registered Democrats still outnumbered Republicans (49.3% to 38.5%), but, because the Democratic Party did not field a candidate, the Republican candidate ran unopposed. In 2000, 46.8% of the registered voters were Democrat, and 38.3% were Republican. The Republican candidate for Congress won that year's general election, however, with 57.0% of the vote; the Democratic candidate garnered only 43.0%.

(5) CD 22 (as drawn in the congressional map effective in 1996): in 1996, 44.1% of registered voters were Democrat, and 41.1% were Republican. That year, however, in the general election for Congress, the Republican candidate won with 61.9% of the vote; the Democratic candidate received only 38.1%. By 2000, the margin between Republican and Democratic registered voters had widened— 42.2% were Democrats and 37.9% were Republican. In the general election, however, the Republican candidate still prevailed, barely winning with 50.1% of the vote.

[86] Of course, registered Democrats often vote for Republicans as a result of non-political factors—like incumbency, a factor which every expert testified holds heavy sway over voters. This truth, however, further proves our point: registered Democrats do not always vote Democratic.

[87] Indeed, in closing arguments, counsel for Congressman Peter Deutsch, plaintiff-intervenor, admitted as much.

[88] The actual contests were (1) in 2000: U.S. President, U.S. Senator, State Treasurer, and State Commissioner of Education; (2) in 1998: U.S. Senator, State Governor, Secretary of State, State Attorney General, State Comptroller, State Treasurer, State Commissioner of Education, and State Commissioner of Agriculture; (3) in 1996: U.S. President; (4) in 1994: U.S. Senator, State Governor, Secretary of State, State Attorney General, State Comptroller, State Treasurer, State Commissioner of Education, and State Commissioner of Agriculture; (5) in 1992: U.S. President and U.S. Senator.

[89] One might argue that the plaintiffs and intervenors submitted the voting index as proof that registered Democrats are politically cohesive. The voting index, however, does not identify voters by registration; it only looks at actual performance. Without such a correlation, the index is worthless in demonstrating how registered Democrats vote. In newly-enacted CD 4, for example, 47.7% of the registered voters are Democrat, and 38.7% are Republican. The average vote share of Democrats, however, is only 39.8%; whereas, the Republican average vote share is 60.2%.

[90] In CD 22, the average Democratic vote share is 56.3%, and the Republican share is 43.7%. The average Democratic vote share in CD 10 is 51.8%; whereas, the Republican vote share is only 48.2%.

[91] The official results of the elections from the Florida Department of State Division of Elections, see supra note 36, reveal the following:

(1) CD 10 (as drawn in the congressional map effective in 1996): in 1992, Bill Young, the Republican candidate, won the general election with 56.6%; the Democratic candidate received only 43.4%. Republican Young won every subsequent election, running unopposed in 1994, winning in 1996 by a margin of 33.2%, again running unopposed in 1998, and winning in 2000 with 75.7% of the vote.

(2) CD 22 (as drawn in the congressional map effective in 1996): in 1992, Republican E. Clay Shaw won the election with 52.0% of the vote, beating Democrat Gwen Margolis (who received only 37.1%) by 14.9%. Representative Shaw continued to win throughout the decade, succeeding in the 1994 general election with 63.4% of the vote, the 1996 election with 61.9%, the 1998 election with 100%, and the 2000 election with 50.1%.

[92] In CD 1, for example, the average Republican vote share is 62.9%, but in 1992, a Democrat won 52% of the vote in the congressional election. In CD 15, however, the average Republican vote share is 56.7%, but in 1992, a Democrat won the congressional election with 51% of the vote.

[93] The Democrats are also geographically compact in congressional districts 3, 17, and 23, the majority-minority districts judicially created to satisfy the requirements of section 2 of the Voting Rights Act. See DeGrandy v. Wetherell, 794 F.Supp. 1076, 1087-88, 1090 (N.D.Fla.1992). The plaintiffs and intervenors do not allege, obviously, that these districts were politically gerrymandered in violation of the Equal Protection Clause.

[94] The Deutsch intervenors discussed an "alternative plan," one that had been proposed in the Florida Senate by Senator Campbell (S33C0009). That plan yielded twelve seats with Democratic vote index majorities and thirteen seats with Republican vote index majorities. However, this plan—and others like it—are insufficient in proving geographic compactness because they do not demonstrate that the "Democrats" in those plans are politically cohesive, nor that the "Republicans" bloc vote usually to defeat Democratic candidates. In other words, alternative plans are helpful in establishing geographic compactness only if there is already proof that the groups highlighted in the alternative plan are identifiable, i.e., that they are politically cohesive.

[95] The main policy behind most legislatively-drawn redistricting will likely be incumbency protection. Even if the plaintiffs have established that this was the guiding principle in the instant redistricting, however, this would not weigh significantly in favor of finding the resulting plan an unconstitutional gerrymander. Rather, the Supreme Court has repeatedly recognized incumbency protection as a legitimate state goal. See, e.g., Hunt v. Cromartie, 526 U.S. 541, 549-51, 119 S.Ct. 1545, 1550-51, 143 L.Ed.2d 731 (1999); Karcher v. Daggett, 462 U.S. 725, 740, 103 S.Ct. 2653, 2663, 77 L.Ed.2d 133 (1983).

[96] Indeed, the evidence proved the opposite: Congressman Deutsch, a Democrat, testified that he has, over the course of his tenure as a representative, consistently represented all of his constituents, without regard to political affiliation. This, he explained, is what every representative seeks to do.

Plaintiff Maurer makes the argument that we discredited in subpart III.C.2. In particular, he argues that a Republican-controlled Congress will be unresponsive to his political views, which, in his words, parallel those of a "bleeding heart liberal." He contends that he "and millions of others will lose on the following illustrative issues if there be [sic] a Republican U.S. House of Representatives: (a) an impeached Democratic President; (b) No chance for an Employment Non-Discrimination Act (ENDA); (c) No chance for an attempted ban against religious discrimination by federally-funded religious service providers; (d) No chance for domestic partner benefits for District of Columbia employees; (e) No chance to join international treaties on control of chemical and biological weapons ...." Maurer's argument, however, is tantamount to holding that "it is invidiously discriminatory for a [political subdivision] to elect its delegation by a majority vote based on party or candidate platforms and so to some extent predetermine legislative votes on particular issues." Whitcomb, 403 U.S. at 155, 91 S.Ct. at 1875. That the majority may reject the views of the minority is "inherent in government by elected representatives." Id.

[97] This is the number of registered voters in 2000, as reported by the Division of Elections, Florida Department of State. See supra note 36.

[98] In the Florida House in 1992, there were 74 Democrats and 46 Republicans. The flip-flop began slowly: in 1994, there were 71 Democrats and 49 Republicans. By 1996, Republicans had a majority with 61 representatives, leaving 59 Democrats. In 1998, the gap was wider with 66 Republicans and 54 Democrats. The total in 2000 was 77 Republicans and 43 Democrats.

The Florida Senate had 21 Democrats and 19 Republicans in 1992. In 1994, the Republicans edged ahead with 22 senators, leaving 18 Democrats. The Republicans gained one more seat in 1996, and by 1998, had 25 seats with only 15 left for Democrats. In 2000, the division was the same: 25 Republicans and only 15 Democrats.

[99] Plaintiff-intervenor, Congressman Deutsch, makes three other arguments. First, he asserts that this case is distinguishable from Bandemer because Bandemer "presented a claim that Indiana's exercise of its reserved power to apportion its own legislature violated the Equal Protection Clause"; whereas, "[t]his case presents a claim that Florida's exercise of its delegated power under Article I to [redistrict] violates the Equal Protection Clause." Deutsch Mem. in Opp'n to Defs.' Mots. for J. on the Pleadings [D.E. 134 at 3] (emphasis in original) [hereinafter Deutsch Mem.]. For this proposition, Deutsch cites U.S. Term Limits v. Thornton, 514 U.S. 779, 115 S.Ct. 1842, 131 L.Ed.2d 881 (1995), which, he claims, confirms "that a more stringent standard must be applied in judging a state's exercise of its delegated powers in adopting a congressional redistricting plan than the standard that would apply in a case such as Bandemer that involved a challenge to a state exercising its reserved powers in redistricting its own legislature." Deutsch Mem. at 4 (emphasis in original). We need not address this argument fully because, even under a lower standard of proving discriminatory effects, the plaintiffs' and intervenors' claim would fail. That less exacting standard would, at a minimum, require more than mere disproportional representation. See, e.g., Whitcomb, 403 U.S. at 153, 91 S.Ct. at 1874; White, 412 U.S. at 765-66, 93 S.Ct. at 2339. Disproportional representation is, however, all that the plaintiffs and intervenors have shown here.

Deutsch's second argument relies on Article I, section 2, of the Constitution, which, Deutsch asserts, "limit[s] the power of state legislatures to put in place barriers that would interfere with the power of the people to choose their representatives." Deutsch Mem. at 2. Indeed, the relevant portion of Article I, section 2, states that "Members [of the House of Representatives] [shall be] chosen ... by the People of the several States." U.S. CONST. art. I, § 2. From this language, Deutsch claims that Article I, section 2 itself is violated because the Florida legislature has gerrymandered so pervasively that it has "limit[ed] the choice of the people of Florida by creating at least seventeen districts in which a Republican will be chosen, and not more than eight districts in which a Democrat will be chosen." Deutsch Mem. at 2. Deutsch, however, has failed to prove the assumption on which this Article I, section 2, claim is made—that the partisan victories in each district are absolute. See supra part III.D.2.a. Moreover, Deutsch has pointed us to no case in which political gerrymandering satisfied the Equal Protection Clause but violated Article I, section 2.

Finally, Deutsch argues that the predominant use of political affiliation is, in and of itself, a violation of the Equal Protection Clause. Deutsch attempts to find precedential support, citing Shaw v. Reno, 509 U.S. 630, 113 S.Ct. 2816, 125 L.Ed.2d 511 (1993), in which the Supreme Court held that using race as the predominant redistricting factor is impermissible. Deutsch's argument is foreclosed, however, by the overwhelming precedent acknowledging that race and ethnicity are special categories that require heightened scrutiny from the federal courts. See, e.g., Cromartie, 526 U.S. at 546, 119 S.Ct. at 1548-49; Bush v. Vera, 517 U.S. 952, 964, 116 S.Ct. 1941, 1954, 135 L.Ed.2d 248 (1996). Moreover, the Court has upheld redistricting schemes challenged under Shaw because the Court found that the legislature was motivated not by race but by political affiliation. See, e.g., Cromartie, 526 U.S. at 551, 119 S.Ct. at 1551.

[1] Under the two-tiered equal protection approach embraced in a series of Supreme Court decisions, the right to vote and to support the party of one's choice could be classified as a fundamental right, requiring lines drawn on this basis to undergo strict scrutiny. Later cases have placed diminished reliance on the two-tiered approach, see generally Gerald Gunther, In Search of Evolving Doctrine on a Changing Court: A Model for a Newer Equal Protection, 86 Harv. L.Rev. 1 (1972), and it is clear, in any event, that strict scrutiny does not apply to the drawing of districts for partisan political reasons. See, e.g., Easley v. Cromartie, supra. This underscores that drawing district lines is different from other uses of political affiliation as a basis for government action.

[2] To be sure, there is a difference between racial vote dilution and political gerrymandering. In racial vote dilution cases, the class at issue is defined by a characteristic (race) that exists separate and apart from voting behavior. The Gingles factors are designed in part to identify the circumstances under which that characteristic is politically significant, that is, to bridge the gap between a nonpolitical characteristic (race) and politics (the opportunity to cast a meaningful vote). Thus, for example, when blacks are "politically cohesive," their placement within the various districts affects their opportunity to have a meaningful impact on the political process. In political gerrymandering cases, in contrast, the class is defined by a characteristic (voting behavior) that is part and parcel of the political process. If the affected class is defined as individuals who vote for Democratic candidates, there is no gap to be bridged between that behavior (voting) and its impact on politics. For this reason, the subsidiary Gingles factors may apply in differing degrees in political gerrymandering cases as contrasted with racial vote dilution cases, but the ultimate issue is the same: whether a group of voters has been denied a meaningful opportunity to affect the political process.

[3] This is some indication that the lines have been drawn to favor Republicans, but this is not as conclusive as might appear at first blush. Other things being equal, Democratic voters are often more concentrated geographically, while Republican voters are often more dispersed. Thus, for example, inner-city districts often have an overwhelming majority of Democratic voters; this "wastes" Democratic votes not because of political gerrymandering, but simply as the result of demographics. It thus would not be surprising to find that, with an even split of voters statewide, more than half of the districts would still have a majority of Republican voters, without regard to political gerrymandering. On a national scale, the same concept was illustrated by the 2000 presidential election, where Mr. Gore received more popular votes nationwide than Mr. Bush, but Mr. Bush carried 30 states. This record does not indicate what the split of congressional districts in Florida would have been had the lines been drawn without regard to partisan politics, but any such non-partisan plan clearly would have been better for Democrats than the plan as adopted.