Rodriguez v. Pataki

308 F. Supp. 2d 346, 2004 U.S. Dist. LEXIS 4010, 2004 WL 503748

• Court: District Court, S.D. New York
• Decided: March 15, 2004
• Docket: 02 CIV. 618(RMB), 02 CIV. 3239(RMB)
• Status: Published

Opinion

OPINION AND ORDER

PER CURIAM.

TABLE OF CONTENTS

I. INTRODUCTION.351

II. BACKGROUND.354

III. ONE-PERSON, ONE-VOTE CLAIM (COUNT I) CO ZD CO

A. Legal Standards. CO ZD CO

B. Analysis . hD <D CO

IV. VOTING RIGHTS ACT CLAIMS. t-H tr-CO

A. Section 2 Legal Standard. t — I t-CO

B. Long Island: Nassau County and Suffolk County (Counts V and VI) CO t> CO

1. Suffolk County (Count VI). C"“ t— CO

2. Nassau County (Count V). r-H co CO

(a) The First Gingles Factor: Majority-in-a-District and the Potential to Elect. CO CO to

(b) Racial Polarization and the Second and Third Gingles Preconditions. CO 00

i) The Methodology of Dr. McDonald. CO 00 -a

ii) Elections Analyzed . 00 CO oo

iii) The Second Gingles Precondition: Cohesion. CO ZD to

iv) The Third Gingles Precondition: White Bloc Voting Sufficient Usually to Defeat the Black Candidate of Choice CO CO to

(e) Electability. CO CO rf*.

i) Recompilation. CO CD cn

ii) Supplemental Electability Analysis. CO ZD

iii) Turnout, Warming, and Other Statistical and Anecdotal Evidence.400

C. The Bronx (Count III). 404

1. Introduction. 404

2. Background to Plaintiffs’ Proposed District 36. 406

3. The First Gingles Precondition. 408

(a) The Appropriate Measure. 409

(b) Electability. 410

i) Recompilation. 410

ii) Supplemental Electability Analysis. 412

4. The Second Gingles Precondition. 413

5. The Third Gingles Precondition. 422

6. Totality of Circumstances. 426

(a) Substantial Proportionality. 427

(b) Legislative Policies Underlying the Contested Practices 429

(c) The Degree of Racially Polarized Voting and Political Partisanship. CO CO

(d) Official Discrimination. CO

(e) Socioeconomic Disparities. CO ÜT

(f) Electoral Mechanisms. CO 05

(g) Other Totality Factors CO Oi

7. Conclusion . 437

D. The Statewide Vote-Dilution Claim (Count II).437

E. The LVRC Intervenors: Senate District 31 (Count IV).437

F. The Rivers/Barbour Intervenors: Congressional District 17 (Count VIII).441

V. EQUAL PROTECTION CLAIM: SENATE DISTRICT 34 (COUNT VII).444

A. Applicable Law.445

B. Analysis .446

1. Senate District 34 .447

2. 2002 Redistricting.448

3. The Plaintiffs’ Case.449

(a) Shape and Boundary Lines.449

(b) Information Available to the Legislators in 2002.452

(c) Wolpoff Case Submissions.454

(d) Burgeson 2001 Memoranda.456

4. Traditional Districting Principles.457

5. Politics/Race: Cromartie II.459

VI. CONCLUSION, .460

I. INTRODUCTION

These consolidated actions include constitutional and statutory challenges to the State Senate and congressional redistriet-ing plans enacted by the New York State Legislature in April 2002 (following the 2000 census) and precleared by the United States Department of Justice in June 2002. Seven of the eight counts in the Joint and Consolidated Amended Complaint, dated January 24, 2003 (“Complaint”), pertain to the 2002 New York State Senate redistricting plan (“Senate Plan”), and one count pertains to New York’s 2002 congressional redistricting plan (“Congressional Plan”). Two of the eight counts raise constitutional challenges under the Fourteenth Amendment, U.S. Const. amend. XIV, § 1: a one-person, one-vote challenge to the Senate Plan as a whole and a racial gerrymandering challenge to Senate District (“SD”) 34. Six counts raise challenges under section 2 of the Voting Rights Act of 1965, 79 Stat. 437, as amended, 42 U.S.C. § 1973(b) (the “VRA”), including challenges by the lead plaintiffs to redistricting in the Bronx, Long Island, and the state as a whole, as well as challenges by plaintiffs-intervenors to SD 31 and Congressional District (“CD”) 17.

On November 6, 2003, this three-judge District Court concluded that the plaintiffs had raised no triable issues of material fact with respect to Counts I, II,.IV,, VI, and VIII of the Complaint and granted summary judgment to the defendants on those claims. We indicated that an opinion explaining the decision would follow. Following trial, the Court has concluded that the plaintiffs have failed also to establish the claims set forth in Counts III, V, and VII by a preponderance of the evidence. Our opinion with'respect to those counts on which we granted summary judgment together with our findings of fact and conclusions of law with respect to all counts are detailed below, but some of the overarching considerations that inform our decision are as follows:

First, New York’s 2002 redistricting laws are well within the purview and political prerogative of the State Legislature. See, e.g., Miller v. Johnson, 515 U.S. 900, 915, 115 S.Ct. 2475, 132 L.Ed.2d 762 (1995) (“Federal-court review of districting legislation represents a serious intrusion on the most vital of local functions.... Electoral districting is a most difficult subject for legislatures, and so the States must have discretion to exercise the political judgment necessary to balance competing interests.”); see also Georgia v. Ashcroft, 539 U.S. 461, 123 S.Ct. 2498, 2511-12, 156 L.Ed.2d 428 (2003); Md. Comm. for Fair Representation v. Tawes, 377 U.S. 656, 676, 84 S.Ct. 1429, 12 L.Ed.2d 595 (1964).

Second, the 2002 Senate Plan reflects traditional districting principles including: maintaining equality of population, preserving the “cores” of existing districts, preventing contests between incumbents, and complying with the requirements of the Voting Rights Act, See Marylanders for Fair Representation, Inc. v. Schaefer, 849 F.Supp. 1022, 1056 (D.Md.1994) (three-judge court); see generally Larios v. Cox, 300 F.Supp.2d 1320 (N.D.Ga.2004) (three-judge court).

Third, the 2002 redistricting continues New York’s check and balance in its bicameral Legislature. In the State Assembly, which has been dominated by Democrats since 1974, six seats were gained by Democrats in 2002 and the balance of Democratic to Republican assemblyper-sons changed from 97/53 to 103/47. In the State Senate, which has been dominated by Republicans since 1966, the balance of Republican to Democratic senators shifted from 36/25 to 38/24, including the post-election change in party affiliation (from Democratic to Republican) of one senator.

Fourth, the Senate Plan reflects less than a 10% deviation in population between any two Senate districts. See Brown v. Thomson, 462 U.S. 835, 842, 103 S.Ct. 2690, 77 L.Ed.2d 214 (1983) (“[A]n apportionment plan with a maximum population deviation under 10% falls within [the] category of minor deviations.”). The plaintiffs do not assert that racial discrimination accounts for the population deviation. (See Transcript of .Oral Argument of Motions for Summary Judgment, dated Nov. 4, 2003, at 65 (“Summ. J. Tr.”).) Rather, they argue that the Senate Plan discriminates geographically, favoring underpopulated “upstate” districts over overpopulated “downstate” districts, and that the single Senate seat added by the Legislature in 2002 should have been located south of Putnam County. 1

Fifth, politics surely played a role in redistricting in New York in 2002 — -as it does in most every jurisdiction. See, e.g., Gaffney v. Cummings, 412 U.S. 735, 753, 93 S.Ct. 2321, 37 L.Ed.2d 298 (1973) (“Politics and political considerations are inseparable from districting and apportionment. ... The reality is that districting inevitably has and is intended to have substantial political consequences.”). For example, the Republicans’ 62-seat Senate Plan, which became the plan that was enacted, paired four incumbent Democrats to run against each other. The plan initially proposed by the Democrats for a 61-seat Senate paired twenty incumbents, seventeen of whom were Republican. At the same time, the Senate and Assembly joined political forces in developing the State’s 2002 Congressional Plan (which was required by the census returns to reflect the loss of two seats from New York’s congressional delegation) to pah-two Democrats and two Republicans, resulting in the loss of one member from each party.

Sixth, the Senate Plan achieves substantial proportionality of minority representation in the Bronx. “[D]ilution may be found to be absent under the totality of the circumstances when the protected minority groups ‘constitute effective voting majorities in a number of districts ... substantially proportional to their share in the population.’ ” Session v. Perry, 298 F.Supp.2d 451, 477 (E.D.Tex.2004) (quoting Johnson v. De Grandy, 512 U.S. 997, 1024, 114 S.Ct. 2647, 129 L.Ed.2d 775 (1994)). Four of the five Bronx Senate districts are majority-minority districts, including three Hispanic districts and one black district. Hispanics and blacks combined are 76.1% of the Bronx citizen voting age population (“CVAP”) and 77.5% of the voting age population (“VAP”). 2

Seventh, the relief sought by the plaintiffs in this case cannot be justified under the Voting Rights Act. For example, in Suffolk County, the plaintiffs propose to create a new minority-coalition “influence” district, even though blacks and Hispanics would constitute only 40.2% of the VAP and 33.7% of the CVAP in the proposed district and the plaintiffs concede that the proposed district would not provide minor ities with the ability to elect candidates of choice. 3 In Nassau County, the plaintiffs propose to create a new “ability to elect” district by gathering substantially all of the county’s black population in one district. Nonetheless, in the proposed Nassau County district, blacks would constitute less than 37% of the VAP and CVAP and, we conclude, would still not have the ability to elect candidates of their choice. The plaintiffs propose to create a fifth majority-minority district in the Bronx by reducing the minority populations in all four majority-minority districts and dismantling the only majority-white district. They argue that the Voting Rights Act requires that this additional majority-minority district be created even though minorities have more than substantially proportionate representation in the Bronx. The claims by the plaintiffs-intervenors similarly seek to create a supermajority Latino district and a Hispanic-black minority-coalition district, respectively, but cannot satisfy the prerequisites for showing vote-dilution under section 2.

In short, the plaintiffs have failed to establish their claims that the 2002 State Senate or congressional redistricting plans violated the Fourteenth Amendment of the United States Constitution or the Voting Rights Act.

II. BACKGROUND

Based upon the 2000 census results, which reflected New York’s lower population growth relative to that of other states, the State Legislature was required to reduce the State’s congressional seats from 31 to 29. See Act of June 5, 2002, 2002 N.Y. Laws 86; see also U.S. Const. art. I, § 4; 2 U.S.C. § 2a. The 2000 census figures, which showed population gains in some New York counties and losses in others (see Pls.’ Ex. (“PX”) 6 tbl. 3), also obligated the Legislature to “readjust” State Senate and Assembly districts so that each district would contain, consistent with applicable districting principles, “an equal number of inhabitants, excluding aliens, and be in as compact form as practicable.” N.Y. Const, art. III, § 4.

The entity responsible for developing redistricting plans for the New York State Legislature is the State Legislative Task Force on Demographic Research and Reapportionment (“LATFOR”), which was initially created in 1978. 4 In connection with the 2002 redistricting, LATFOR began holding public hearings throughout the State, including hearings held in Syracuse, Binghamton, each of the five boroughs of New York City, Suffolk County, Westchester County, Rochester, and Buffalo, in May-July 2001, and received written comments and proposals from the public regarding Senate, Assembly, and congressional districts. During March 2002, LATFOR received additional public comment during a second round of hearings throughout the State (Stip-¶ 13), including hearings held in Buffalo, Rochester, Brooklyn, Queens, the Bronx, Manhattan, Suffolk County, and Albany. Despite its efforts, by the end of 2001, LATFOR had not yet proposed plans for the state legislative or congressional districts.

On January 25, 2002, the plaintiffs, who include black, Hispanic, and white New York voters, filed companion lawsuits against various New York State officials, including Governor George E. Pataki, in the New York State Supreme Court, New York County (“Allen ” case), and in this Court (“Rodriguez ” case), seeking declaratory and injunctive relief against the use of the State’s then-existing congressional and state legislative districts. (Stip. ¶¶ 8-9); Allen v. Pataki, No. 101712/02 (N.Y.Sup.Ct.) (Cahn, J.); Rodriguez v. Pataki, No. 02-618 (S.D.N.Y. filed Jan. 25, 2002). The lawsuits alleged, among other things, that, “based on the 2000 Census, those districts violate the United States Constitution and the Voting Rights Act of 1965 as amended.” (See Rodriguez, Compl. J 1 (Jan. 25, 2002).) 5 The plaintiffs requested that each court “set a reasonable deadline for state authorities to enact redistricting plans and obtain [United States Department of Justice] pre-clearance thereof’ and adopt and promulgate new districts in the event of a failure by the Legislature to act in time for the 2002 elections. {Id. at 17.)

The two cases proceeded along parallel tracks until approximately May 20, 2002 when Allen was removed to this Court and then consolidated with Rodriguez. The cases have been heard by this three-judge Court because of the assertion of interrelated constitutional and Voting Rights Act claims. See 28 U.S.C. § 2284(a); Order at 1, Rodriguez, No. 02-618 (RMB), 2002 WL 818086 (Apr. 8, 2002) (granting “plaintiffs’ request for the appointment of a three-judge district court pursuant to 28 U.S.C. § 2284(a), which provides that ‘[a] district court of three judges shall be convened ... when an action is filed challenging the constitutionality of the apportionment of congressional districts or the apportionment of any statewide legislative body’ ”); see also Page v. Bartels, 248 F.3d 175, 190 (3d Cir.2001) (“[B]ecause statutory Voting Rights Act challenges to statewide legislative apportionment are generally inextricably intertwined with constitutional challenges to such apportionment, those claims should be considered a single ‘action’ within the meaning of § 2284(a).”).

Meanwhile, on February 14, 2002, LAT-FOR publicly released preliminary proposals for new Senate and Assembly districts — but not for congressional districts. (StipJ 10.) The Senate Plan, which was developed and supported by the Republican Senate majority and which was ultimately enacted, proposed to increase the size of the Senate from 61 to 62 seats. It also added three new majority-minority districts within New York City: a majority-Hispanic district principally in Manhattan (with part in the Bronx), a majority-black district in Brooklyn, and a majority- Hispanic district in Queens. The Senate Plan’s ideal population for each district is 306,072 persons; its maximum deviation— that is, the difference between the most populated and the least populated districts — is 9.78%; and the average deviation from the ideal district size is 2.22%. (StipJH 44-45, 47.) Senate Democrats devised two alternative plans, one of which was a 61-seat plan prepared by the Senate Minority and submitted to LATFOR in late 2001 and early 2002 (“the Senate Minority 61-seat plan”) (see Aff. of Mark Burgeson, dated Nov. 12, 2003, ¶ 3 (“Burgeson Aff.”) 6 ), and the other was a 62-seat plan first made public by (then) LATFOR Co-Chairman and Democratic Sen. Richard Dollinger immediately prior to the enactment of the Senate Plan on April 8, 2002 (“Dollinger 62-seat plan”). 7 The Dollinger 62-seat plan proposed adding a fifth majority-minority district in the Bronx (44.8% Hispanic VAP; 18.9% black VAP; 27.4% white VAP) (Stip. ¶ 73; Def. Ex. (“DX”) 67) and eliminating the majority-white district included in the Senate Plan. At the public hearings in the Bronx both before and after the LATFOR Task Force released its redistricting plan for public comment in February 2002, no speaker had ever suggested the creation of an additional Hispanic district in the Bronx or a majority-Hispanic/black “coalition” district in Bronx/Westchester; and prior to April 8, 2002, no one had ever publicly suggested a fourth Hispanic district. (Stip-¶¶ 72,151.)

The 2002 Assembly Plan (“Assembly Plan”) was developed by the Democratic Assembly. It maintains the number of Assembly districts at 150. The Assembly Plan’s maximum deviation is 9.43% and its average deviation is 2.67%. (Id. ¶¶ 46, 48.) In the Assembly Plan, the seven majority-Hispanic districts in the Bronx have an average Hispanic VAP of approximately 59%. (Id. ¶ 158.) In March 2002, Thomas Doty et al. filed a petition in the New York State Supreme Court, Tioga County, challenging the proposed Assembly districts. Justice Herman Cahn of that court ordered that the Doty action be consolidated with the. Allen action, and the issues involved in Doty’s application appear to have been resolved prior to the removal of Allen to federal court. 8

At a public meeting held on April 8, 2002, LATFOR adopted final proposals for Senate and Assembly districts. (Id. ¶ 16.) At the same meeting, Sen. Dollinger stated that he was not, in fact, submitting his 62-seat plan as a proposal for the LATFOR Task Force to consider and adopt: “I do not intend to move its adoption. If I can convince you that you should adopt this plan’s underlying principles, you will still undoubtedly wish to make many changes. The plan I put before you is a legal benchmark.” (DX 133 (Tr. of LATFOR Public Meeting, dated Apr. 8, 2002, at 17).)

Two days later, on April 10, 2002, the Legislature enacted Assembly Bill No. 11014, which incorporated LATFOR’s proposed new Senate and Assembly districts. All three Hispanic senators representing Bronx districts (Sens. Olga Mendez (SD 28), Pedro Espada (SD 32), and Efrain Gonzalez (SD 31)) voted in favor of the Senate Plan (see DX 135), along with Sen. Guy Yelella (SD 34), who is white. Sen. Ruth Hassell-Thompson, the only black senator from the Bronx, did not vote in favor of the Senate Plan but testified before the vote that she “supported the lines that were drawn that would configure her district.” (StipY 107.) On April 22, 2002, the Legislature passed Senate Bill No. 7300, which included certain chapter amendments to the redistricting legislation. Governor Pataki signed Assembly Bill No. 11014 into law (Chapter 35 of the Laws of 2002) on April 22, 2002, and he signed Senate Bill No. 7300 into law on April 24, 2002 (Chapter 38 of the Laws of 2002). (Id. ¶¶ 17-19.)

On April 25, 2002, the plaintiffs amended their complaint to challenge the newly enacted Senate Plan and also to seek judicial intervention with respect to the looming congressional redistricting impasse. (Id. ¶ 20.) See Rodriguez, Am. Compl. ¶¶ 215, 277-78 (Apr. 25, 2002). On the same day, the Allen plaintiffs moved in state court for a preliminary injunction prohibiting the use of the Senate Plan for the 2002 elections and seeking to impose a court-ordered plan in its place. (StipY21.) On May 9, 2002, following a hearing, Justice Cahn denied the plaintiffs’ application in Allen, finding that neither injunctive relief nor an expedited trial was warranted. See Allen, Order at 2, 5 (May 9, 2002).

By order dated April 26, 2002, this Court appointed former United States District Judge Frederick B. Lacey as Special Master, pursuant to Fed.R.Civ.P. 53, to assist in developing a redistricting plan dividing New York into 29 congressional districts. Rodriguez, 207 F.Supp.2d 123, 125 (2002) (“[T]he ‘eleventh hour’ is upon us, if indeed it has not already passed. It is therefore necessary ... to prepare for the possibility that this Court will be required to adopt an appropriate [congressional] redistricting plan.”). 9 Following a series of comprehensive public hearings, and after conferring with leading redistricting experts, on May 13, 2002, Special Master Lacey filed his Report and Plan for congressional redistricting with this Court (“Lacey Plan”). On May 20, 2002, Referee Kenneth Bialkin filed a similar congressional redistricting plan in the state court action. See Rodriguez v. Pataki, Nos. 02-618 (RMB), 2002 WL 1058054, at *3 (S.D.N.Y. May 24, 2002) (three-judge court). On May 24, 2002, this Court adopted the Lacey Plan and ordered its use for the 2002 congressional elections, stating “Special Master Lacey’s Plan meets all applicable legal requirements and ... its immediate adoption is required to ensure a timely and orderly New York State Congressional election process.” Id. at *4. The Court specifically noted that it was “willing, even eager, to accommodate timely state action and ... open to the possibility of withdrawing the Plan we are adopting if the State were to enact an appropriate and lawful plan of its own that allows for a full, fair, and orderly election process.” Id. at *8. Shortly thereafter, on June 5, 2002, the Legislature did adopt a congressional redistricting plan of its own and passed Senate Bill No. 7356, which Governor Pataki signed into law the same day (Chapter 86 of the Laws of 2002). 10

Precleamnce

Because several New York counties, namely, New York, Bronx, and Kings (Brooklyn) Counties, are “covered” jurisdictions under section 4(b) of the Voting Rights Act, 42 U.S.C. § 1973b(b), the legislature’s 2002 Senate, Assembly, and Congressional Plans were each submitted to the United States Department of Justice for “preclearance.” 11 (See, e.g., PX 27 at 1 (“The 2002 Senate districts in [the covered] boroughs, as well as those in non-covered jurisdictions, were designed to avoid any retrogression in minority voting strength. By increasing the number of electable districts without compromising minority electoral opportunities in the preexisting districts, the 2002 plan clearly enhances minority voting strength.”).) On June 17, 2002, the Department of Justice precleared the Senate and Assembly Plans, including specifically the increase in Senate districts from 61 to 62. (See, e.g., DX 138.) The Department of Justice pre-cleared the Congressional Plan on June 25, 2002. (Stip.f 35.) After the Congressional Plan was precleared, this Court withdrew the Lacey Plan. See Rodriguez, 2002 WL 1334733, at *1 (June 25, 2002) (three-judge court). 12

In November 2002, the first New York elections were held under the new state legislative and congressional redistricting plans. The Democrats had a net gain of six seats in the Assembly, giving them a 103/47 majority. The Republicans had a net gain of one seat in the Senate (in addition to the post-election change to the Republican party of Sen. Mendez), increasing their majority to 38/24. The congressional redistricting, which had paired two incumbent Democrats (Reps. Louise Slaughter and John LaFalce) and two incumbent Republicans (Reps. Benjamin Gil-man and Sue Kelly), resulted in the loss of one Democratic congressional seat and one Republican congressional seat after Reps. LaFalce and Gilman each decided to retire rather than run against another incumbent. 13

The Joint and Consolidated Amended Complaint

On January 24, 2003, the plaintiffs and plaintiffs-intervenors 14 filed a Joint and Consolidated Amended Complaint (“the Complaint”) that stated the following causes of action:

• Count I — the Senate Plan violates the “one person, one vote” requirement of the Fourteenth Amendment (Compl. ¶ 278 (“The Legislature did not make an honest or good-faith effort to construct Senate districts as nearly of equal population as is practicable.”));

• Count II — the Senate Plan violates section 2 of the Voting Rights Act (id. ¶ 299 (“By failing to draw additional districts in which minority voters would have the opportunity to elect the candidate of their choice, the State Senate Plan dilutes the votes of African-American and Latino voters.”));

• Count III — Bronx-based Senate districts violate section 2 of the Voting Rights Act (id. ¶ 305 (“Three discrete discriminatory practices resulted in the Legislature’s failure to draw a majority-minority district in northern Bronx and southern Westchester counties.”));

• Count IV — SD 31 in Manhattan/Bronx violates section 2 of the Voting Rights Act (id. ¶ 314 (“The 31st district in the State Senate Plan has the result and effect of denying or abridging the right of Latinos to vote .... ”));

• Count V — Nassau County Senate districts violate section 2 of the Voting Rights Act (id. ¶ 316 (“By fragmenting into multiple districts politically cohesive, compact African-American and Latino communities in Nassau County, who are sufficiently numerous to form a majority-minority district, the State Senate Plan has the result of diluting the voting power of [these] voters .... ”));

• Count VI — Suffolk County Senate districts violate section 2 of the Voting Rights Act (id. ¶ 323 (“By fragmenting into multiple districts the politically cohesive, compact African-American and Latino communities in Suffolk County, who are sufficiently numerous to form a district in which minority voters would have the opportunity to elect the candidate of their choice, the State Senate Plan has the result of diluting the voting power of [these] communities .... ”));

• Count VII — SD 34 in the Bronx/West-chester is a racial gerrymander in violation of the Fourteenth Amendment (id. ¶ 336 (“The predominant purpose of the lines of Senate District 34 was to include non-Hispanic white persons on the basis of race and segregate those non-Hispanic white persons from minority communities in adjoining majority-minority districts.”)); and

• Count VIII — CD 17 in the Bronx, Westchester, and Rockland counties violates section 2 of the Voting Rights Act (id. ' ¶ 347 (“As drawn and implemented, the newly enacted Congressional District 17 has the result and effect of diluting the voting strength of Black voters ... in Bronx County and throughout the City of New York.”)). 15

The defendants filed their Answer to the Complaint on July 18, 2003, denying each of the plaintiffs’ claims.

On March 7, 2003, the defendants moved to dismiss Counts I, II, VI, VIII, and IX of the Complaint pursuant to Fed.R.Civ.P. 12(b)(6). On June 24, 2003, following extensive briefing and oral argument, the Court denied the motion. Rodriguez v. Pataki, 274 F.Supp.2d 363, 364 (S.D.N.Y.2003) (three-judge court) (“Taking the allegations in the ... Complaint as true as the Court must upon a motion to dismiss, it is appropriate to allow discovery to proceed rather than to resolve these counts at this time. The Court is not here ruling upon the ultimate merits of the parties’ respective claims.” (citations omitted)). Extensive discovery ensued and was completed on or about September 26, 2003.

Summary Judgment Motion

On October 3, 2003, the defendants moved for summary judgment pursuant to Fed.R.Civ.P. 56 with respect to all pending claims, and the plaintiffs thereafter cross-moved for summary judgment on Count I. In addition to the extensive briefing, 16 the Court heard lengthy oral argument on November 4, 2003. On November 6, 2003, the Court granted summary judgment in favor of the defendants with respect' to Counts I, II, IV, VI, and VIII of the Complaint, and advised the parties that a full opinion incorporating .the counts decided on summary judgment would issue following the trial of Counts III, V, and VII. Rodriguez, Nos. 02-618 & 02-3239, 2003 WL 22853022, at *1 (S.D.N.Y. Nov.6, 2003) (three-judge court).

Trial

The trial was held before this Court between November 20 and November 25, 2003. It was preceded by the submission of extensive written filings by all parties which included, among other things, some 273 proposed exhibits; 252 stipulated facts (with subparts); legal memoranda; and proposed findings of fact and conclusions of law. Because no jury was involved, and in order to expedite the trial, the parties also submitted 243 pages of direct testimony by affidavit or declaration in advance of trial. 17

The plaintiffs’ witnesses included: Professor Michael P. McDonald, an expert on voting behavior; Professor Andrew A. Beveridge, an expert on demographics and statistically-based social science data; Professor Ronald Hayduk, an expert, on voter participation; Dr. Lisa Handley, an expert on redistricting and voting rights; Professor John Powell, an expert on the socioeconomic status of minorities; State Senator and Minority Leader David A. Paterson; Councilwoman Dorothy L. Goosby, Town of Hempstead (Nassau County); Roger Corbin, Nassau County Legislator; former New York State Assemblywoman Barbara Patton; and Donald Shaffer, a civil rights attorney. The defendants’ witnesses included: Professor Harold W. Stanley, an expert on voting behavior and elections; Professor Stephan Thernstrom, an expert on social, political, historical, and economic issues; and Mark Burgeson, an expert on redistricting and the assistant to the (Republican) Co-Chairman of LATFOR. 18

Following trial, on December 8, 2003, the parties submitted 454 pages of additional proposed findings of fact and conclusions of law; on December 11, 2003, they filed responses to each other’s post-trial submissions. The parties also submitted letters describing recent redistricting decisions in other jurisdictions. 19 Our decision is based upon a close review of the extensive record.

Legal Standards for Summary Judgment and Trial

With respect to Counts I, II, IV, VI, and VIII, we have applied the standards for granting summary judgment based on the papers submitted to the Court and the record developed with respect to the mo tions for summary judgment. Summary judgment may not be granted unless “the pleadings, depositions, answers to interrogatories, and admissions on file, together with the affidavits, if any, show that there is no genuine issue as to any material fact and that the moving party is entitled to a judgment as a matter of law.” Fed.R.Civ.P. 56(c); see also Celotex Corp. v. Catrett, 477 U.S. 317, 106 S.Ct. 2548, 91 L.Ed.2d 265 (1986).

In determining whether summary judgment is appropriate, a court must resolve all ambiguities and draw all reasonable inferences against the moving party. See Matsushita Elec. Indus. Co. v. Zenith Radio Corp., 475 U.S. 574, 587, 106 S.Ct. 1348, 89 L.Ed.2d 538 (1986) (citing United States v. Diebold, Inc., 369 U.S. 654, 655, 82 S.Ct. 993, 8 L.Ed.2d 176 (1962)). But summary judgment is appropriate “when the nonmovant fails to demonstrate that there is sufficient summary judgment evidence to allow a reasonable fact finder to find in its favor on all essential issues as to which it would bear the burden of proof at trial.” Chen v. City of Houston, 206 F.3d 502, 505 (5th Cir.2000) (discussing summary judgment standards in context of racial gerrymandering claim); see also Marylanders for Fair Representation, Inc. v. Schaefer, 849 F.Supp. 1022, 1030 (D.Md.1994) (three-judge court) (explaining that to survive summary judgment dismissal plaintiff must show “there is sufficient evidence from which a reasonable factfinder could find in its favor”).

The substantive law governing the case will identify those facts that are material and “only disputes over facts that might affect the outcome of the suit under the governing law will properly preclude the entry of summary judgment.” Anderson v. Liberty Lobby, Inc., 477 U.S. 242, 248, 106 S.Ct. 2505, 91 L.Ed.2d 202 (1986). In the context of voting rights cases, a court’s summary judgment analysis should account for “the sensitive nature of redistricting and the presumption of good faith” owed to state legislatures. See Miller v. Johnson, 515 U.S. 900, 916, 115 S.Ct. 2475, 132 L.Ed.2d 762 (1995); see also id. at 916-17, 115 S.Ct. 2475 (cautioning courts, in context of racial gerrymandering claim, to “recognize these principles, and the intrusive potential of judicial intervention into the legislative realm, when assessing under the Federal Rules of Civil Procedure the adequacy of a plaintiffs showing at the various stages of litigation” (citing Fed.R.Civ.P. 56; Celotex, 477 U.S. at 327, 106 S.Ct. 2548)); Chen, 206 F.3d at 505.

With respect to the claims in this case, Count I is a “one person, one vote” challenge under the Equal Protection Clause. For such claims, where the population deviation in the legislature’s plan is sufficiently minor and the plan is prima facie constitutional — as, for the reasons explained herein, it is in this case — the plaintiffs must produce evidence that raises a genuine issue of fact as to whether the plan was adopted based on “unconstitutional or irrational” reasons. Marylanders, 849 F.Supp. at 1032; see also Baines v. Masiello, 288 F.Supp.2d 376, 386-87 (W.D.N.Y.2003). Counts II, IV, VI, and VIII are vote-dilution claims under section 2 of the Voting Rights Act. On a motion for summary judgment, such claims are viewed in light of the three threshold requirements articulated in Thornburg v. Gingles, 478 U.S. 30, 106 S.Ct. 2752, 92 L.Ed.2d 25 (1986). See Marylanders, 849 F.Supp. at 1045 (explaining that three “Gingles preconditions” act as threshold for summary judgment).

With respect to Counts III, V, and VII, which were tried, we have carefully reviewed the extensive record and assessed the demeanor and credibility of the witnesses at trial. Our decision with re spect to those counts includes our Findings of Fact and Conclusions of Law as required by Fed.R.Civ.P. 52(a). The plaintiffs bear the burden of proving their claims by a preponderance of the evidence. See, e.g., Easley v. Cromartie, 532 U.S. 234, 258, 121 S.Ct. 1452, 149 L.Ed.2d 430 (2001) (providing standard for burden of proof in racial gerrymandering claim); Miller, 515 U.S. at 916, 115 S.Ct. 2475; see also Voinovich v. Quilter, 507 U.S. 146, 155-56, 113 S.Ct. 1149, 122 L.Ed.2d 500 (1993) (finding district court in error,, in section 2 vote-dilution claim, for placing burden of justifying apportionment on State); id. at 157, 113 S.Ct. 1149 (stating that, under Gingles, “plaintiffs claiming vote dilution ... must prove three threshold conditions,” and “plaintiffs 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”).

III. ONE-PERSON, ONE-YOTE CLAIM (COUNT I)

A. Legal Standards

The “one person, one vote” principle is grounded in the Equal Protection Clause of the Fourteenth Amendment. Reynolds v. Sims, 377 U.S. 533, 84 S.Ct. 1362, 12 L.Ed.2d 506 (1964). It prohibits the dilution of individual voting power by means of state districting plans that allocate legislative seats to districts of unequal populations and thereby diminish the relative voting strength of each voter in overpopulated districts. In Reynolds, the Court held that “the Equal Protection Clause requires that the seats in both houses of a bicameral state legislature must be apportioned on a population basis.” The Court required states to “make an honest and good faith effort to construct districts ... as nearly of equal population as is practicable.” Id. at 577, 84 S.Ct. 1362.

While the Supreme Court has held that absolute population equality is required for congressional districts, Karcher v. Daggett, 462 U.S. 725, 732-33, 103 S.Ct. 2653, 77 L.Ed.2d 133 (1983), districting plans for state legislative seats require only “substantial” population equality. See Gaffney v. Cummings, 412 U.S. 735, 748, 93 S.Ct. 2321, 37 L.Ed.2d 298 (1973). The Court has recognized that minor deviations from absolute population equality may be necessary to permit states to pursue other legitimate and rational state policies. See Reynolds, 377 U.S. at 577-81, 84 S.Ct. 1362; see also Mahan v. Howell, 410 U.S. 315, 321-22, 93 S.Ct. 979, 35 L.Ed.2d 320 (1973). Particular state policies that justify minor deviations from ábsolute population equality generally include “making districts compact, respecting municipal boundaries, preserving the cores of prior districts, and avoiding contests between incumbent Representatives.” Karcher, 462 U.S. at 740, 103 S.Ct. 2653.

Because the promotion of these important state policies will often necessitate “minor deviations” from absolute population equality, the Court has held that such minor deviations, alone, are insufficient to establish a prima facie ease of invidious discrimination. Voinovich v. Quilter, 507 U.S. 146, 160-62, 113 S.Ct. 1149, 122 L.Ed.2d 500 (1993). In Brown v. Thomson, 462 U.S. 835, 842, 103 S.Ct. 2690, 77 L.Ed.2d 214 (1983), the Court held that redistricting plans with a maximum population deviation below ten percent fall within the category of minor deviations that are insufficient to establish a prima facie violation of the Equal Protection Clause. “Thus, a redistricting plan with a maximum deviation below ten percent is. prima facie constitutional and there is no burden on the State to justify that- devia tion.” Marylanders for Fair Representation, Inc. v. Schaefer, 849 F.Supp. 1022, 1031 (D.Md.1994) (three-judge court); see Holloway v. Hechler, 817 F.Supp. 617, 623 (S.D.W.Va.1992) (three-judge court), aff'd mem., 507 U.S. 956, 113 S.Ct. 1378, 122 L.Ed.2d 754 (1993); Fund for Accurate & Informed Representation, Inc. v. Weprin, 796 F.Supp. 662, 668 (N.D.N.Y.) (three-judge court), aff'd mem., 506 U.S. 1017, 113 S.Ct. 650, 121 L.Ed.2d 577 (1992); Gorin v. Karpan, 788 F.Supp. 1199, 1201 (D.Wyo.1992) (three-judge court); Cosner v. Dalton, 522 F.Supp. 350, 357 n. 11 (E.D.Va.1981) (three-judge court).

Compliance with Brown’s “ten percent rule” does not end the inquiry. There is still a question of how the “ten percent rule” dovetails with Reynolds and its progeny, which require a “good faith effort” by the state to achieve “as nearly of equal population as is practicable.” Reynolds, 377 U.S. at 577, 84 S.Ct. 1362. For example, a three-judge court in Hastert v. State Board of Elections, 777 F.Supp. 634, 645 (N.D.Ill.1991), suggested that “minute population deviations remain legally significant.” The court in Corbett v. Sullivan, 202 F.Supp.2d 972, 987 n. 7 (E.D.Mo.2002) (citing Karcher, 462 U.S. at 738-40, 103 S.Ct. 2653), held that “[e]ven deviations smaller than the census margin of error must be the result of a good faith effort to achieve population equality.”

The defendants’ summary judgment brief argued, in substance, that the “ten percent rule” effectively creates a “safe harbor” that forecloses all challenges to redistricting plans in which the maximum deviation is below that percentage. Supporters of this per se rule point to the fact that nearly no state districting plan with a maximum deviation below ten percent has ever been struck down by a court as violating population equality. Moreover, the language of some recent decisions seem to support the per se rule. The Weprin court, for example, indicated that plaintiffs who fail to demonstrate a maximum population deviation in excess of ten percent cannot survive a motion to dismiss. See Weprin, 796 F.Supp. at 668 (“[AJbsent credible evidence that the maximum deviation exceeds 10 percent, plaintiffs fail to establish a prima facie case of discrimination under [the one-person, one-vote] principle sufficient to warrant further analysis by this [c]ourt.”). But see Cecere v. County of Nassau, 274 F.Supp.2d 308, 311-12 (E.D.N.Y.2003) (finding claims where deviation is less than ten percent justiciable if some other discriminatory purpose is shown over and above obvious political motivations); Licht v. Quattrocchi, 449 A.2d 887 (R.I.1982) (striking plan with five-percent deviation).

The plaintiffs concede that if the maximum population deviation between districts in a redistricting plan is under ten percent, the state has no burden to justify that deviation. The plaintiffs argue, however, that they may successfully challenge a plan with a maximum deviation below ten percent if they can prove that the “minor deviation” does not result from the promotion of other legitimate state policies, but rather from an impermissible or irrational purpose. We think that Brown, Mahan, Gaffney, and Abate v. Mundt, 403 U.S. 182, 187, 91 S.Ct. 1904, 29 L.Ed.2d 399 (1971), lend support to the proposition that the “ten percent rule” is not meant to protect a state that is systematically disadvantaging groups of voters with no permissible rational justification for the disproportion.

We conclude, with Marylanders, 849 F.Supp. at 1032, that a plan within the “ten percent rule” is not per se immune from judicial review. No decision explicitly adopts the per se rule. Weprin — the case that comes closest to stating the rule — did not involve a plaintiff claiming unconstitutional or irrational state policies. See Weprin, 796 F.Supp. at 668. Thus, “if the plaintiff can present compelling evidence that the drafters of the plan ignored all the legitimate reasons for population disparities and created the deviations solely to benefit certain regions at the expense of others,” a one-person, one-vote action will lie even with deviations below ten percent. See Legislative Redistricting Cases, 331 Md. 574, 629 A.2d 646, 657 (1993); see also Licht, 449 A.2d at 887 (finding deviation of five percent to violate one-person, one-vote requirement because deviation “negate[d] the effects of reapportionment”).

Moreover, in light of recent technological changes, there is a reason not to allow the state systematically to dilute the votes of certain classes of citizens simply because the state is able to keep its discrimination within a ten-percent deviation. The powerful computer programs of today allow states to manipulate districting lines to alter voting patterns within a district with a high degree of precision. Under these circumstances, we see no reason to give a state operating within the ten-percent margin immunity from all review as to whether it is acting irrationally or under-' taking invidious discrimination. The benefit of flexibility to pursue legitimate state policies that states receive under the “ten percent rule” since Brown carries with it a responsibility not to use the rule to frustrate the very purpose of the decennial census and systematically discriminate against a group of voters.

Appropriate review can be achieved under the burden-shifting regime upheld in Marylanders where

a plaintiff [can], with appropriate proof, successfully challenge a redistricting plan with a maximum deviation below ten percent. To prevail, though, the plaintiffs have the burden of showing that the deviation in the plan results solely from the promotion of an uncon-, stitutional or irrational state policy. Thus, the plaintiffs ... must demonstrate ... that the asserted unconstitutional or irrational state policy is the actual reason for the deviation. See Karcher, 462 U.S. at 740-44, 103 S.Ct. 2653. In addition, the plaintiff must prove that the minor population deviation' is not caused by the promotion of legitimate state policies.

Marylanders, 849 F.Supp. at 1032 (emphasis added). If the burden on the plaintiffs in minor-deviation cases were anything less than this substantial showing, then the plaintiffs would be able to challenge any minimally .deviant redistricting scheme based upon scant evidence of ill will by district planners, thereby creating costly trials and frustrating the purpose of Brown’s “ten percent rule.”

B. Analysis

The maximum deviation between Senate districts in the Senate Plan is 9.78%, while the average deviation of the Senate districts from the ideal district population is 2.22%. 20 Because the plaintiffs have failed to establish a prima facie ease of unconstitutional discrimination under Brown, the defendants have no burden to justify the plan’s minor deviation. Marylanders, 849 F.Supp. at 1033 (citing Voinovich, 507 U.S. at 160, 113 S.Ct. 1149; Brown, 462 U.S. at 842-43, 103 S.Ct. 2690).

We denied the defendants’ initial motion to dismiss this count of the Complaint, see Rodriguez v. Pataki, 274 F.Supp.2d 363 (S.D.N.Y.2003), to give the plaintiffs an opportunity to meet their burden to show that the minimal deviation results solely from an unconstitutional or irrational state purpose rather than from other state policies recognized by the Supreme Court to be appropriate reasons for deviations. Such policies, announced in Karcher, include “making districts compact, respecting municipal boundaries, preserving the cores of prior districts, and avoiding contests between incumbent [rjepresenta-tives.” Karcher, 462 U.S. at 740, 103 S.Ct. 2653. Following discovery, the defendants moved for summary judgment, arguing that the plaintiffs have failed to raise disputed issues of fact that, if resolved in their favor, could support a finding that they carried their burden. The plaintiffs cross-moved for summary judgment on this count, arguing that they have met their burden with uncontradicted evidence.

The plaintiffs assert that the Senate redistricting scheme impermissibly and arbitrarily discriminates against “downstate” residents (defined by the plaintiffs as Senate Districts 10-38 — notably excluding the Long Island-based districts, Senate Districts 1-9) by systematically overpopulating all of those districts and systematically underpopulating all of the “upstate” districts (defined by the plaintiffs as Senate Districts 39-62). These discriminatory actions, the plaintiffs allege, result in a single Senate district being retained by voters located in the “upstate” area when it should have been eliminated and an additional “downstate” district created. 21 The plaintiffs maintain that this qualifies as invidious discrimination because Reynolds itself disallowed discrimination against rural as opposed to urban residents of the state and thus made “regional discrimination” actionable. The plaintiffs also argue, as they must to meet their burden, that the defendants’ systematic overpopulation and underpopulation lacked any rational or constitutionally permissible purpose. By implication, the plaintiffs suggest that racial bias may have animated the plan because all fourteen majority-minority Senate districts were overpopulated and are “downstate,” where most of the state’s minority population lives. 22 Finally, the plaintiffs urge that the defendants made no “honest and good faith effort,” as required under Reynolds, to achieve real population equality, but rather adhered to the “ten percent rule” (just barely) because they believed it to be a “safe harbor.”

It is effectively undisputed that in their redistricting plan the defendants were plainly mindful of the “ten percent rule,” as demonstrated by various memoranda of the chief architect of the plan, Mark Burgeson. No invidious purpose can be inferred from such conduct alone; indeed, Brown invites adherence to the “ten percent rule.” Consistent with Marylanders and Farnum v. Burns, 561 F.Supp. 83, 93 (D.R.I.1983) (three-judge court), we find that an express objective of staying within a ten-percent deviation while pursuing other legitimate goals provides no support to the plaintiffs’ claim of invidious or arbitrary discrimination or of bad faith. See Marylanders, 849 F.Supp. at 1034 (reasoning that objective of staying within ten-percent window rather than achieving absolute population equality “merely recognize[s] the flexibility that the State had in order to accommodate other legitimate state policies” and “demonstrates nothing more than the objective of crafting a plan with constitutional population equality”).

To be sure, Vigo County Republican Central Committee v. Vigo County Commissioners, 834 F.Supp. 1080 (S.D.Ind.1993), held that merely trying to stay below ten-percent deviation could not be considered good faith. But Vigo arose under idiosyncratic facts, where the plan’s total deviation when the litigation started was 37%; only to avoid losing in court did the planners consciously try to eliminate pre-existing bad faith by drawing a plan within the ten-percent parameter. See id. at 1085-86. The Vigo court essentially found that the defendants there did too little too late — a finding inapposite here. In sum, in the case before us, the defendants’ conscious use of the “ten percent rule” cannot, without more, support an inference that no legitimate state policies accounted for a minor deviation in a districting plan or that adherence to the “ten percent rule” was a mere pretext for impermissible considerations.-

The plaintiffs have not produced evidence of irrational or unconstitutionally discriminatory behavior by the Legislature, notwithstanding their reliance and emphasis upon Burgeson’s memorandum of July 20, 2001, which the plaintiffs claim reveals the defendants’ intention to discriminate against the “downstate” region. (Netburn Decl., Ex. 39 (Memorandum from Mark Burgeson to Sen. Skelos, dated July 20, 2001).) In that memorandum, Burgeson expresses a preference against increasing the Senate to 63 seats; he suggests that the only place to add a district to a 62-seat plan would be on Long Island to “combine politically undesirable areas.” (Id. at 1.) He further emphasizes that an extra seat cannot be added “upstate” because those districts are already drawn “light, to avoid migration downstate.” (Id.) The plaintiffs argue vigorously that the memorandum’s references to “politically undesirable” areas prove invidious discrimination.

Putting aside the plaintiffs’ questionable assumption that Burgeson’s motives are a proxy for those of the Legislature, a fair reading of the memorandum reveals many permissible redistricting considerations. The memorandum shows that LATFOR was interested in contiguity, compactness, preserving the cores of existing districts, desiring not to pit incumbents against one another, respecting then-current political subdivisions and county lines, and staying within the ten-percent-deviation parameter of Brown. (See generally id.); cf. Marylanders, 849 F.Supp. at 1034. 23 Indeed, this contemporaneous memorandum as sists the defendants at least as much as it assists the plaintiffs because it plainly invokes the permissible policies that Karcher contemplates. Moreover, to the extent that it refers to areas that would be represented by Democrats as “politically undesirable,” we cannot understand what invidious discrimination this phrase supposedly signals: there is no political or racial gerrymandering claim embedded in this count, nor is it surprising that a memorandum to the Republican State Senate in control of redistricting would describe a potential Democratic district as comprised of “undesirable” voters. Finally, it was plainly Democrats on Long Island in Senate Districts 1-9 who were potentially disadvantaged by this decision, not people in the “downstate” region, defined by the plaintiffs as Districts 10-38. Accordingly, the memorandum does not help the plaintiffs meet their burden; it evinces permissible motives and fails to show that the deviations resulted solely from impermissible considerations.

On the record before us, still other reasons support summary judgment in favor of the defendants apart from the fact that the plaintiffs failed to show that the deviation was not caused by the promotion of court-approved state policies. First and foremost, the plaintiffs only beg the question in repeatedly asserting that the pattern of overpopulation and underpopulation “reflects an illegitimate effort to overrepresent an entire region of the State in order to maintain its ascendancy in the Senate, at the expense of another large region which the 2000 Census data showed had grown much more substantially over the past decade.” (Pis. Mem. at 9.) How can we conclude, given the “ten percent rule” we reaffirm today, that the minimal underpopulation of “upstate” districts was driven by regional discrimination rather than other permissible considerations? Just as in the racial context where courts must deal with the overlap of racial identity and partisan identification and still conclude that the districting is racially — and not simply politically— drawn to find an equal protection violation, see, e.g., Easley v. Cromartie, 532 U.S. 234, 121 S.Ct. 1452, 149 L.Ed.2d 430 (2001), so in the one-person, one-vote context must the plaintiffs who challenge a plan with less than a ten-percent deviation present some evidence that the dis-tricting can be traced to impermissible considerations. In New York State, the traditional correlation between “upstate” districts and Republican political identification (21 out of 24 incumbent Senators upstate are Republican) means that the plaintiffs here needed to proffer more than a mere assertion of a Senate conspiracy for “upstate” ascendancy to meet their burden of showing a violation of the one-person, one-vote principle. 24

Second, to the extent that the plaintiffs seek to use the regional aspect of their claim as a proxy for a claim to that a group of voters were systematically disadvantaged, their proposed definitions of “upstate” and “downstate” are self-serving and defective. Cf. Reynolds, 377 U.S. at 562, 84 S.Ct. 1362 (stating that principle of population equality in districting is aimed to protect citizens and not geographic areas: “Legislators are elected by voters, not farms or cities or economic interests.”). The plaintiffs identified the “downstate” region as Districts 10-38, including New York City, Rockland. County, most of Westchester County, and two towns in Orange County. (Comply 106.) They defined upstate as Districts 39-62, including the northern portion of Westchester, most of Orange County, and all of the counties north and west of Westchester and Orange Counties. (Id.) We are unable to discern why certain border counties that reach north of New York City — and why only portions of two of those counties — are classified as “downstate” on any basis other than that they are overpopulated. Moreover, the Long Island districts (1-9), while decidedly “downstate” geographically, are conveniently excluded from plaintiffs’ “downstate” calculus because they are not overpopulated; they are populated at close to the ideal level. The deposition of the plaintiffs’ witness, Todd Breitbart, confirmed that at least he is “not aware of any generally accepted practice for defining upstate and downstate.” (Stip-¶ 51.) In sum, the plaintiffs have offered no persuasive basis for embracing their selective definition of state regions, which appears to be tailored to suit their litigation strategy-

Finally, the affidavits on summary judgment establish that if every district were apportioned with perfect equality, the difference in “downstate” representation from what was accorded under the enacted plan would be insignificant. New York City would have been entitled to 26.2 seats as compared with the 26 seats accorded to New York City under the enact-, ed plan (with a s.eat defined as representing a district controlled or predominantly controlled by city-based voters). See Marylanders, 849 F.Supp. at 1035 n. 12 (discussing concept of “regional discrimination” and dismissing argument where county “controlled] precisely the number of State Senators that its population indicates it should control”).

We- consider one final point. The defendants, in response to the plaintiffs’ allegation of systematic “downstate” overpopulation, urge us to reassess this allegation in light of population variants: citizen voting age population (“CVAP”) and voting age population (“WAP”), instead of total population (which includes ineligible voters). Using these alternative counting methods, both parties stipulate that the New York City districts within the “downstate” districts would then be substantially underpopulated rather than overpopulated. According to the plaintiffs’ CVAP figures, the New York City districts in the Senate Plan are underpopulated by 12.0% and “upstate” districts are overpopulated by 15.4%. With respect .to registered voters, the Weight óf one New York City resident’s vote, depending on the district of comparison, is worth 29.9% to 63.6% more than an “upstate” citizens vote. (See Stip. ¶¶ 16-17, 22.) Although total population figures are the generally accepted basis for redistricting calculations, the practical effect- of the Senate Plan on those who actually vote shines a very different light on the plaintiffs’ argument that the Senate Plan discriminates against New York City and its voters: the underpopulated “upstate” districts have more eligible citizens and actual voters. The practical effect of the Senate Plan; then, is to dilute the votes of “upstate” residents, not those who reside “downstate.” ■

It does not appear that the New York legislature employed CVAP data in creating its 2002 Senate Plan, 25 and thus the issue is not whether the court should defer to the state’s decision to use a measure other than total population. See, e.g., Burns v. Richardson, 384 U.S. 73, 94-96, 86 S.Ct. 1286, 16 L.Ed.2d 376 (1966) (upholding Hawaii’s use of registered voter statistics in apportionment because “[t]otal population figures may ... constitute a substantially distorted reflection of the distribution of state citizenry” and because use of registered voter statistics “substantially approximated [the apportionment that] would have appeared had state citizen population been the guide”). This is also not a situation where a state’s apportionment plan is within the ten-percent window under one measure but is challenged for being beyond it under a different measure. See Daly v. Hunt, 93 F.3d 1212, 1222 (4th Cir.1996) (involving issue of “the determinative statistic for measuring compliance with the one person, one vote principle” where plan complied with total-population proportionality, but it allegedly “produced an unacceptably high deviation in terms of voting-age population”). In raising the issue of the CVAP distribution in New York State, we are not attempting to promote any particular theory of proportional representation. 26 We simply find that under the circumstances of this case, including the fact that the plan is prima facie constitutional, we do not need to ignore the reality that the overpopulation of New York City districts has not, in fact, diluted the voting strength of “downstate” voters.

Even accepting the plaintiffs’ argument that the Senate Plan was designed to draw upstate districts “light, to avoid migration [of a Senate seat] downstate” (see Netburn Decl. Ex. 39, at 1 (Burgeson Memo, dated July 20, 2001)), we find no constitutional harm. The overall population deviation is within the ten-percent margin of Brown, and the plan is prima facie constitutional. The plan promotes the traditional principles of maintaining the core of districts and limiting incumbent pairing. In this case, the overall effect of the deviation is only one seat (actually, two-thirds of a seat) in a 62-seat Senate. 27 Accordingly, we granted the defendants’ motion for summary judgment on Count I, the population equality count of the Complaint, and we denied the plaintiffs’ cross-motion for summary judgment on the same count. No genuine issues of material fact warranted a trial on this count, and the plaintiffs failed to meet their burden.

IV. VOTING RIGHTS ACT CLAIMS

The Complaint asserts six separate claims of vote dilution pursuant to section 2 of the Voting Rights Act of 1965, 79 Stat. 437, as amended, 42 U.S.C. § 1973, including five State Senate district challenges and one congressional district challenge. The State Senate claims, in the order discussed herein, relate to Suffolk County (Count VI); Nassau County (Count V); the Bronx and a small portion of .West-chester County (Count III); the Senate Plan in its entirety (Count II); and Manhattan and a small portion of the Bronx (Count IV). The congressional claim relates to CD 17, which is located in Bronx, Westchester, and Rockland Counties (Count VIII).

A. Section 2 Legal Standard

Section 2(a) of the Voting Rights Act (“VRA”) provides that “[n]o voting qualification or prerequisite to voting or standard, practice, or procedure shall be imposed or applied by any State ... 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 membership in a language minority group].” 42 U.S.C. § 1973(a). 28 Section 2(b) provides that a “denial or abridgement” occurs where,

based on the totality of circumstances, it is shown- that the political processes leading to nomination or election in the State,... 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 ... 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.

§ 1973(b). To establish- a section 2 violation, plaintiffs must first establish by a preponderance of the evidence three threshold conditions articulated in Thornburg v. Gingles, 478 U.S. 30, 106 S.Ct. 2752, 92 L.Ed.2d 25 (1986):

(1) that the minority group is “sufficiently large and geographically compact to constitute a majority in a single-member district”;

(2) that [the minority group] is “politically cohesive”; and

(3) that “the white majority vot[es] sufficiently as a bloc to enable it ... usually to defeat the minority’s preferred candidate.”

Growe v. Emison, 507 U.S. 25, 40, 113 S.Ct. 1075, 122 L.Ed.2d 388 (1993) (quoting Gingles, 478 U.S. at 50-51, 106 S.Ct. 2752).

With respect to the first Gingles precondition, “[t]he reason that a minority group ... must show, as a threshold matter, that it is sufficiently large and geographically compact to constitute a majority in a single-member district is this: Unless minority voters possess the potential to elect representatives in the absence of the challenged structure or practice, they cannot claim to have been injured by that structure or practice.” Id. at 50 n. 17, 106 S.Ct. 2752. “Because the very concept of vote dilution implies — and, indeed, necessitates — the existence of an ‘undiluted’ practice against which the fact of dilution may be measured, a § 2 plaintiff must also postulate a reasonable alternative voting practice to serve as the benchmark ‘undiluted’ practice.” Reno v. Bossier Parish Sch. Bd., 520 U.S. 471, 480, 117 S.Ct. 1491, 137 L.Ed.2d 730 (1997); accord Holder v. Hall, 512 U.S. 874, 881, 114 S.Ct. 2581, 129 L.Ed.2d 687 (1994).

With respect to the second Gingles precondition, “if the minority group is not politically cohesive, it cannot be said that the [challenged structure, practice, or procedure] thwarts distinctive minority group interests.” Gingles, 478 U.S. at 51, 106 S.Ct. 2752. Proof of cohesion in actual elections is required. See Sanchez v. Colorado, 97 F.3d 1303, 1312 (10th Cir.1996); Gomez v. City of Watsonville, 863 F.2d 1407, 1415 (9th Cir.1988); see also NAACP v. City of Niagara Falls, 65 F.3d 1002, 1018 (2d Cir.1995).

With respect to the third Gingles precondition, a plaintiff must demonstrate that elections are racially polarized — that “there is a consistent relationship between the race of the voter and the way in which the voter votes,” Gingles, 478 U.S. at 53 n. 21, 106 S.Ct. 2752 (internal quotations marks and alterations omitted)' — and that “the white majority votes sufficiently as a bloc to enable it ... usually to defeat the minority’s preferred candidate.” Gingles, 478 U.S. at 51, 106 S.Ct. 2752; see also id. at 55-58, 63, 106 S.Ct. 2752; Jenkins v. Red Clay Consol. Sch. Dist. Bd. of Educ., 4 F.3d 1103, 1123 (3d Cir.1993) (“The correct question is not whether white voters demonstrate an unbending or unalterable hostility to whoever may be the minority group’s representative of choice, but whether, as a practical matter, the usual result of the bloc voting that exists is the defeat of the minority-preferred candidate.”).

The three Gingles preconditions are necessary — but not sufficient — to establish a section 2 violation. “[I]f Gingles so clearly identified the three [preconditions] as generally necessary to prove a § 2 claim, it just as clearly declined to hold them sufficient in combination, either in the sense that a court’s examination of relevant circumstances was complete once the three factors were found to exist, or in the sense that the three in combination necessarily and in all circumstances demonstrated dilution.” Johnson v. De Grandy, 512 U.S. 997, 1011, 114 S.Ct. 2647, 129 L.Ed.2d 775 (1994). Rather, a plaintiff asserting vote dilution must also 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, 122 L.Ed.2d 500 (1993). The relevant inquiry may include the list of factors set forth in the United States Senate Judiciary Report (“Senate Report”) accompanying the 1982 bill amending section 2 of the Voting Rights Act, S.Rep. No. 97-147, at 2 (1982), 1982 U.S.C.C.A.N. 177, 177-78 (“Senate Factors”), but this is neither a “comprehensive nor exclusive” list. Gingles, 478 U.S. at 44-45, 106 S.Ct. 2752; see S.Rep. No. 97-147 at 28-29, 1982 U.S.C.C.A.N. at 206-07 (“[The Senate Factors] often will be the most relevant ones, [but] in some ... cases other factors will be indicative of the alleged dilution”). 29 “[T]he ultimate conclusions about equality or inequality of opportunity were intended by Congress to be judgments resting on comprehensive, not limited, canvassing of relevant facts.” De Grandy, 512 U.S. at 1011, 114 S.Ct. 2647.

B. Long Island: Nassau County and Suffolk County (Counts V and VI)

The plaintiffs have brought section 2 vote-dilution claims challenging the Senate Plan with respect to two counties on Long Island: Nassau County and Suffolk County. The plaintiffs contend that in each county, despite the fact that the minority population has been rapidly growing, district lines continue to divide minority comT munities. Thus, the plaintiffs claim, the districts in Nassau and Suffolk Counties present a classic case of “cracking,” where minority voting strength is diluted by dividing the minority population across several districts and thus submerging in each district the minority community in 'a white majority.

While the black and Hispanic populations have indeed grown significantly on Long Island, particularly during the decade prior to the 2000 Census, in neither Suffolk County nor Nassau County is either minority group large enough itself to form a majority in a Senate district. The plaintiffs proposed two districts, one in Suffolk County and one in Nassau County, as black-Hispanic coalition districts. The first district was Plaintiffs’ Proposed District 4 in Suffolk County, which would encompass minority populations from portions of the current SDs 3, 4, and 8. 30 It would have a white VAP of 56.4%, a non-Hispanic black VAP of 17.5%, a Hispanic VAP of 22.7%, and a combined black and Hispanic VAP of 40.2%. See infra Table 1. Plaintiffs’ Proposed District 8 would include sections from all four Nassau County Senate districts — SDs 6, 7, 8, and 9 — and would have a white VAP of 40.7%, a non-Hispanic black VAP of 34.5%, a Hispanic VAP of 19.8%, and a combined black and Hispanic VAP of 54.3%. Id. However, the combined black and Hispanic CVAP would be only 47.6% because in those districts almost half of Hispanics of voting age are not citizens. Id.

The voting age populations of the Senate districts under the 2002 Senate Plan that comprise Plaintiffs’ Proposed Districts 4 and 8 are as follows:

_Table 1: Voting Age Populations fob Suffolk: AND Nassau CouNties 31 _

Non-Hispanic Non-Hispanic Hispanic Hispanic + District 32 _White VAP Black VAP 33 VAP Black VAP

2002 Senate Plan_

SD 3(S)_74,25%_7.09%_15.59%_22.68%

SD 4(S)_76.41%_8.77%_11,92%_20.69%

SD 6(N)_69.04%_15.87%_11,15%_27.02%

SD 7(N)_73.07%_8.36%_9.26%_17.62%

SD 8(N/S)_71.85%_15.06%_10.31%_25.37%

SD 9(N)81.44%5.67%8.48%14.15%

Suffolk County:

Proposed District 56.38% 17.53% 22.65% 40.19%

4 34 (CVAP)_(63.U%)_(18.11%)_(15.61%)_(33.72%)

Nassau County:

Proposed District 40.65% 34.51% 19.81% 54.32%

8 35 (CVAP)(18.20%)(36.85%)(10.76%)(17.61%)

As pleaded, the plaintiffs’ Long Island vote-dilution claims face a number of obstacles in their attempt to satisfy the first Gingles precondition, which, when read literally, requires that a minority group constitute a majority of the relevant population in a single-member district. Gingles, 478 U.S. at 49, 106 S.Ct. 2752 (“First, the minority group must be able to demonstrate that it is sufficiently large and geographically compact to constitute a majority in a single-member district.”).

The first problem arises from the fact that the Complaint treated blacks and Hispanics as the equivalent of a single minority group by aggregating the populations of the two minority groups to meet the majority-in-a-district requirement. While the defendants argue that minority groups cannot be aggregated for purposes of a section 2 claim, the Supreme Court has left the question open. See Growe v. Emison, 507 U.S. 25, 41, 113 S.Ct. 1075, 122 L.Ed.2d 388 (1993). The Court of Appeals for the Second Circuit, however, has permitted blacks and Hispanics to be combined for the purposes of complying with the first Gingles precondition, provided the groups are shown to be politically cohesive. See Bridgeport Coalition for Fair Representation v. City of Bridgeport, 26 F.3d 271, 276 (2d Cir.), vacated and remanded on other grounds, 512 U.S. 1283, 115 S.Ct. 35, 129 L.Ed.2d 931 (1994); see also Arbor Hill Concerned Citizens Neighborhood Ass’n v. County of Albany, 281 F.Supp.2d 436, 445 (N.D.N.Y.2003) (citing Bridgeport in stating that Second Circuit law “assumes diverse minority groups can be combined to meet VRA litigation requirements” crafted by Gingles). But see Nixon v. Kent County, 76 F.3d 1381, 1390-92 (6th Cir.1996) (en banc) (ruling against aggregating minority groups to meet first Gingles condition). In any event, it is plain that to the extent minorities can be aggregated, cohesion between groups cannot be presumed, and the plaintiffs bear the burden of proving such cohesion. See Growe, 507 U.S. at 41, 113 S.Ct. 1075.

The more substantial legal question for the purposes of the Nassau and Suffolk County claims is whether a vote-dilution claim can succeed if the relevant minority population is not a majority in the plaintiffs’ proposed districts. That is, the Nassau and Suffolk County claims raise the issue of whether “coalition” or “influence” claims are cognizable under the VRA. Gingles addressed a claim where “the plaintiffs alleged and attempted to prove that their ability to elect the representatives of their choice was impaired by the selection of a multimember electoral structure.” Gingles, 478 U.S. at 46 n. 12, 106 S.Ct. 2752. The Court declined “to consider whether section 2 permits, and if it does, what standards should pertain to, a claim brought by a minority group, that is not sufficiently large and compact to constitute a majority in a single-member district, alleging that the use of a multimember district impairs its ability to influence elections.” Id. As Justice O’Connor pointed out in her Gingles concurrence, the plurality opinion left open a question about the nature of the distinction between majority-minority districts and “influence” districts, as well as whether plaintiffs could make a claim where they showed an ability to elect candidates of choice without being a majority of the population. Id. at 89 n. 1, 106 S.Ct. 2752 (O’Connor, J., concurring). Since Gingles, the Supreme Court has three times declined to reach the issue of whether vote-dilution claims are viable when the minority group is less than a majority of voters in a proposed district. See Johnson v. De Grandy, 512 U.S. 997, 1008-09, 114 S.Ct. 2647, 129 L.Ed.2d 775 (1994); Voinovich v. Quilter, 507 U.S. 146, 158, 113 S.Ct. 1149, 122 L.Ed.2d 500 (1993); Growe, 507 U.S. at 41 n. 5, 113 S.Ct. 1075. 36

Although courts often use the term “influence-dilution” or “influence district” to refer to any district where the minority group would not be a majority of voters, there are two kinds of districts that are conceptually different. A “coalition” or “crossover” district is one where members of the minority group are not a majority of the relevant voting population but nonetheless have the ability to elect representatives of their choice with support from a limited but reliable white crossover vote. See Hall v. Virginia, 276 F.Supp.2d 528, 533-34 & nn.7-10 (E.D.Va.2003) (defining terms and contrasting “coalition districts” with “influence districts”). 37 Coalition or crossover districts are also referred to as “performance” districts, “effective” districts, or “ability to elect” districts. See id. at 533 nn.7-9. An “influence” district is where minority voters may not be able to elect a candidate of choice but can play a substantial, although not decisive, role in the electoral process. See id. at 534 n. 10; see also Georgia v. Ashcroft, 539 U.S. 461, 123 S.Ct. 2498, 2513, 156 L.Ed.2d 428 (2003). Both of these districts are distinct from majority-minority districts, which a literal reading of Gingles requires. Most courts have adopted the literal reading of Gingles and have rejected claims where the minority group does not constitute a majority of voters in a single-member district. See Hall, 276 F.Supp.2d at 536, 538 (listing cases and stating that bright-line majority-in-a-district rule is well-established).

Plaintiffs’ Proposed District 4 in Suffolk County was presented as a pure influence-dilution claim, where no ability for minorities to elect candidates of choice was alleged, even if the black and Hispanic populations are aggregated. For the reasons we will explain, this Court follows the clear majority of courts and therefore rejects the Suffolk County influence-dilution claim. We thus granted summary judgment dismissing Count VI.

In Plaintiffs’ Proposed District 8 in Nassau County, blacks and Hispanics, if aggregated, would constitute 54.3% of the VAP and 47.6% of the CVAP — thus straddling the Gingles majority threshold. See supra Table 1. In Count V of the Complaint, the plaintiffs alleged that blacks and Hispanics should be aggregated and that together they had the ability to elect candidates of their choice in Proposed District 8. (Comply 317.) The Court denied summary judgment dismissing Count V of the Complaint both before trial and when the motion was renewed at trial. All par ties presented evidence on this claim at trial. At trial, the plaintiffs abandoned their contention that blacks and Hispanics voted cohesively in Nassau County, and they reframed their argument to treat Proposed District 8 as a black-only “ability to elect” district with a 34.5% black VAP. Even assuming, for the sake of argument, that such a crossover claim could be cognizable, we find that the plaintiffs have failed to prove that the Senate Plan violated section 2 of the VRA with respect to the districting scheme in Nassau County.

1. Suffolk County (Count VI)

The proposed Suffolk County district would connect a stretch of minority communities from the Nassau/Suffolk border eastward through Islip. (See PX 1 map 15.) Plaintiffs’ Proposed District 4 would encompass the minority population currently distributed across three Senate districts, including: the minority population in the eastern part in the Town of Islip in SD 3; the minority population in the greater part of the Town of Babylon-, which is now in SD 4; and the minority community in the western part of the Town Babylon, which is currently in SD 8, a district that bridges Nassau and Suffolk Counties. (Compare id. map 14c (Black and Hispanic Population and Senate Plan 2002), ivith, id. map 15 (Black and Hispanic Population and Revised Plaintiffs’ Plan).) Proposed District 4 would have a white VAP of 56.4%, a non-Hispanic black VAP of 17.5%, a Hispanic VAP of 22.7%, and a combined black and Hispanic VAP of 40.2%. See supra Table 1. The CVAP of blacks and Hispanics combined would be 33.7%. See id.

At the argument of the motions for summary judgment, the plaintiffs conceded that they could not satisfy the majority-in-a-district requirement. (See Summ. J. Tr. at 44.) The plaintiffs stated that their proposed district in Suffolk County would be an “influence district.” They admitted that the case for a violation was weaker for influence districts than for crossover districts under the VRA, but they argued that because there is evidence of intentional discrimination, the influence claim should be allowed. However, we find no basis for concluding that the creation of such a district is required, nor that the Senate Plan violated section 2 of the VRA because it failed to create a minority “influence” district. Moreover, the plaintiffs have failed to present any evidence to support a finding of intentional discrimination.

Although the Supreme Court has left the issue open, federal courts have nearly unanimously interpreted the first Gingles precondition strictly and have rejected any claim where the minority group does not constitute a majority of the relevant population in the proposed district. 38 The Courts of Appeals for the Sixth, Seventh, and Ninth Circuits, have expressly rejected “influence” claims while the Courts of Appeals for the Fifth and Eleventh Circuits have similarly enforced a strict majority-in-a-district requirement. See Cousin v. Sundquist, 145 F.3d 818, 828 (6th Cir.1998) (announcing that section 2 violations cannot “consist of an impairment of the minority’s ability to influence the outcome of fhe election rather than determine it”); Romero v. City of Pomona, 883 F.2d 1418, 1424 & n. 7 (9th Cir.1989) (refusing to. reopen case decided prior to Gingles in order to consider ability-to-influence claim, and stating, “We are aware of no successful section 2 .voting rights claim ever made without a showing that the minority group was capable of a majority vote in a designated single district.”), overruled on other grounds, Townsend v. Holman Consulting Corp., 914 F.2d 1136 (9th Cir.1990); McNeil v. Springfield Park Dist., 851 F.2d 937, 947 (7th Cir.1988) (describing Gingles, majority-minority threshold as having “the apparent objective of sharpening section 2’s focus,” and refusing to consider claim that plaintiffs’ ability to influence, rather than win, elections was impaired); see also Valdespino v. Alamo Heights Indep. Sch. Dist., 168 F.3d 848, 851-53 (5th Cir.1999) (calling Gingles factors “a bright line test” requiring proof that minority group “constitute[s] more than 50% of the relevant population in their demonstration district,” and denying “ability to elect” claim for district with approximately 48% Hispanic CVAP); Negron v. City of Miami Beach, 113 F.3d 1563, 1569 (11th Cir.1997) (requiring that minority group constitute majority of citizen voting age population and rejecting districts where Hispanics had voting age majority but were 41 to 48% of CVAP). But see Vecinos De Barrio Uno v. City of Holyoke, 72 F.3d 973, 979 n. 2, 990-91 (1st Cir.1995) (noting that first Gingles factor may need to be reconfigured for influence claims and directing district court to address plaintiffs’ influence-dilution claim because “potential influence is relevant to a determination of whether the [minority] group lacks a meaningful opportunity to participate in the electoral system”).

Several district courts have also- found that Gingles prohibits influence-dilution claims. See, e.g., DeBaca v. County of San Diego, 794 F.Supp. 990, 996-97 (S.D.Cal.1992) (rejecting influence claims), aff'd, 5 F.3d 535, 1993 WL 379838 (9th Cir.1993); Turner v. Arkansas, 784 F.Supp. 553, 568-72 (E.D.Ark.1991) (three-judge court) (citing McNeil, 851 F.2d at 947, and refusing to accept “ability to influence” claim because doing so would undermine Gingles preconditions), aff'd mem., 504 U.S. 952, 112 S.Ct. 2296, 119 L.Ed.2d 220 (1992); Hastert v. State Bd. of Elec., 777 F.Supp. 634, 652-54 (N.D.Ill.1991) (three-judge court) (emphasizing utility of objective rule and refusing to accept plaintiffs’ argument for “adoption of a definition of minority influence district that has no statistical bounds as to voter group size”); cf. Fund for Accurate & Informed Representation, Inc. v. Weprin, 796 F.Supp. 662, 672 (N.D.N.Y.) (three-judge .court) (declining to resolve issue, because plaintiffs could not meet burden of proof on cohesion and racial bloc voting preconditions), aff'd mem., 506 U.S. 1017, 113 S.Ct. 650, 121 L.Ed.2d 577 (1992).

We agree with the nearly universal opinion of federal courts that section 2 of the VRA does not require the creation of influence districts where minority voters will not be able to elect candidates of choice. Section 2(b) of the VRA protects the ability of minority groups “to participate in the political process and to elect representatives of their choice.” VRA § 2(b), 42 U.S.C. § 1973(b) (emphasis added). While Gingles left open the possibility of influence claims, Justice Brennan’s plurality opinion emphasized: “Unless minority voters possess the potential to elect representatives- in the absence of the challenged structure or practice, they cannot claim to have been injured by that structure or practice.” Gingles, 478 U.S. at 50 n. 17, 106 S.Ct. 2752; see also id. at 48, 106 S.Ct. 2752 (requiring plaintiffs to “prove that the use of a [voting practice or procedure] operates to minimize or cancel out their ability to elect their preferred candidates”). If a minority population is too small to elect candidates of choice in a reconfigured district even with the assistance of reliable crossover voters, then it is the size of the population and not the voting practice or procedure that is preventing the minority group from electing representatives of their choice. Dilution of the ability to influence representatives is not an injury cognizable under section 2(b) of the VRA. See id. at 50 n. 17, 106 S.Ct. 2752.

In addition to being at odds with the structure of the statute, the plaintiffs’ influence claim, as presented, has no standards and would be judicially unmanageable. “Influence” cannot be clearly defined or statistically proved. See, e.g., Hastert, 777 F.Supp. at 652-54 (refusing to open “Pandora’s box” by accepting influence claims where influence is poorly defined and has no adjudicable standards). As Justice Souter recognized in De Grandy, “some dividing [of minorities] by district lines ... is virtually inevitable.” De Grandy, 512 U.S. at 1015, 114 S.Ct. 2647. Allowing influence claims would open the door for a legal challenge any time a minority population could be shifted to increase the minority population in a nearby district. It would open the door for cases like this one, where the plaintiffs are arguing that the defendants had an affirmative obligation to create a district that has never existed in order to unite all minority communities in a particular region to maximize the proportion of a minority in at least one district.

Moreover, this would be a particularly inopportune case to recognize influence claims under section 2 because there is no evidence that the failure to create a district such as Plaintiffs’ Proposed District 4 has diluted meaningful influence for minority voters in Suffolk County. While blacks and Hispanics combined would be 40.2% of the VAP and 33.7% of the CVAP in Plaintiffs’ Proposed District 4, see supra Table 1, the plaintiffs cannot prove that blacks and Hispanics vote cohesively in Suffolk County. 39 Even if the groups were cohesive, statistics offered by the plaintiffs’ own expert undermine any viable argument that blacks and Hispanics could form a meaningful and influential voting bloc.

The plaintiffs’ expert, Dr. Michael McDonald, examined Suffolk County State Senate elections from 1996-2002 and estimated the turnout rates for whites, blacks, and Hispanics in each election for the three districts constituting Plaintiffs’ Pro posed District 4. (See Netburn Decl. Exs. 5, 6 (raw data and summaries of analysis by Dr. McDonald).) Dr. McDonald estimated that whites turned out to vote for those elections at an average estimated rate of 43.0%. (See id. Ex. 6 (first Electability Analysis table).) The average estimated turnout rate for blacks was 12.0% and for Hispanics it was only 1.4%. (See id.) Based on these turnout rates and the demographics for Proposed District 4, whites would, on average, constitute an estimated 90.0% of the actual voters in the proposed district. Blacks would constitute an estimated 7.9% of voters, and Hispanics would constitute only 1.2% of voters. Combined, the minorities would be 9.1% of the voters in elections in the proposed district. (See Defs. SOF ¶ 175; Pis. SOF' Response ¶ 175 (disputing defendants’ calculations only insofar as they used average turnout rates over sixteen elections instead of maximum black and Hispanic turnout rates).) The minority groups would not be able to form a voting bloc large enough to have meaningful influence. This is particularly true because the Republican incumbents for SD 3 and 4 — one of which would likely run in the proposed district — have obtained an average of approximately 65% of the vote in the last several elections. (See Netburn Decl. Ex. 5 app. 3, at 1, 8, 16-17, 24-25 (reporting election results).)

The plaintiffs argue that the Gingles factors should be relaxed to allow the influence claim for Suffolk County because there was intentional discrimination in drawing the district lines. See Garza v. County of Los Angeles, 918 F.2d 763, 770-71 (9th Cir.1990) (relaxing Gingles conditions where there was evidence of intentional discrimination but stating that plaintiffs must still prove injury). Although the 1982 Amendments to the VRA directed courts to look at the effects of redistricting, the plaintiffs argue that Congress did not employ an effects test at the expense of prohibiting practices that were intentionally discriminatory.

Even if intentional discrimination were a basis for relaxing the Gingles preconditions, the plaintiffs have presented no evidence of intentional discrimination. The thrust of the plaintiffs’ argument is not that the drawing of the Senate district lines in Suffolk County was based on intentional discrimination. It is plain that with minor exceptions, the district lines did not change between the 1992 and 2002 redistricting plans. (Compare Netburn Decl. Ex. 1 map 14c, with id. map 15.) Indeed, the lines affecting the minority population have not changed since the 1980s district-ing plan. (Compare id. map 13c, with id. maps 14c, 15.) The plaintiffs’ argument, then, is that the defendants discriminated by refusing to change the district lines to unite the minority communities that have grown over time.

The plaintiffs rely again on the memo to Sen. Skelos from Mark Burgeson, head of the Majority Redistricting Office. (See Netburn Decl. Ex. 39 (Memorandum from Mark Burgeson to Sen. Skelos, dated July 20, 2001) (“Burgeson Memo”); Summ. J. Tr. at 45^46.) In that memo, discussed earlier with respect to Count I, Burgeson discussed various options for creating a new district that would bring the total number of Senate districts to 63. One option involved creating a majority-minority district by connecting minority areas in Elmont, which is on the west border Nassau County, through Brentwood in the Town of Islip in the middle of Suffolk County. (Burgeson Memo at 1.) This district, Burgeson noted, would be vulnerable to an attack for racial gerrymandering under Shaw v. Reno and would also not be useful for Republicans. Burgeson also stated that he received testimony that the minority areas in Hempstead (Nassau County) should be united, and he raised the option of a majority-minority district bridging Nassau and Queens Counties that could accomplish that goal. (Id. at 1-2.)

The memo, however, ultimately recommended against a 63-seat plan. Burgeson found that “the only reason to go to 63 is to strengthen the Long Island delegation by combining politically undesirable areas in the extra district.” (Id. at 1.) He explained that there is no other place where it would be possible to strengthen the surrounding districts solely by adding another district. Plainly, the political majority would have been strengthened in existing districts by removing voters not of that party from current districts and placing them in a new district. However,.Burge-son concluded that adding another district could have resulted in the loss of a Republican incumbent depending on how it was done, and it probably would not have provided an extra Republican seat. The memo therefore shows only that Burgeson was considering options that he later rejected, largely for reasons of politics and redistricting demographics.

The plaintiffs deduce an alleged intent to discriminate primarily from two phrases in the memo. (See Summ. J. Tr. at 38-42.) Burgeson refers to a possible district uniting minority populations in Nassau County as “carving out politically undesirable areas and at the same time demonstrating] sensitivity to testimony” received at. public hearings. (Burgeson Memo at 1.) He later commented that the majority-minority district stretching from the border of Nassau and Queens Counties into Suffolk County would not-include most minority areas in Hempstead, and thus those areas “would still need attention.” (Id. at 2.) It is clear from the memo that the districting was politically motivated and that Burgeson was conscious of the impact on minority groups. But consciousness of minority groups is not evidence of intentional discrimination. See Miller v. Johnson, 515 U.S. 900, 916, 115 S.Ct. 2475, 132 L.Ed.2d 762 (1995) (“Redistricting legislatures will ... almost always be aware of racial demographics; but it does not follow that race predominates in the redistricting process.”) Indeed, it would be impossible in the face of the Voting Rights Act not to be aware of how redistricting affects minority populations. Most importantly, the memo was directed to a consideration of whether the redistricting committee should have created an additional district under a 63-seat plan — a plan that was never adopted. The memo says nothing about whether the current district lines were drawn or maintained for a discriminatory purpose.

In sum, the plaintiffs have failed to establish that the current Senate district lines in Suffolk County violate section 2 of the VRA. While they have asserted that an “influence” district should be created, the VRA does not require such districts. Moreover, the proposed district is of such slight minority influence that Gingles, which left open the possibility, cannot be read to accommodate such a district. The plaintiffs argue that the Gingles preconditions should be relaxed in the face of intentional discrimination, but they have failed to present evidence creating a material issue of fact that there was such discrimination. Summary judgment was therefore granted dismissing the plaintiffs’ claim with respect to Suffolk County.

2. Nassau County (Count V)

The plaintiffs contend that in Nassau County, the Senate Plan violates section 2 of the VRA because it has “cracked” Nassau County’s black population across four Senate districts — SDs 6 through 9— thus diluting the vote of black voters. The plaintiffs initially attempted to argue that blacks and Hispanics should be treated as a single, cohesive minority group. But at trial the plaintiffs conceded that they could not prove that blacks and Hispanics voted cohesively.' {See Trial Tr. (“Tr.”) at 57-60.) The turnout rate for and number of Hispanic voters was too- low to make reliable statistical estimates for black and Hispanic cohesion. The plaintiffs thus re-framed their claim as based solely on the dilution of black voting strength. Plaintiffs’ Proposed District 8 was depicted as an “effective” black district with blacks constituting 34.5% of the VAP and 36.9% of the CVAP. See infra Table 2.

_Table 2: Demogeaphics of Plaintiffs’ PROPOSED DistRict 8 40 _

Nassau County: Black +

Proposed District 8 White_Black_Hispanic_Hispanic

VAP 40.7% 34.5% 19.8% 54.3%

(CVAP) (18.2%)(86.9%)(10.8%)(1.7.6%)

(a) The First Gingles Factor: Majority-m-a-District and the Potential to Elect

In Plaintiffs’ Proposed District 8, black voters would not be a majority of the VAP or CVAP. The plaintiffs thus cannot satisfy the literal requirements of the first Gingles precondition. See Gingles, 478 U.S. at 49, 106 S.Ct. 2752 (“First, the minority group must be able to demonstrate that it is sufficiently large and geographically compact to constitute a majority in a single-member district.” (emphasis added)).. 41 The plaintiffs argue that their claim should be cognizable because black voters in their proposed district would be able to form an effective coalition district where they would have “the potential to elect representatives” of their choice with crossover votes from the white majority. Id. at 50 n. 17, 106 S.Ct. 2752. The defendants urge the Court to reject the plaintiffs’ Nassau County claim at the threshold because it is now clear that the plaintiffs cannot prove that there is a cohesive minority population that is a majority in the proposed district.

As explained earlier, the Supreme Court has declined to rule on whether “coalition/crossover” or “influence” claims are cognizable under the Voting Rights Act. See, e.g., Growe, 507 U.S. at 41 n. 5, 113 S.Ct. 1075; Gingles, 478 U.S. at 46 n. 12, 106 S.Ct. 2752. Unlike an influence claim such as the one presented for Suffolk County, a crossover claim at least fits the general structure of section 2(b) of the VRA: It alleges that but for the voting standard or practice, the members of the minority group would have the ability to “elect” representatives of their choice. See VRA § 2(b), 42 U.S.C. § 1973(b); Gingles, 478 U.S. at 48, 106 S.Ct. 2752 (explaining that section 2 requires minority members to “prove that the use of a [challenged] electoral structure operates to minimize or cancel out their ability to elect their preferred candidates”). In Voino vich, the Supreme Court noted that if a crossover claim were allowed, the first Gingles factor would have to be modified or eliminated to determine whether the voting practice deprived the protected group of “the possibility of being a sufficiently large minority to elect their candidate of choice with the assistance of crossover votes from the white majority.” Voinovich, 507 U.S. at 158, 113 S.Ct. 1149 (“[T]he Gingles factors cannot be applied mechanically and without regard to the nature of the claim.”); see also De Grandy, 512 U.S. at 1008, 114 S.Ct. 2647 (referring to first Gingles condition as requiring “a sufficiently large minority population to elect candidates of its choice”). The plaintiffs thus argue that their vote dilution claim should proceed because blacks in Proposed District 8 would allegedly have the potential to elect candidates of choice.

Just as with “ability to influence” claims, courts have nearly unanimously rejected “ability to elect” claims and have continued to require that the minority group constitute a majority of voters in a sufficiently compact district. See, e.g., Hall, 276 F.Supp.2d at 536, 538 (collecting cases and noting that plaintiffs’ coalition claim was distinct from an influence claim, but maintaining “the well-established and objective rule requiring a majority-minority district” under the first Gingles precondition); see also Valdespino, 168 F.3d at 852-53 (rejecting plaintiffs’ argument that to satisfy first Gingles condition they need to show only a potential to elect); Parker v. Ohio, 263 F.Supp.2d 1100, 1104 (S.D.Ohio 2003) (three-judge court) (dismissing “influence” claim where minority group was alleged to be large enough “to effectively influence elections, getting their candidate of choice elected”), aff'd mem., — U.S. —, 124 S.Ct. 574, 157 L.Ed.2d 426, 2003 WL 22171264 (2003). Only two federal courts to address the issue specifically have taken a “functional” approach allowing “ability to elect” claims, and the validity of both decisions is in serious doubt. See Metts v. Murphy, 347 F.3d 346 (1st Cir.), opinion withdrawn and reh’g en banc granted, No. 02-2204, 2003 WL 22434637 (1st Cir. Dec. 3, 2003), 2003 U.S.App. LEXIS 24313; Armour v. Ohio, 775 F.Supp. 1044, 1050-52, 1059 & n. 19 (N.D.Ohio 1991) (three-judge court) (allowing claim where blacks had ability to elect candidates of choice, but basing its decision in part on finding that Gingles did not apply to single-member districts); 42 see also Martinez v. Bush, 234 F.Supp.2d 1275, 1320 n. 56, 1320-23 (S.D.Fla.2002) (three-judge court) (doubting that first Gingles factor was intended to be “literal, mathematical requirement”). 43

Courts adopting the majority-in-a-district requirement have emphasized the need for a bright-line rule to act as a gatekeeper. See, e.g., Valdespino, 168 F.3d at 852; Hall, 276 F.Supp.2d at 538 (stating that “objective rule” ensures that judges are not required to make “subjective estimated] of minority voter influence”). As the Court of Appeals for the Seventh Circuit explained, the majority-in-a-district requirement provides “a choice of clear rules over muddy efforts to discern equity,” and it “reins in the almost unbridled discretion that section 2 gives the courts.” McNeil, 851 F.2d at 942. The bright-line rule effectuates- the judicial duty to enforce, voting rights while at the same time recognizing, as the Supreme Court instructed, that “the Constitution leaves with the States primary responsibility for apportionment of their federal congressional and state legislative districts.” Growe, 507 U.S. at 34, 113 S.Ct. 1075. Without a majority-minority threshold, any redistricting scheme could be challenged whenever disparate minority communities could be pooled together in sufficient numbers to create some’ potential to elect. See Parker, 263 F.Supp.2d at 1108 (Graham, J., concurring); see also McNeil, 851 F.2d at 942 (stating that Gingles “shields the courts from meritless claims and ensures that clearly meritorious claims will survive summary judgment”); Hastert, 777 F.Supp. at 654.

Both sides in this case have argued that the Supreme Court’s decision in Georgia v. Ashcroft, 539 U.S. 461, 123 S.Ct. 2498, 156 L.Ed.2d 428 (2003), proves their point. In Ashcroft, the Supreme Court found that Georgia’s redistricting plan did not violate section 5 of the VRA even though it broke up safe majority-minority districts into districts where minority voters would have to form coalitions in order to elect candidates of their choice. The plaintiffs argue that the decision shows that such coalition districts are viable under the VRA. Moreover, the plaintiffs argue, as Ashcroft recognized, such districts reflect the fact that “minority voters are not immune from the obligation to pull, haul, and trade to find common political ground, the virtue of which is not to be slighted in applying a statute meant to hasten the waning of racism in American politics.” De Grandy, 512 U.S. at 1020, 114 S.Ct. 2647, quoted by Ashcroft, — U.S. at —, 123 S.Ct. at 2512.

However, Ashcroft was a section 5 case, and the basis for the Court’s decision was that “[s]ection 5 gives States the flexibility to choose one theory of effective representation over the other.” Ashcroft, — U.S. at —, 123 S.Ct. at 2512. That is, states have the flexibility to choose between safe majority-minority districts, which guarantee minority representation but tend to isolate minorities by removing them from the remaining areas, and “coalitional” districts, which provide less security but enable minority groups to establish a broader political base and potentially affect more elections. See id.; id. at 2512-13 (O’Connor, J., concurring). Ashcroft recognizes that, at times, voting may not necessarily be so polarized so as to require majority-minority districts. While Ashcroft allows crossover districts under section 5, its reasoning does not broaden the power of federal courts under section 2 of the VRA to require state legislatures to protect or create such “ability to elect” districts. See Session v. Perry, 298 F.Supp.2d 451, 480-81 (E.D.Tex.2004) (three-judge court).

The Gingles opinions are directed at a key question in section 2 claims that is left unanswered by the statute: If a section 2 claim is that minority voting strength is being diluted, “how should undiluted minority strength be measured?” Gingles, 478 U.S. at 88, 106 S.Ct. 2752 (O’Connor, J., concurring). Gingles generally set that norm in terms of the ability to elect candidates of choice in the absence of the challenged structure or practice. Id. at 50, n. 17, 106 S.Ct. 2752 (Brennan, J.). More specifically, the first Gingles precondition set the baseline as when the minority group is sufficiently numerous and com pact to form a majority in a district. See id. at 90-91, 106 S.Ct. 2752 (O’Connor, J., concurring); Hastert, 777 F.Supp. at 653 (explaining that first Gingles factor provides “an objective brightline measure for determining whether an injury has in fact occurred”). In this case, the plaintiffs presented Proposed District 8 as the baseline for measuring vote dilution. With a YAP of 34.5%, the plaintiffs struggled to present evidence that with the right combination of turnout and crossover support, blacks in Plaintiffs’ Proposed District 8 would have the potential to elect representatives of their choice. The plaintiffs needed to prove electability not just to show that there was an available remedy, but also to prove that there was an injury, namely, the denial of a cohesive minority’s ability to elect a candidate of its choice.

A more complete description of Plaintiffs’ Proposed District 8 and of its derivation illustrates the hazards of deviating from a requirement of the majority-in-a-district benchmark, and it raises questions about whether Proposed District 8 should qualify as the baseline measure for “undiluted” black voting strength in Nassau County. The theory of the vote-dilution claim in Nassau County identifies cracking as the unlawful voting practice. At the core of the plaintiffs’ claim is the contention that Senate redistricting plans have fragmented the black community into four Senate districts, thus dividing and diluting black voting strength. As a descriptive matter, the black communities in Hempstead, Lakeview, Uniondale, Roosevelt, Freeport, and Baldwin have been bisected by the line dividing SDs 6 and 8. Cf. Goosby v. Town Bd., 180 F.3d 476, 484 (2d Cir.1999) (stating that black population in Town of Hempstead has been concentrated in those six communities). This is evident from the maps showing the districts and demographics in Nassau County from the 1970s through the current plan. (See PX 1 maps 11,12a, 13a, 14a.)

However, the black population in SDs 6 and 8 alone is not large enough to create an effective Senate district. While SDs 6 and 8 account for around 70% of the black population in the plaintiffs’ proposed district, there are only about 76,000 members of the black community in the two districts combined. (See DX 171.) Under the enacted and proposed districting schemes, Long Island Senate districts include about 306,000 people. (Id.) If the plaintiffs presented a district that united the black voters from Senate Districts 6 and 8 but then drew the rest of its population from predominantly white neighborhoods, the black voting age population would be closer to 25%. In order to reach a 34.5% VAP, the plaintiffs included Westbury and Elmont from SD 7, as well as the areas of North Valley Stream and Valley Stream from SD 9.

On its face, there has not been a history of cracking a minority population across four Senate districts, and traditional redistricting principles do not demand or indicate that the minority communities in all of those areas should be united in one district. Westbury has had a substantial minority population since the 1970s. (See PX 1 map 11.) It is located on the southern border of North Hempstead Town, and for the past thirty years it has been included in SD 7, which has encompassed all of North Hempstead Town in every districting plan in evidence. (See id. maps 11, 12A, 13A, 14A; Tr. at 308 (Beveridge).) Elmont, on the border of Queens County, is in the Town of Hempstead and is one of the few areas outside the Town of North Hempstead that the legislature has placed in SD 7. (See PX 1 maps 11, 12A, 13A, 14A.) While Elmont currently has a significant minority population, it was almost entirely white as of 1970, when it was included in SD 7. (See id. map 11; Tr. at 308 (Beveridge).) Likewise, North Valley Stream and Valley Stream, which have been part of SD 9 since the 1980s redistricting cycle, have a growing minority population that only began to emerge in the 1990s. (See PX 1 maps 12A, 13A, 14A.) It is not reasonable to claim that but for an illicit districting practice, the black communities in Westbury, Elmont, North Valley Stream, and Valley Stream would be joined with the black communities in and around Hempstead, Roosevelt, and Freeport.

Rather, the plaintiffs have gathered together in one proposed district substantially all of the black population of Nassau County. (See PX 6 tbl. 16 (showing that black population would not be more than three percent in Plaintiffs’ Proposed Districts 6, 7, and 9); Tr. at 308-09 (Bever-idge).) Even so, the plaintiffs still proffer a district with only 34.5% black VAP, which, as will be seen, is not sufficient to show that black voters have the potential to elect candidates of choice. This is not the use of section 2 of the Voting Rights Act to remedy a prohibited practice of vote dilution but rather a proposed use of section 2 to maximize minority voting strength in Nassau County. 44 As the Supreme Court succinctly stated in De Grandy, “Failure to maximize cannot be the measure of § 2.” De Grandy, 512 U.S. at 1017, 114 S.Ct. 2647.

In any event, assuming arguendo, as the Supreme Court has done, that the first Gingles precondition could be satisfied without the minority group constituting a majority in a compact district, the plaintiffs would still be required to prove as part of the first Gingles precondition that the minority group is sufficiently numerous to elect representatives of their choice in that district. The plaintiffs have alleged that blacks in their proposed district would fulfill this requirement. See Voinovich, 507 U.S. at 158, 113 S.Ct. 1149 (observing that “the first Gingles precondition ... would have to be modified” to accommodate crossover claims). When Growe assumed that minority-coalition claims were feasible, it found a “higher-than-usual need for the second of the Gingles showings,” and it noted that proof of cohesion across minority groups was “all the more essential.” Gr owe, 507 U.S. at 41, 113 S.Ct. 1075. This is because “unless the Gingles preconditions are satisfied, there can be neither an injury nor a remedy.” Id. at 40-41, 106 S.Ct. 2752. In this case, to demonstrate an injury based on alleged cracking, if the plaintiffs were to be allowed to proceed without meeting the majority-in-a-district threshold, they would nevertheless have to prove that blacks are sufficiently numerous in their proposed district to elect candidates of choice.

In order to determine whether Plaintiffs’ Proposed District 8 fulfills a modified first Gingles precondition requiring a “potential to elect,” we must first make findings of fact with respect to the turnout rates and voting patterns of whites and blacks. Because these findings address the issue of minority cohesion and racial polarization in Nassau County voting, we will also address the second and third Gingles factors in the context of those findings. Having done so, we will then consider the plaintiffs’ arguments on electability.

(b) Racial Polarization and the Second and Third Gingles Preconditions

The second Gingles precondition requires the plaintiffs to show that the minority group is politically cohesive while to satisfy the third Gingles precondition, the plaintiffs must demonstrate that “the white majority votes sufficiently as a bloc to enable it ... usually to defeat the minority’s preferred candidate.” Gingles, 478 U.S. at 51, 106 S.Ct. 2752; see Goosby v. Town Bd., 180 F.3d 476, 491 (2d Cir.1999). The second and third factors are thus related under the concept of racial polarization. To show that white bloc voting usually defeats the minority-preferred candidate, it is usually necessary to show that minority voters are sufficiently cohesive in preferring certain candidates. It is then necessary to show that whites similarly vote as a bloc, and in doing so, they usually defeat the minority candidate of choice.

To satisfy the second and third Gingles preconditions, the plaintiffs relied upon the expert testimony of Dr. Michael McDonald, an assistant professor of government and politics at George Mason University whose research focuses on voting behavior and research methodology. (Decl. of Dr. Michael McDonald (“McDonald Deck”) ¶ 1.) In assessing voting patterns in this case, Dr. McDonald relied primarily on the Ecological Inference (“El”) method developed by Dr. Gary King and supplemented that analysis with bivariate ecological regression (“BER”), and extreme case analysis (also called homogenous precinct analysis). (Id. ¶ 3.) The El and BER methods are related ways for political scientists to make inferences about the voting behavior of individuals based on aggregate voting data. See Georgia v. Ashcroft, 195 F.Supp.2d 25, 69 (D.D.C.2002), vacated on other grounds, 539 U.S. 461, 123 S.Ct. 2498, 156 L.Ed.2d 428 (2003).

i) The Methodology of Dr. McDonald

Ecological inference generally refers to “the statistical process of drawing inferences about individuals from aggregate data.” See Bruce M. Clarke & Robert Timothy Reagan, Fed. Judicial Ctr., Redistricting Litigation: An Overview of Legal, Statistical, and Case-Management Issues 53 (2002) (hereinafter “FJC”); see id. at 53-64 (providing explanation of various statistical methods used to show racially polarized voting). Ecological regression is one method for drawing such inferences and has been called the “standard technique” most commonly used in voting rights cases. See Reed v. Town of Babylon, 914 F.Supp. 843, 851 (E.D.N.Y.1996); see also id. at 851-52 (explaining statistical methods for estimating voter behavior). Bivariate Ecological Regression (“BER”) was used by Dr. Benjamin Grofman in Gingles. See Gingles, 478 U.S. at 52-53 & n. 20, 106 S.Ct. 2752.

While Dr. McDonald did ecological regression analyses, he primarily relied on a different Ecological Inference method developed in 1997 by Dr. Gary King. See FJC at 58-63. Dr. McDonald referred to Dr. King’s method simply as “El.” El has not been tested much in litigation, but it offers the prospect of improving on ecological regression in analyzing racial polarization. In Ashcroft, the district court relied on El calculations and accepted that El “was generally .considered to be the most accurate method of calculation.” See Ashcroft, 195 F.Supp.2d at 70 n. 36. The defendants presented the expert testimony of Dr. Harold Stanley, a Professor of American Politics and Political Economy at Southern Methodist University. (Stanley Aff. ¶ 1.) Dr. Stanley, asserted that there is an ongoing debate about the reliability of Dr. King’s El methodology, which requires numerous underlying assumptions. (Id. ¶¶ 32-35.) He also questioned the reliability of some of Dr. McDonald’s estimates, especially with respect to the estimates for Hispanics and the analyses of black-Hispanic cohesion. (See id. ¶¶ 14-17.) Because the plaintiffs no longer rely on black-Hispanic cohesion and estimates of Hispanic voting patterns in their Nassau County claim, those criticisms are not material for this count. But see infra note 118 and accompanying text.

However, Dr. Stanley reviewed Dr. McDonald’s work and found that the analyses were properly performed. The defendants have not submitted conflicting statistical analyses and have instead advanced alternative interpretations of the evidence provided by Dr. McDonald. Dr. Lisa Handley, another expert for the plaintiffs, confirmed that Dr. McDonald’s analyses were performed correctly. (Deck of Dr. Lisa Handley (“Handley Deck”) ¶ 3.) We will therefore accept the statistical analyses of the election results to which all parties have agreed.

Dr. McDonald’s analysis was directed at identifying racial polarization, which is loosely defined as when whites and the relevant minority group have different preferred candidates or “candidates of choice.” A minority-preferred candidate generally is the candidate that receives the most votes from the relevant minority group. Dr. McDonald, however, adopted a more rigorous standard for identifying elections as racially polarized based on his standard for judging whether the relevant minority group is “politically cohesive.” See Gingles, 478 U.S. at 51, 106 S.Ct. 2752 (second Gingles precondition). Dr. McDonald defined cohesion as occurring in an election when 60% or more of the voters of a group vote for the same candidate. (McDonald Deck ¶ 8.) In adopting the 60% threshold for cohesion, Dr. McDonald defined racial bloc voting as occurring where at least 60% of a minority group’s voters prefer one candidate and at least 60% of white voters prefer a different candidate. For purposes of the Nassau County claim, Dr. McDonald characterized an election as racially polarized only where blacks and whites cohesively supported different candidates of choice. 45

ii) Elections Analyzed

To determine whether racial polarization occurred in Nassau County, Dr. McDonald analyzed sixteen State Senate elections: the 1996, 1998, 2000, and 2002 elections for SDs 6, 7, 8, and 9. Those elections are referred to as “endogenous” because they occurred in the jurisdiction at issue and for the office at issue. Of those sixteen endogenous elections, four involved minority candidates. “Exogenous” elections cover the jurisdictions or geographic areas at issue but are statewide or local elections for offices not at issue. See Cano v. Davis, 211 F.Supp.2d 1208, 1235 n. 29 (C.D.Cal.2002) (providing definition of endogenous and exogenous elections). As acknowledged by all experts in the case, endogenous elections are the most probative elections for determining whether minority groups are politically cohesive and for determining whether racial bloc voting by whites usually defeats the minority-preferred candidate for the office in issue. (See McDonald Decl. ¶ 5.) While Dr. McDonald noted that endogenous elections featuring a minority candidate are the most probative, he testified that endogenous elections without minority candidates are also important and cannot be ignored. (Id. ¶ 6.) The Court of Appeals for the Second Circuit has confirmed that in addition to considering cases where white bloc voting defeats a minority candidate, courts must also consider elections where white bloc voting defeats a white candidate cohesively supported by the relevant minority group. See NAACP v. City of Niagara Falls, 65 F.3d 1002, 1015-16 (2d Cir.1995).

In addition to the sixteen endogenous elections, Dr. McDonald analyzed ten exogenous statewide general elections and reported estimated voting statistics for each of the four Senate districts for each statewide election. For Nassau County, there were no primary elections for State Senate during the period studied. Dr. McDonald did analyze the exogenous 1998 Democratic primary for Lieutenant Governor because a minority candidate was involved. He also reported findings for the 1988 Democratic presidential primary, which included Jesse Jackson. Neither primary election is probative for the purposes of the Nassau County claim. 46

In summary form, the cohesion and white bloc voting results for the sixteen endogenous elections analyzed are as follows:

Table 3: Sixteen Endogenous Elections in Nassau County 47

Black White Bloc Result Source

Election 48 Candidates 49 Cohesion Voting for BPC 50 Page 51

2002 SD 6 Hannon/St. George * Yes Yes Defeated C<3

2002 SD 7 Balbonl/Murray * Yes Yes Defeated CO

2002 SD 8 Fuschilllo/Dash * Yes Yes Defeated ^

2002 SD 9 Skelos/Glassberg * Yes Yes Defeated lO

2000 SD 6 Hannon/Keefe * Yes Yes Defeated 9

2000 SD 7 Balboni/Brooks * Yes Yes Defeated 10

2000 SD 8 Fuschillo/Santos * Yes Yes Defeated 11

2000 SD 9 Skelos/Oppenheiraer Yes Yes Defeated 12

1998 SD 6 Hannon/Macagone * Yes Yes Defeated OO rH

1998 SD 7 Balboni/Brooks * Yes Yes Defeated 05 t — I

1998 SD 8 Fuschillo/Hagan * Yes Yes Defeated O <N]

1998 SD 9 Skelos/Waterman * Yes Yes Defeated 1 — I <M

1996 SD 6 Hannon/Langberg * Yes Yes Defeated 26

1996 SD 7 Tully/Nachbar * Yes Yes Defeated 27

1996 SD 8 Levy/McNight * Yes Yes Defeated 28

1996 SD 9 Skelos/Miller * Yes Yes Defeated 29

In the sixteen endogenous State Senate elections, whites voted as a bloc (60% or more in favor of the same candidate) in all sixteen elections and in all sixteen elections the black-preferred candidate was defeated. See supra Table 3. On average, an estimated 94.2% of blacks voted for the black-preferred candidate. (See DX 149.) Dr. McDonald also examined the rate of white crossover support and found, on average, that an estimated 30.4% of white voters supported the black-preferred candidate. (See id.) Dr. McDonald thus deduced that whites voted as a bloc at an average rate of 69.6% against the black-preferred candidate.

Of the sixteen State Senate elections, four included minority candidates. 52 The results for those four elections were as follows:

Table 4: ENDqgenous Elections with MINORITY Candidates

Election Candidates/Results Black Support for Minority Candidate White Support for Minority Candidate

2002 SD 6 Hannon (R) (White) 61.1% St. George (D) (Black) 35.1% 96.1% 23.9%

2002 SD 8 Fuschillo (R) (White) 67.9% Dash (D) (Black) 32.1% 94.0% 21.1%

2000 SD 8 Fusehillo (R) (White) 57.4% _Santos (D) (Hispanic) 42.6%_95.7%_31.5%

1996 SD 8 Levy (R) (White) 66.2% McNight (D) (Black) 33.5%97.4%24.7%

(See DX 147.) In these four elections, whites voted as a bloc against the minority candidate at an average estimated rate of 74.7%. However, it should be noted that the 2002 elections occurred after this litigation was begun, and there is some evidence that those candidates ran in part to establish a litigation track record. (Tr. at 79-80); see Gingles, 478 U.S. at 76, 106 S.Ct. 2752 (noting that elections conducted after lawsuit was filed may involve special considerations).

Voting patterns in the ten exogenous statewide general elections in Nassau County were more mixed. They are as follows:

Table 5: Ten Statewide General Elections 53

Election Candidates Black Cohesion 54 White Bloc Voting Result for BPC 55 Source Page

2002 Governor Pataki/McCall* 56 Yes Yes "Defeated" 55-58

2002 Att’y General Spitzer* / Irrizary Yes No "Won" 63-66

2002 Comptroller Faso / Hevesi* Yes 2/4 Yes (SD 6, 8) 57 Defeated 86-89

2000 U.S. President Bush / Gore* Yes No Won 94-97

2000 U.S. Senate Clinton* / Lazio Yes Yes Defeated 102-05

1998 U.S. Senate Schumer* / D’Amato Yes 2/4 Yes (SD 6, 8) Defeated 110-13

1998 Governor Pataki / Vallone* Yes Yes Defeated 117-20

1998 Att’y General Spitzer* / Vacco Yes 2/4 Yes (SD 6, 8) 58 Defeated 125-28

1998 Comptroller McCall* / Blakeman Yes No Won 71-74

1996 U.S. President Clinton* / Dole Yes No Won 133-136

Six out of the ten statewide exogenous elections, therefore, showed signs of racial polarization. In three of those six elections, only two out of four Senate districts were racially polarized. However, in all three elections, the two polarized districts were SDs 6 and 8, which, when combined, constitute around 70% of the black population in Proposed District 8. (See DX 171.)

The exogenous elections with minority candidates were limited in number and produced ambiguous results. For example, Carl McCall, who is black, easily won the election for Comptroller in 1998, and there was no racial polarization. But in 2002 when he ran against the incumbent for Governor, the top leadership position in the state, he was resoundingly defeated in an election with the highest degree of white bloc voting recorded for all of the Nassau County contests. (See DX 151 (showing that white voters in the four Nassau Senate districts supported McCall at rates of 14.2%, 20.2%, 15.1%, and 18.7%).) The 2002 Attorney General election matched Spitzer, a white Democrat against Irizarry, a Hispanic Republican challenger. While there was no white bloc voting in that race, a statewide election with a Hispanic Republican says little about Nassau County voters’ reactions to black and black-preferred candidates in State Senate elections. 59

iii) The Second Gingles Precondition: Cohesion

With respect to the second Gingles factor, the plaintiffs conceded during the trial that they could not prove cohesion between blacks and Hispanics. (See Tr. at 57-60.) It is undisputed, however, that blacks vote cohesively in Nassau County. Blacks voted cohesively for the same candidate in all sixteen State Senate elections analyzed. See supra Table 3. On average, in those elections an estimated 94.2% of blacks backed the same candidate. (DX 149.) The pattern of black cohesion in exogenous elections was equally clear. See supra Table 5.

iv) The Third Gingles Precondition: White Bloc Voting Sufficient Usually to Defeat the Black Candidate of Choice

All parties agreed that endogenous elections are the most probative, and the pattern in the endogenous elections is clear. As a purely descriptive matter, in every Nassau County State Senate election over the past eight years, blacks have voted cohesively as a bloc for one candidate, whites have voted cohesively as a bloc for the other candidate, and the black-preferred candidate has lost. (See supra Table 2.) It is true, however, that the results of the sixteen endogenous State Senate elections can be described equally by the fact that in each of the State Senate elections, the Republican candidate, who was almost always the incumbent, won the election, and the majority of black voters supported the losing Democratic challenger.

The defendants’ expert, Dr. Stanley, did not challenge the voting patterns but argued that the cause of the voting patterns is not race. Dr. Stanley claimed that “Dr. McDonald’s analyses establish that any divergence between white and minority voters ... is better correlated with partisanship than race.” (Stanley Aff. ¶ 39.) The plaintiffs’ evidence, Dr. Stanley contends, “essentially consists of the fact that whites in Senate elections supported well-entrenched Republican incumbents who defeated all challengers, white or minority, and the fact that Governor Pataki, another popular incumbent who was easily reelected, defeated Carl McCall in the 2002 gubernatorial election. The Senate election contests reveal no race-based disadvantage since the incumbents defeated all opponents of any race or ethnicity.” {Id. ¶ 45.)

The thrust of the defendants’ argument is that the plaintiffs have not established the third Gingles precondition because they have not shown that the voting behavior in the four Senate districts in Nassau County is correlated with race rather than partisanship. They argue that the voting patterns in the Senate districts simply reflect partisan voting with blacks supporting the Democratic challenger who consistently loses to the Republican, usually incumbent, candidate. They essentially argue that it is the plaintiffs’ obligation to establish the three Gingles preconditions and that the third condition requires not simply white bloc voting that usually defeats the minority preferred candidate, but “legally significant” white bloc voting correlated with race rather than partisanship. See Gingles, 478 U.S. at 56, 106 S.Ct. 2752 (“[T]he question of whether a given district experiences legally significant racially polarized voting requires discrete inquiries into minority and white voting practices .... And, in general, a white bloc vote that normally will defeat the combined strength of minority support plus white ‘crossover’ votes rises to the level of legally significant white bloc voting.”) The plaintiffs respond that the third Gingles precondition does not require the plaintiffs to exclude partisanship as a cause of the voting patterns at issue and that inquiries into partisanship should be considered as part of the totality of the circumstances inquiry after the Gingles preconditions have been satisfied.

The Court of Appeals for the Second Circuit considered the issue of partisanship in connection with the Gingles factors in Goosby v. Town Board, 180 F.3d at 491. In Goosby, the plaintiffs challenged the practice of using at-large voting to elect the members of the Town Board of Hemp-stead, and the dispute centered on whether legally significant racial bloc voting usually defeated the minority-preferred candidate. Dr. Andrew Beveridge and Dr. McDonald testified for the plaintiffs. Id. at 488-89. On behalf of the defendants, Dr. Stanley testified that the patterns of racial polarization were better explained by partisanship than racial animus: a majority of voters in Hempstead were white Republicans while blacks overwhelmingly voted for the Democrat who lost on partisan grounds. Id.

In Goosby, the Court of Appeals accepted the plaintiffs’ arguments and agreed with the Court of Appeals for the Fourth Circuit that it was necessary.“to consider the political partisanship, argument under the ‘totality of the circumstances’ analysis rather than as part of the third Gingles precondition.” Goosby, 180 F.3d at 493; see Lewis v. Alamance County, 99 F.3d 600, 615 n. 12 (4th Cir.1996). This approach has also been adopted by the Courts of Appeals for the First, Seventh, and Eleventh Circuits, although it has been rejected by the Court of Appeals for the Fifth Circuit. Compare Milwaukee Branch of the NAACP v. Thompson, 116 F.3d 1194, 1199 (7th Cir.1997); Vecinos De Barrio Uno v. City of Holyoke, 72 F.3d 973, 983 (1st Cir.1995); and Nipper v. Smith, 39 F.3d 1494, 1524-25 (11th Cir. 1994); with League of United Latin Am. Citizens v. Clements (“LULAC"), 999 F.2d 831, 891-92 (5th Cir.1993) (en banc) (finding that plaintiffs had not satisfied third Gingles precondition because “partisan affiliation, not race, caused the defeat of the minority-preferred candidate”). The Court of Appeals for the Sixth Circuit in Clarke v. Cincinnati declined to consider whether the third Gingles precondition requires that “minority-preferred candidates’ lack of success is ‘somehow tied to race.’” Clarke v. Cincinnati, 40 F.3d 807, 813 n. 2 (6th Cir.1994) (quoting LULAC, 999 F.2d at 850). But Clarke did reject the contention that the third Gingles factor involves a “colorblind approach to the white bloc voting issue.” Id. at 812.

Goosby is, of course, controlling. The raw statistics establish that in the sixteen endogenous Senate elections, the black-preferred candidate was usually defeated and whites voted as a bloc for the prevailing candidate. See supra Table 3. Of those sixteen elections, four involved minority candidates, and in all four elections the minority candidate was defeated. See supra Table 4.

However, this case illustrates the danger of relaxing the first Gingles precondition and allowing plaintiffs to proceed on less than a showing that the minority at issue would be a majority in the proposed district. In Goosby, there was no question that blacks would constitute a majority in the new proposed district — 52.47% of the VAP. Goosby, 180 F.3d at 489. When that requirement is relaxed and when partisanship cannot be considered in connection with the third precondition, the third precondition can be effectively eviscerated. The smaller the racial minority is, providing that it supports the minority political party in a district, the more likely it becomes that the racial minority will always be outvoted by the majority, which is both white and consists of members of the majority political party in that district. Those results may say nothing about “legally significant” white bloc voting, but only support the unremarkable proposition that groups that cannot muster a majority of voters will lose an election because they have fewer votes.

Having recognized the problems with allowing crossover claims such as this one, we will nonetheless consider the plaintiffs’ claim that their proposed district would provide blacks in Nassau County an ability to elect candidates of choice.

(c) Electability

As the Supreme Court has explained, section 2 of the VRA “places at least the initial burden of proving an apportionment’s invalidity squarely on the plaintiffs shoulders.” Voinovich, 507 U.S. at 155, 113 S.Ct. 1149. The Court of Appeals for the Seventh Circuit warned that relaxing the bright-line majority-in-a-distriet rule would invite courts to “build castles in the air,” and it thus declared that proof of “the potential to elect must be solid and substantial. It cannot be speculative.” McNeil, 851 F.2d at 944; see id. at 943-44 (recognizing that various factors may affect potential to elect, including crossover voting, and enforcing majority-in-a-district rule to ensure “clarity and uniformity”). In order to proceed with their vote-dilution claim, the plaintiffs must prove that black voters have the ability to elect candidates of their choice in Plaintiffs’ Proposed District 8. One reason for this proof is that it is not the role of courts to intrude upon a state legislature’s responsibility in redistricting based on slight mathematical adjustments the calculation of which is largely beyond the expertise of judges and better suited to legislatures. See, e.g., Latino Political Action Comm. v. City of Boston, 784 F.2d 409, 412 (1st Cir.1986) (Breyer, J.) (stating that projec tions of future electoral results rely on “the very finest of political judgments about possibilities and effects — judgments well beyond [judicial] capacities”).

In addition, the plaintiffs must prove electability because electability is acting as a substitute for the first Gingles precondition, which provides a measure for whether there has been an injury. See Gingles, 478 U.S. at 48, 50 n. 17, 106 S.Ct. 2752 (stating that potential to elect “in the absence of the challenged practice” is necessary show vote-dilution). The plaintiffs’ arguments that the proposed district may be generally “competitive” or that the population will grow and in the future provide the ability to elect are beside the point. The purpose of looking^ into electability is not only to determine whether there is a possible remedy, but also to determine whether a legislative plan diluted a minority group’s voting strength and impaired its potential to elect candidates of choice. See Growe, 507 U.S. at 40-41, 113 S.Ct. 1075 (explaining that unless Gingles conditions are established, “there neither has been a wrong nor can be a remedy”).

To prove that Proposed District 8 is an effective black district, the plaintiffs relied largely on the expert testimony of Dr. McDonald, who performed “recompilation” and “supplemental electability” analyses. In addition, the plaintiffs have offered analyses by Dr. Handley and anecdotal evidence from witnesses such as Sen. David Paterson, the minority leader of the State Senate who represents a district in Manhattan. The defendants have challenged the plaintiffs evidence based primarily on the testimony of their expert, Dr. Stanley.

i) Recompilation

Recompilation analysis does not involve statistical estimates, complicated regression analyses, or margins of error. {See McDonald Decl. ¶ 62.) . It simply looks at the votes actually cast in the voting precincts in each Senate district and then recompiles a vote count for those precincts that would comprise the proposed district. The purpose of the method is to see whether a minority-preferred candidate who garnered a minority of votes in the current Senate districts would have “prevailed” if the districts were drawn according to the plaintiffs’ plan. 60

For example, the 2002 election for Governor matched George Pataki (R) against Carl McCall (D), who is black and was the black-preferred candidate. In the actual election, Pataki defeated McCall in SDs -6-9 by an average of 27)489 vote across the four districts. (PX 82.) 61 Dr. McDonald recompiled the votes cast in the portions of SDs 6-9 that comprise Plaintiffs’ Proposed District 8. He found that if Plaintiffs’ Proposed District 8 had existed, McCall would have received 32,496 votes in the hypothetical district, compared to Pataki’s 28,011 votes. (Id.) McCall, the black-preferred candidate, would thus have “won” in Plaintiffs’ Proposed District 8 by 4485 votes. The plaintiffs seek to extrapolate such ré- suits to argue that a hypothetical State Senate candidate like McCall would lose in any of the current districts but could nonetheless win the proposed district with minimal white crossover voting.

' There are two particularly relevant features of recompilation analysis generally and as employed by Dr. McDonald. First, the analysis can only be performed by looking at elections that span the entire jurisdiction of the proposed the district. Recompilation analysis thus depends strictly on exogenous elections and does not work with endogenous State Senate elections that are specific to any particular district. Dr. McDonald presented recompilation evidence for six statewide general elections and the 1988 presidential primary involving Jesse Jackson (although countywide elections could theoretically be used for recompilation analysis as well).

Second, Dr. McDonald only performed this analysis for elections that exhibited patterns of alleged racially polarized voting similar to the patterns found in State Senate elections — namely, elections where there was cohesive black support for one candidate and white bloc voting for the other candidate. The reason for doing this is to provide a better estimate of whether a black-preferred candidate could win a polarized State Senate race in the Plaintiffs’ Proposed District 8. If Dr. McDonald included exogenous elections where there was no white bloc voting and substantial white support for the black-preferred candidate, it would appear easier for blacks to elect their candidate of choice. In contrast, a highly polarized statewide election would provide a more rigorous test for whether the black community could elect their candidate of choice in the proposed district with minimal white crossover support. To make it clear: the recompilation electability analysis presumes racially polarized voting and is not evidence of racially polarized voting.

Dr. McDonald examined six racially polarized statewide elections from 1996-2002 (minority-preferred candidate named first): (1) the 2002 election for Governor with Carl McCall versus George Pataki; (2) the 2002 election for Comptroller with Alan Hevesi versus John Faso; (3) the 2000 election for United States Senate with Hillary Clinton versus Rick Lazio; (4) the 1998 election for United States Senate with Chuck Schumer versus Alfonse D’Amato; (5) the 1998 election for Governor with Peter Vallone versus George Pa-taki; and (6) the 1998 election for Attorney General with Eliot Spitzer versus Dennis Vacco. (PX 68, 82). In each election, the black-preferred candidate (also a Democrat) obtained fewer votes than the Republican candidate in each Nassau County Senate district. However, in five of the six elections, the black-preferred candidate would have received more votes than the Republican candidate in the Proposed District 8. 62 In the six elections, the black-preferred candidate received an estimated 13.4%, 31.3%, 30.2%, 29.2%, 14.8%, and 27.5% of the white vote, respectively. (PX 82.) White crossover vote for the sixteen State Senate elections analyzed for racial bloc voting averaged 30.4%. (See DX 149.)

The plaintiffs particularly emphasize the McCall/Pataki race for Governor in 2002 as the most rigorous test for electability in their proposed district. They consider it the most rigorous because only 13.4% of white voters in Proposed District 8 voted for McCall, but McCall would have received a plurality (49.7%) of the votes in the hypothetical district. (PX 69.) McCall’s plurality, therefore, is primarily attributable to cohesive support from the black community.

The recompilation method and the evidence presented at trial, however, have substantial flaws and, in the end, are unpersuasive. Recompilation relies on elections that, according to all parties, are not particularly probative. Although both parties have consistently emphasized that endogenous, minority-white elections are the most probative, Dr. McDonald’s recompilation analysis relies solely on statewide elections, most of which have no minority candidate. As Dr. McDonald himself explained, exogenous elections tend to lack probative value because elections “are going to be different across different levels of jurisdictions.” (Tr. at 187 (McDonald).) Such elections will involve different issues, types of candidates, and media attention, as well as different kinds of campaigning and fundraising. (Id.) Moreover, the exogenous elections — particularly high-profile races for Governor or United States Senate — do not account for the local effects of State Senate incumbency or possible “roll-off’ that occurs with lower profile elections that are further down the ballot. Furthermore, as Dr. Stanley pointed out, recompilation does not take into account important factors, such as the rates of turnout, crossover, and cohesion, which are used to generate Dr. McDonald’s supplemental electability analyses. (See Stanley Aff. ¶ 110). And the relative ease with which the minority-preferred candidates succeed in the plaintiffs’ recompiled elections is inconsistent with Dr. McDonald’s supplemental electability analysis and the anecdotal evidence presented at trial.

ii) Supplemental Electability Analysis

Dr. McDonald’s second, supplemental electability analysis involves estimating support for a minority preferred candidate based on the sixteen State Senate general elections in SDs 6-9 in the years 1996, 1998, 2000, and 2002. The method estimates support by looking at (1) the percent of the VAP that is black and white in the proposed district, (2) the estimated turnout rates for black and white voters in the State Senate elections, (3) the estimated degree of black support for black-preferred candidates (black cohesion), and (4) the estimated degree of white support for black-preferred candidates (white crossover). This electability analysis, however, fails to provide sufficient evidence to prove that Proposed District 8 is an electable district for black-preferred candidates both because the methodology is unreliable and because the results are ambiguous at best and unfavorable to the plaintiffs at worst.

Dr. McDonald’s data and supplemental electability analysis for the sixteen State Senate elections is summarized as follows:

Table 6: Data for the Supplemental Electability Analysis: Turnout and Crossover Rates in Sixteen Senate Elections 63

[[Image here]]

Even taken at face value, the supplemental electability analysis undermines the plaintiffs’ claim that the black community has the ability to elect a candidate of choice in the proposed district. For the sixteen State Senate elections, on average, an estimated 48.8% of the white voting age population turned out for the State Senate elections while only 17.3% of blacks turned out to vote in those elections. (DX 149.) The black-preferred candidate was supported by blacks at very high rates — 94.6% cohesion, on average — -and generated, on average, support from 30.4% of white vot ers. (Id.) When Dr. McDonald extrapolated those numbers for a district with a voting age population of 40.7% white and 34.5% black, he found that the black-preferred candidate would receive only 45.3% of the vote. 64 The white-preferred candidate, presumably the Republican incumbent, would receive 54.7% and handily win the election in the Plaintiffs’ Proposed District 8. (Id.) Dr. McDonald explicitly acknowledged that he is on record elsewhere testifying that 45.8% is not a “competitive district.” (Tr. at 105.)

The plaintiffs also failed to prove the reliability of Dr. McDonald’s electability analysis. Dr. McDonald admitted that he could not speak to how competitive the proposed district would be without performing a competitive race analysis, jvhich he did in a case in which he testified in Arizona. (Tr. at 105-07.) Such an analysis would have included the standard error for estimates of how partisan a district has to be in order for the minority-preferred candidate to win. As Dr. McDonald explained, the numbers for turnout, cohesion, and crossover support are “all based on statistical estimates, so there is going to be some degree of uncertainty to it.” (Tr. at 108.) The standard error, or margin of error, provides a range of election results and indicates with 95% confidence that any given election will fall within that range. Generally, a smaller standard makes for more rehable and informative estimates— the uncertainty is more contained. Dr. McDonald, however, did not produce confidence intervals and provide standard errors for his electability analysis, although he testified it would have been possible to do so. (Tr. at 107-09.)

Instead of performing a competitiveness analysis and providing a standard error, Dr. McDonald relied on what he called a “pseudo sort of analysis.” (Tr. at 107.) He created a “best case scenario” using the highest levels for turnout and support for the black-preferred candidate for whites, blacks, and Hispanics found in the sixteen Senate elections. Based on those numbers, he estimated that the black-preferred candidate would garner 54.5% of the vote in the plaintiffs’ proposed district. (DX 14.9.) Dr. McDonald testified that the best case scenario analysis “kind of is similar to what these 95 percent intervals are” in that 54.5% would represent the upper bound of the estimated vote for a black-preferred candidate. (Tr. at 108.)

Although blacks would be able to elect their candidate of choice under the “best case scenario,” the plaintiffs provided no evidence that this scenario was ever likely to occur, and the history of recent elections provides no support for its likelihood. For example, the best case scenario assumes 37.7% white crossover support for the black-preferred candidate — the highest level of white crossover found in the sixteen elections. (See DX 149.) The likely Republican candidate in the plaintiffs’ proposed district would be Sen. Hannon, the incumbent of the current SD 6, which constitutes much of the proposed district. (See Tr: at 538 (Paterson).) The white crossover rate in the 2002 election, in which Hannon ran against a black candidate, was 23.9%. (See DX 149.) The average crossover rate for the last four State Senate elections involving Hannon is 30.3% — very close to the average for all sixteen elections. Moreover, if Dr. McDonald were truly offering the “best case scenario” as an upper bound estimate, he should have provided a “worst case scenario” as a lower bound estimate. 65 He did not do so, and Dr. Stanley persuasively testified that there is no professionally acceptable way of knowing how much confidence one should place in Dr. McDonald’s results. (See Stanley Aff. ¶¶ 134; see generally id. ¶¶ 116-21.) The best case scenario thus reflects little more than the cherry-picking of favorable numbers that produce a favorable — but ultimately artificial — result.

The supplemental electability analysis for Plaintiffs’ Proposed District 8 has failed to prove that the proposed district would enable blacks to elect candidates of their choice. The reliability of the analysis is doubtful, and the plaintiffs simply failed to perform analyses that might have produced more probative results. Moreover, the plaintiffs’ own electability projections cast doubt on blacks’ ability to elect a senator of choice even in the proposed district.

in) Turnout, Warming, and Other Statistical and Anecdotal Evidence

The supplemental electability analysis points to the crucial role of the turnout rate in determining whether blacks will be able to elect candidates of choice while being less than a majority of the VAP. In Proposed District 8 blacks would constitute 34.5% of the VAP. 66 If the turnout rates remain at the levels estimated by Dr. McDonald, blacks would account for only 22.9% of the voters in State Senate elections. (See Stanley Aff. ¶ 122; see also DX 149 (giving turnout-rate data used by Dr. Stanley to calculate share of voters).) Whites would comprise 76.3% of State Senate voters in the plaintiffs’ proposed district. (Stanley Aff. ¶ 122.) With those numbers, any black-preferred candidate could be elected only by a coalition consisting of more white voters than black voters. (Id.)

The critical importance of turnout rates was also highlighted by Dr. Handley, the plaintiffs’ second electability expert. She analyzed voting patters in general elections to the Nassau County Legislature for Legislative Districts 1 and 2. 67 (See Handley Decl. ¶¶ 12-13.) Over 98% of the population in those districts would be in Plaintiffs’ Proposed District 8, and the two legislative districts account for about half of the population that would be included in the proposed district. (Id.) State Senate districts have substantially more population that Nassau County legislative districts. Dr. Handley determined that for the legislative districts in the areas she analyzed, a 34% black VAP would be sufficient to provide black voters with an opportunity to elect minority-preferred candidates. Dr. Handley recognized that in the larger area that Dr. McDonald analyzed, he found higher minority percentages required to elect minority-preferred candidates. (Id. ¶ 13.) 68

Dr. Handley identified the relative turnout rates of blacks and whites as an important reason for the different estimates. In the County legislative elections she analyzed, blacks have been successful in electing candidates of choice because blacks turn out for those off-year elections at rates roughly one-and-a-half times higher than whites. (Id. ¶ 14; Tr. at 148-49 (McDonald).) This turnout ratio was the basis for her conclusion that, in the districts she analyzed, blacks would be able to elect candidates of choice with only 34% of the VAP. According to Dr. Handley’s analysis, if whites and blacks turned out at equal rates, blacks would need to constitute 42% of the VAP. (Tr. at 149 (McDonald).)

Under the estimates calculated by Dr. McDonald, blacks turned out for State Senate elections at an average rate of 17.3% while whites turned out at an average rate of 48.8% — nearly three times higher than blacks. (DX 149.) While Dr. Handley’s testimony supports the theory that black turnout will increase if they can elect candidates of choice, her conclusions show that black turnout would need to increase dramatically in order for Plaintiffs’ Proposed District 8 to be an “effective” black district. Indeed, the plaintiffs recognize the need to account for the gap in white and black turnout. They argue that the allegedly unfair districting schemes have suppressed voter participation for State Senate elections and that uniting the black community would spark an increase in interest and greater turnout. Because of these factors, they contend, the supplemental electability analysis actually underestimates black voting strength, and the Court should understand the analysis as showing generally that with increased turnout, the proposed district would be effective.

In arguing that black turnout would increase in the proposed district, the plaintiffs introduced testimony by Dr. McDonald and Dr. Handley on an alleged “warming effect.” The warming effect is a sociological phenomenon where creating a competitive district with a solidified minority community promotes participation in the political process and raises turnout rates. (See McDonald Decl. ¶¶ 79, 95-97; Handley Decl. ¶¶ 15 — 16; Tr. at 136-39 (McDonald).) Senator Paterson testified from a more personal and political perspective that unfair districting practices created a downward spiral of political participation where voters feel alienated because they have no chance of success and thus withdraw from the political process. (See generally Paterson Aff. ¶¶ 35-42.) However, if members of Nassau’s black community perceive that “they are part of vibrant campaign and can vote for a winning candidate,” they will get more involved in the political process, (id. ¶ 41.) Concentrating the black community in one district would also help increase voter mobilization efforts by allowing party members to devote their resources to just one, rather than four, Senate elections. (See id. ¶¶ 39-42.) "

However, there is no reliable evidence that any “warming effect” would be sufficient to make Plaintiffs’ Proposed District 8 a district in which blacks could elect candidates of choice. Dr. McDonald acknowledged that his testimony as to any warming effect was “speculation,” largely based on his general work, in political science and analyses of elections across the country. (Tr. at 138.) Upon questioning by the Court, he further admitted that warming in Nassau County was more of a subjective idea than a statistical one. (Tr. at 133.) Indeed, the plaintiffs did not perform any analyses comparing the turnout rates of minority communities before and after obtaining the ability to elect candidates of choice in different elections or in other locations in the state. There was no analysis of the average turnout rate across the state or the country for blacks when they constitute around one-third or 34% of the district.

To support the warming effect argument, Dr. McDonald recompiled the turnout rates for voters in the areas that comprise Plaintiffs’ Proposed District 8, and he compared the recompiled turnout rate for statewide and State Senate elections. Dr. McDonald found that on average, black voters who would be included in the proposed district turned out for their respective State Senate elections at an estimated rate of only 16.1%. (McDonald Decl. ¶ 80.) The recompiled turnout estimate for whites averaged 49.6%. For the statewide elections, however, the average estimated turnout rate for blacks jumped to around 28 to 29% while for whites it went down to around 42%. 69 (Id.) Dr. McDonald and Dr. Handley argue that these figures show that when black voters perceive that they can elect candidates of choices — as in the statewide elections' — 'they will vote in greater numbers.

The recompiled statistics do not provide persuasive evidence of warming or that turnout will increase sufficiently to create an effective district. 70 The increased turnout rates can be attributed more to the fact that elections for Governor and United States Senator, for example, draw more attention than elections for state legislators. That is, the rise in turnout for the statewide elections is likely attributable to “roll-off’ — the tendency for people to cast votes for the most high profile elections at the top of the ballot while “rolling off’ and leaving the ballot blank for the more obscure elections. (See Tr. at 144 (McDonald).) Both media attention and the physical position on the ballot affect the voting rates for a particular election, and roll off as a phenomenon is particularly prevalent among minority voters. (Id. at 145.) Dr. McDonald also conceded that he could not make judgments about the roll-off in New York because it would depend on the structure of the ballot, which he understood was “quirky” but which he had not studied. (Id.) The plaintiffs have provided no evidence beyond broad conclusory and unpersuasive statements that the higher statewide voting rates for blacks are attributable to the perceived ability to elect a candidate of choice. Moreover, even accepting the recompiled turnout rates for the statewide elections, whites still turned out at a rate roughly one-and-a-half times higher than blacks. According to Dr. Handley’s analysis, the white/blaek turnout ratio would have to be the opposite for a 34% black voting age population to be sufficient for an effective district. (See Handley Deck ¶ 14; Tr. at 148-50 (McDonald).)

Ultimately, as Dr. McDonald recognized, the ability of statistics to predict the future is limited, and election results will inevitably depend on “[w]hat candidates will be offered to the voters and how that might affect their voting patterns.” (Tr. at 146.) In particular, Dr. McDonald acknowledged that incumbent candidates tend to have a substantial advantage, and it would be more difficult for a non-incumbent to win if one of the Republican incumbents ran in Proposed District 8. (Tr. at 130.) While Dr. McDonald did not specifically analyze any incumbency advantage for plaintiffs’ proposed district, in a different case in Arizona, he estimated a 9.4% advantage for incumbents in congressional elections, though it was less for state legislative elections. (Tr. at 130.)

In fact, the parties acknowledged that the candidate likely to run for the Republicans in the plaintiffs’ proposed district is Sen. Hannon, the incumbent for SD 6, which constitutes the largest share of the plaintiffs’ proposed district. Sen. Paterson credibly testified for the plaintiffs that Sen. Hannon, who has served in the State Senate for around fourteen years, would likely win in the plaintiffs’ proposed district because he is the incumbent. (Tr. at 538.) Sen. Paterson further testified that Sen. Hannon would likely continue to win the district, and the opportunity to elect a black-preferred candidate would arise only after Sen. Hannon retires. (Tr. at 540-41.)

Virtually every federal court to consider the issue has interpreted Gingles as requiring the minority group to be a majority in a sufficiently compact district. If that test were applied here, the plaintiffs would plainly fail to establish a violation of section 2 in Nassau County. Even if the first Gingles factor were applied flexibly to accommodate crossover or “ability to elect” districts, the plaintiffs would have to prove that their proposed district would provide blacks with the ability to elect candidates of choice. Cf. Growe, 507 U.S. at 41, 113 S.Ct. 1075. Given that the black VAP in the proposed district is nowhere near a majority, the plaintiffs had the burden to prove by credible and persuasive evidence that blacks would have the ability to elect candidates of choice in the Plaintiffs’ Proposed District 8.

The plaintiffs have failed to satisfy their burden. Their evidence is often questionable in its reliability, and even if taken at face value, much of it actually casts doubt on the' effectiveness of the plaintiffs’ proposed district. Even the plaintiffs admit that for blacks to have the ability to elect candidates of choice in the proposed district there would have to be a dramatic increase in black voter turnout. While the plaintiffs presented intuitive theories suggesting that voting rates would increase in general, by the admission of the plaintiffs’ own expert, estimating the degree of warming, if any, would be speculative. And at best, the district would be competitive only in the future, after the white-preferred incumbent retires. The danger, of course, in moving away from the Gingles majority-minority requirement is that it invites claims that are “based on quite speculative foundations.” McNeil, 851 F.2d at 944. If the first Gingles requirement were to be relaxed, there must be more compelling evidence of the minority’s ability to elect than the plaintiffs have presented in this case for Nassau County.

Because we do not find that there is sufficient evidence to prove that Plaintiffs’ Proposed District 8 is an effective black district, the drawing of SDs 6, 7, 8, and 9 could not have caused an injury that is cognizable under the VRA. We therefore do not need to reach the totality of the circumstances inquiry, which is considered only after the Gingles preconditions are satisfied. 71

C. The Bronx (Count III)

1. Introduction

Count III is a black-Hispanie “minority-coalition claim” which advocates the creation of a fifth majority-minority district in the Bronx (Plaintiffs’ Proposed District 36) based upon alleged Hispanic and black cohesion. 72 During the course of the litigation, the plaintiffs have also described this claim in terms of a single (Hispanic) minority. (See, e.g., Pis.’ PreTrial Br. at 14-15 (“Hispanics in the Bronx and Westchester counties ‘have sufficient numbers to constitute a majority in [four] districts. So apportioned, the group will inevitably elect [four] candidates of its choice, assuming the group is sufficiently cohesive.’ ” (quoting Voinovich v. Quilter, 507 U.S. 146, 153, 113 S.Ct. 1149, 122 L.Ed.2d 500 (1993))).)

The plaintiffs charge the legislature with failing to create an additional majority-minority district in the Bronx in 2002 and, thereby, “diluting the voting power of’ Hispanic and black voters. (Comply 304.) Plaintiffs’ Proposed District 36 would have a VAP of 50.1% Hispanic, 23.0% black, and 19.1% white. Its CVAP demographics would be 45.4% Hispanic, 26.0% black, and 23.4% white. 73 The plaintiffs argue that three alleged “discriminatory practices” together gave rise to the Legislature’s failure to draw a fifth (out of five) majority-minority district in the Bronx. (Id. ¶¶ 303-05.) 74 First, the plaintiffs contend that the legislature “packed” minority voters into majority-minority SDs 28, 32, 33, and 36 (id. ¶ 307), and they propose to unpack those districts. 75 Second, the plaintiffs allege that SD 34 is a racially gerrymandered majority-white district (see discussion of Count VII, infra). (Id. ¶ 306.) Third, the plaintiffs contend that New York’s population is distributed disproportionately to the detriment of “downstate.” (Id. ¶ 308; see discussion of Count I, supra.)

The defendants deny the discriminatory practices alleged. They contend that the majority-minority districts are not packed. They point out that the Hispanic majorities in three of these districts were, in fact, reduced in 2002, and they argue that “unpacking” as the plaintiffs propose would raise section 5 retrogression problems. (See Burgeson Aff. ¶ 52 (“The difference between the Revised Plaintiffs’ plan and the 2002 Senate plan is especially significant since the 1990’s plan under the 2000 census, which I understand is the legal benchmark for Section 5 of the Voting Rights Act, had an average Hispanic VAP of almost 62% in the three comparable districts.... [Minority population reductions such as those in the Revised Plaintiffs’ plan are ... problematic -”).) They also contend that the drawing of SD 34 is consistent with preserving the core of the district and protecting incumbents, among other traditional districting principles. The defendants also point out that the Senate Plan’s maximum deviation of less than ten percent is appropriate and that the plaintiffs’ “upstate/downstate” claim did not survive summary judgment. See Rodriguez, 2003 WL 22853022, at *1. 76

A central issue in this claim is whether to define the relevant minority ^ group in terms of a Hispanic-black coalition or Hispanics alone. While courts are divided as to whether minority-coalition claims are cognizable, we follow the Court of Appeals for the Second Circuit in assuming that blacks and Hispanics can be combined in a section 2 claim if evidence is presented to show they are politically cohesive. See Bridgeport Coalition for Fair Representation v. City of Bridgeport, 26 F.3d 271, 276 (2d Cir.) (affirming district court finding of Hispanic-black cohesion but also noting finding that “[i]t is preferable to create majority districts for each group”), vacated and remanded on other grounds, 512 U.S. 1283, 115 S.Ct. 35, 129 L.Ed.2d 931 (1994); Arbor Hill Concerned Citizens Neighborhood Ass’n v. County of Albany, 281 F.Supp.2d 436, 445 (N.D.N.Y.2003) (citing Bridgeport). Compare Concerned Citizens v. Hardee County Bd., 906 F.2d 524, 526 (11th Cir.1990) (“Two minority groups ..: may be a single section 2 minority if they can establish that they behave in a politically cohesive manner.”), with Nixon v. Kent County, 76 F.3d 1381, 1393 (6th Cir.1996) (rejecting minority-coalition suits based on text, history, and purpose of VRA). However, as emphasized in Growe v. Emison, 507 U.S. 25, 41, 113 S.Ct. 1075, 122 L.Ed.2d 388 (1993), the plaintiffs bear the burden of proving that Hispanics and blacks are politically cohesive and should be treated as the relevant minority group for the Gingles analysis and section 2 inquiry. 77

We analyze Count III as both a “minority-coalition claim” and a single-minority claim. But however the claim is styled, we find that thé plaintiffs have failed- to establish vote dilution. Our findings, summarized, are as follows.

The plaintiffs can only establish the first Gingles precondition if we accept either that the plaintiffs can pursue their minori ty-coalition claim or that, counting Hispanics alone, a CVAP of 45.4% and a VAP of 50.1% are sufficient to establish a majority in a district. Even if the plaintiffs rely on Hispanic demographics alone, we find that they may be able to satisfy the first Gin-gles factor because Hispanics would be a majority of the VAP and a near majority of the CVAP, and because evidence supports their ability to elect candidates of choice in the proposed district.

With respect to the second Gingles precondition, the plaintiffs can show that Hispanics alone are politically cohesive. However, the plaintiffs cannot establish a minority-coalition claim based on Hispanic-black cohesion. See Growe, 507 U.S. at 41, 113 S.Ct. 1075. Having made this finding under the second Gingles precondition, we analyze the third Gingles precondition with respect to Hispanics alone. We find that the plaintiffs appear to be able to establish the third Gingles precondition because white bloc voting usually defeats the Hispanic-preferred candidate, although these results may be attributable to partisanship rather than race.

Even if the plaintiffs are able to surmount the Gingles preconditions, they have not shown under a totality of circumstances that the votes of minorities are diluted in the Bronx. Thus, we are not persuaded that any electoral law, practice, or structure has interacted with social and historical conditions “to cause an inequality in the opportunities enjoyed by [minority] and white voters to elect their preferred representatives.” Gingles, 478 U.S. at 47, 106 S.Ct. 2752. Among other totality considerations, we find that the Senate Plan achieves “substantial proportionality” because, among other reasons, Hispanics constitute 42.0% of the CVAP of the Bronx and “control” 60.0% (three of five) of the Senate Plan Bronx districts, and blacks and Hispanics together constitute 76.1% of the CVAP of the Bronx and “control” 80.0% (four of five) of the Senate Plan Bronx districts and, counting. See De Grandy, 512 U.S. at 1014-16, 114 S.Ct. 2647. We also find that the Legislature’s redistricting policies were sound and not “tenuous.”

2. Background to Plaintiffs’ Proposed District 36

The following facts, which are not in dispute, are relevant to our consideration of Count III. Bronx County had a total population of 1,332,650 in 2000, which was: 48.4% Hispanic, 31.2% black, and 14.5% white. The total Bronx VAP was 935,278 in 2000, including: 45.3% Hispanic, 31.0% black, and 17.7% white. The Bronx CVAP was 736,157 in 2000, including: 42.0% Hispanic, 34.1% black, and 20.6% white.

The 2002 Senate Plan includes five Bronx Senate districts, located entirely or primarily in that county, including three that are majority Hispanic (SDs 28, 32, and 33); one that is majority black (SD 36); and one that is majority white (SD 34). SD 28 is represented by Olga Mendez, a Hispanic Republican (formerly a Democrat), and is located primarily in the south and west Bronx (where 63.0% of the district falls) and also encompasses a relatively small part of northeastern Manhattan. Sen. Mendez was first elected in 1978. 78 SD 32 is represented by Ruben Diaz, a Hispanic Democrat, and falls entirely within the Bronx and is concentrated in the south Bronx. Sen. Diaz was first elected in 2002. SD 33 is represented by Efrain Gonzalez, a Hispanic Democrat, and falls entirely within the Bronx and is concentrated in the west Bronx. Sen. Gonza lez was first elected in 1989. SD 36 is represented by Ruth Hassell-Thompson, a black Democrat, and is concentrated in the north Bronx (where 86.8% of the district falls) but stretches from the south Bronx into Mt. Vernon in southern Westchester County. Sen. Hassell-Thompson was first elected in 2000. SD 34 is represented by Guy J. Velella, a white Republican, and is concentrated in the Bronx (where 69.8% of the district falls) but also includes portions of southern Westchester. Sen. Velella was first elected to the Senate in 1986. 79

Under the Senate Plan, SDs 28, 32, 33, 34, and 36 combined contain a total population of 1,306,195, which includes 98.0% of Bronx County’s total population, 14.6% of Westchester County’s total population, and 7.5% of Manhattan’s total population. (See DX 56, 73.) More than four-fifths (86.2%) of the total population of these five districts is located in the Bronx. (See DX 73.) By contrast, Plaintiffs’ Revised Plan would include relatively less Bronx and more Westchester population. The five (entirely or primarily) Bronx districts in Plaintiffs’ Revised Plan — Proposed Districts 31, 32, 34, 35, 36 — contain a total population of 1,240,074, which includes 93.1% of Bronx County’s total population, 16.7% of Westchester County’s total population, and 7.7% of Manhattan’s total population. (See DX 56, 74.) Of the total population of the five districts in Plaintiffs’ Revised Plan, 82.1% is based in the Bronx. (See DX 74.)

Under the 2002 Senate Plan and Revised Plaintiffs’ Plan, the racial and ethnic composition of the Bronx districts measured by VAP and CVAP is as follows:

_Table 7: DemogRaphiCS FOR BRONX DISTRICTS UnpeR 2002 SENATE PlaN 80 _

2002 Senate Plan Districts Incumbent_ White VAP (CVAP) Black VAP (CVAP) Hispanic VAP (CVAP) Black + Hispanic VAP (CVAP)

SD 28 Mendez 12.9% 28.8% 54.0% 82.8%

_(15.6%) (32.8%) (48.4%)_(81.2%)

SD 32 Espada, Jr. 7.2% 30.0% 58.1% 88.1%

_( 7.8%) (32.7%) (56.3%)_(89.0%)

SD 33 Gonzalez, Jr. 13.0% 23.3% 56.7% 80.0%

_(16.0%) (27.0%) (52.5%)_(79.5%)

SD 34 Velella 59.5% 13.3% 20.4% 33.4%

_(62.7%) (13.5%) (19.5%)_(33.0%)

SD 36 Hassell-Thompson 6.5% 65.2% 25.8% (91.0%)

_( 7.7%) (64.6%) (25.6%)_(90.2%)

Table 8: Demographics FOR Proposed District 36 and Other Pbqpqsed Bronx DistRicts

Plaintiffs’ Proposed Districts Incumbent (Corresponding 2002 SD) 81 White VAP (CVAP) Black VAP (CVAP) Hispanic VAP (CVAP) Black + Hispanic VAP (CVAP)

-SD36 iNone7 82 19.1% 23.0% 50.1% 73.1%

_(23.4%) (26.0%) (45.4%) (71.5%)

SD 31 Mendez 10.9% 30.5% 55.1% 85.6%

_(2002 SD 28) , (12.4%) (34.3%) (50.8%) (85.1%)

SD 32 Espada, Jr. 12.5% 30.5% 51.7% ' 82.2%

_(2002 SD 32)_(14.2%) (32:2%) (49.7%)_(81.9%)

SD 34 Gonzalez, Jr./ Velella 83 13.5% 28.4% 53.4% ' 81.8%

_(2002 SDs 33, 34) (16.5%) (32.5%) (47.9%)_(80.4%)

SD 35 Hassell-Thompson 29.6% 51.2% 15.1% 66.3%

'(2002 SD 36) (32.5%) (48.8%) (15.7%) (64.5%)

As noted, the new majority-minority district proposed by the plaintiffs, Proposed District 36, would have the following demographics: 50.1% Hispanic VAP (45.4% CVAP); 23.0% black VAP (26.0% CVAP); and 19.1% white VAP (23.4% CVAP). It would draw its population from five existing districts, including three majority-minority Bronx districts and two majority-white districts, one in the Bronx (SD 34) and one in Westchester (SD 35), as follows: 1.1% of the total population would come from SD 32; 63.2% of the total population would come from SD 33; 3.9% of the total populatiop would come from SD 34; 28.0% of the total population would come from SD 35; and 3.8% of the total population would come from SD 36. (DX 110.) Stated in other terms, Plaintiffs’ Proposed District 36 draws 91.2% of its total population. from two heavily Hispanic areas: the geographic center of SD 33, which is a majority-Hispanic Bronx district, and the portion of SD 35 (located in southwestern Yonkers) that has a Hispanic VAP of 41.4%. (See DX 74.) The Revised Plaintiffs’ Plan includes four Bronx majority-minority districts that have smaller minority majorities than their current counterparts, the fifth district being the proposed new majority-minority District 36.

3. The First Gingles Precondition

The plaintiffs contend that if Hispanics and blacks are counted together, or even if Hispanics are counted alone, they can satisfy the first Gingles precondition because minorities would constitute a majority of the total population and VAP in Proposed District 36. 84 The defendants respond that the first Gingles prong is met only if the proposed district is composed of (a) a majority of (b) the CVAP of (c) a single minority group. They thus argue that Hispanics alone should be counted for the first Gingles factor, and that because Hispanics constitute only 45.4% of the CVAP of Proposed District 36, the plaintiffs fail to meet their burden.

If blacks and Hispanics are counted together, they would easily constitute “a majority in a single-member district.” Growe, 507 U.S. at 40, 113 S.Ct. 1075. Together, they are 73.1% of the VAP and 71.4% of the CVAP in Proposed District 36. See supra Table 8. Nonetheless, because we find, as we explain below in our analysis of cohesion, that black and Hispanic voters are not cohesive in the Bronx, we must evaluate Plaintiffs’ Proposed District 36 as a Hispanic district for the purposes of the first Gingles precondition. In doing so we consider (a) whether VAP or CVAP is the appropriate measure for determining a “majority” under the first Gingles factor and.(b) whether Plaintiffs’ Proposed District 36 would be an “electable” Hispanic district.

(a) The Appropriate Measure

Counted alone, Hispanics would constitute a bare majority (50.1%) of the VAP in Proposed District 36 but less than a majority (45.4%) of the CVAP. Thus, if we were to adhere to a “bright line” majority requirement, the plaintiffs’ majority claim fails on the basis of CVAP but possibly survives on the basis of VAP. Here, too, courts appear to differ. Some have measured “majority” by reference to VAP. See, e.g., Solomon v. Liberty County, 899 F.2d 1012, 1018 (11th Cir.1990) (en banc) (Kravitz, J., specially concurring) (“[T]he undisputed evidence shows that blacks would constitute a[51%] majority of [the proposed district’s] voting age population. That conclusively establishes that blacks are a sufficiently large ... group ... to be eligible for relief under the Voting Rights Act.”). Other courts emphasize CVAP, observing that the Voting Rights Act specifically proscribes the “denial or abridgement of the right of any citizen of the United States to vote on account of race or color.” 42 U.S.C. § 1973(a) (emphasis added); see, e.g., Negron v. City of Miami Beach, Fla., 113 F.3d 1563, 1569 (11th Cir.1997) (stating that “the proper statistic for deciding whether a minority group is sufficiently large and geographically compact is voting age population as refined by citizenship”); France v. Pataki, 71 F.Supp.2d 317, 326 (S.D.N.Y.1999) (criticizing plaintiffs’ plan because “the proper measure of population [for purposes of the first Gingles requirement] is the total citizens eligible to register and vote” and plaintiffs’ plan “failed to take citizenship [of Hispanics] into account” (citing Valdespino v. Alamo Heights Indep. Sch. Dist., 168 F.3d 848, 853 (5th Cir.1999))); Session v. Perry, 298 F.Supp.2d 451, 494 n. 133 (E.D.Tex.2004) (observing that every court to “consider the question, has concluded that the relevant voting population ... is citizen voting age population”); see also supra note 38.

The Supreme Court has not ruled explicitly on the subject. It has advised that there is no “magic parameter” in section 2 cases, De Grandy, 512 U.S. at 1017 n. 14, 114 S.Ct. 2647, and that “the Gingles factors cannot be applied mechanically and without regard to the nature of the claim.” Voinovich, 507 U.S. at 158, 113 S.Ct. 1149. In De Grandy, the Supreme Court explained:

The parties’ ostensibly factual disagreement raises an issue of law about which characteristic of minority populations {e.g., age, citizenship) ought to be the touchstone for proving a dilution claim and devising a sound remedy.... As in the past, we will assume without deciding that even if Hispanics are not an absolute majority of the relevant population in the additional districts, the first Gingles condition has been satisfied in these cases.

De Grandy, 512 U.S. at 1008-09, 114 S.Ct. 2647 (citing Voinovich, 507 U.S. at 154, 113 S.Ct. 1149; Growe, 507 U.S. at 41-42, n. 5, 113 S.Ct. 1075; Gingles, 478 U.S. at 46-47 & n. 12, 106 S.Ct. 2752). We will assume arguendo, on the specific facts presented here and without conclusively deciding the issue of law, that an ethnic minority group constituting a majority (50.1%) of the VAP and a near majority of the CVAP (45.4%) in a single-member district may under some circumstances satisfy the first Gin-gles precondition. See Solomon, 899 F.2d at 1018 (“This holding should not be read to imply an opposite result where blacks do not constitute an outright majority of the [VAP] in any district.... In other cases, blacks may be so close to fifty percent that they would have a realistic chance of electing a representative ....”).

(b) Electability

Because we find no Hispanic-black cohesion and because the plaintiffs’ proposed Hispanic majority is numerically tenuous, it is important for the plaintiffs to show that their proposed district is “electable” for Hispanics. See Gingles, 478 U.S. at 50 n. 17, 106 S.Ct. 2752. The plaintiffs argue that the two electability analyses — recompilation and supplemental electability analysis — conducted by their expert, Dr. McDonald, demonstrate that Proposed District 36 would be an effective district for Hispanics. The defendants question the plaintiffs’ methodology and challenge their conclusion that either analysis reliably demonstrates electability. 85

The plaintiffs contend that, on the basis of “recompiled” elections, Proposed District 36 would provide Hispanic voters with an electable district. As explained with respect to Count V involving Nassau County, in recompilation, results from pri- or (exogenous) elections are re-aggregated or recompiled in an effort to determine whether minority voters would have an opportunity to elect their preferred candidate in a proposed district. “The recompilation of election results entails no statistical estimation .... Recompilation is simply a procedure of aggregating votes from elections[ ] for the area composing a proposed district.” (McDonald Decl. ¶ 62.) 86 For his recompilation analysis, Dr. McDonald used elections that he determined (i) offered all voters the same choice of candidates and (ii) featured a “[Hispanic]-preferred candidate fac[ing] racial bloc voting on the part of white voters, comparable to the racial bloc voting [he] found in [his] analysis of the State Senate elections in the same areas [SDs 34 and 35].” (Id. ¶¶ 61-62.) 87 Dr. McDonald argued that “racially polarized elections” provide the “toughest test” for Hispanic-preferred candidates. (Id. ¶ 60; Tr. at 220 (McDonald); see also Handley Decl. ¶ 20.)

The New York gubernatorial races won by Governor Pataki in 2002 and 1998, among others, were selected by Dr. McDonald for recompilation. (See McDonald Decl. ¶ 65.) Dr. McDonald determined that the Hispanic minority-preferred candidate in both gubernatorial elections (Carl McCall, who is black, in 2002 and Peter Vallone, who is white, in 1998) were favored by an estimated 35% and 39%, respectively, of white voters, but in both cases “won” Proposed District 36 by a large margin. (Id.) Thus, in 2002, while Pataki beat McCall by a large margin statewide (and in SDs 34 and 35), Pataki would have lost to McCall by 8320 votes in the proposed district, under Dr. McDonald’s recompilation analysis. (PX 68.) In 1998, Pataki beat Vallone by a substantial margin statewide (and in SDs 34 and 35) but in Proposed District 36 Pataki would have lost to Vallone by 12,886 votes, according to Dr. McDonald. (McDonald Decl. ¶ 65; PX 68.) Dr. McDonald also reported recompilations of three additional elections that he determined were racially polarized only in SD 34 (not SD 35). Each of these three elections (i.e., the 2000 United States Senate race, the 1998 United States Senate race, and the 1998 Attorney General race) showed that the Hispanic-preferred candidate won by a large margin in Proposed District 36. (Id. ¶ 66.) In the 2000 United States Senate race, the Hispanic-preferred candidate won by 48,041 to 9565; in the 1998 United States Senate race, the Hispanic-preferred candidate won by 28,423 to 8023; and in the 1998 Attorney General race, the Hispanic-preferred candidate won by 24,994 to 6921. (PX 11.) 88

The defendants’ expert, Dr. Stanley, challenged Dr. McDonald’s recompilation methodology on three principal grounds and analyzed certain elections to assess Dr. McDonald’s conclusions. First, Dr. Stanley complained that “for his electability analysis, Dr. McDonald offers predominantly white-white contests in exogenous elections.” (Stanley Aff. ¶ 108.) Second, Dr. Stanley criticized Dr. McDonald’s recompilation because it did not “entail any assessment of turnout, cohesion or white cross-over votingf] the method used to analyze minority success and voting patterns for the existing districts.” (Id. ¶ 110.) Third, Dr. Stanley pointed out that Dr. McDonald’s analysis “included and excluded elections in a biased manner to suggest that the existing districts provide less opportunity than they actually do to elect minority preferred candidates.” (Id. ¶ 109.) 89

ii) Supplemental Electability Analysis

The plaintiffs contend that a second analysis based on seven endogenous elections shows that Hispanics would be able to elect their candidates of choice in Proposed District 36. The defendants counter that the plaintiffs’ expert’s methodology is unreliable and that the analysis says little about Hispanics’ actual ability to nominate and elect candidates of choice.

As explained with respect to the Nassau County claim, Dr. McDonald’s supplemental electability analysis is a statistical estimate based upon turnout and upon his racial bloc voting analysis of endogenous elections. 90 He testified that the seven endogenous elections he analyzed show that the Hispanic-preferred candidate in Proposed District 36 would receive 56.1% of the total vote on average and, in the “best case,” 64.5% of the total vote. (McDonald Decl. ¶ 77.) 91 And, he predicted that based on demographic trends, these electability numbers “will continue to improve during the course of the present decade.” (Id. ¶¶ 45-47.) 92

The defendants’ expert, Dr. Stanley, complained that, “[a]lthough this appears to be a ‘winnable’ district,” Dr. McDonald’s analysis “provides little evidence that the proposed ‘Hispanic district’ 36 would empower Hispanic voters relative to the- existing situation.” (Stanley Aff. ¶ 128.) He estimated that Hispanics would constitute only 16.5% of the actual voters in the proposed district and “would not be able to independently choose the winning candidate.” (Id. ¶ 129.) Dr. Stanley observed that “projecting electoral success in Plaintiffs’ District 36 on elections in Districts 34 and 35 is not reasonable since those districts comprise only 31.9%” of the proposed district. (Id. ¶ 130.) He also observed that, “[m]ost importantly, Dr. McDonald’s [supplemental] electability analysis is based solely on general elections, although the minority-preferred candidate would obviously have to win the Democratic primary, particularly in this predominantly. Democratic area,” and he cites an absence of evidence that a Hispanic candidate would be victorious in such a primary. (Id. ¶ 132.)

While we recognize that there are questions about the reliability of the electability analyses, we conclude that, on balance, the plaintiffs’ data suggest that Proposed District 36 would allow Hispanics to elect candidates of choice even with a 45.4% Hispanic CVAP. We found the electability evidence for Plaintiffs’ Proposed District 8 in Nassau County insufficient because, among other things, the supplemental electability analysis projected that the black-preferred candidate, on average, would have received only 45.3% of the votes in a State Senate election. In contrast, the supplemental electability analysis for Proposed District 36 in the Bronx projects that the Hispanic-preferred candidate would receive an average of 56.1% of the votes in a State Senate election. The recompilation results are also stronger for this claim than the Nassau County claim. For example, the 2002 gubernatorial election involving Pataki and McCall presents the hardest test of electability for a minority-preferred candidate because of the low rate of white crossover. The recompilation results for Proposed District 8 showed McCall receiving 49.7% of the vote and “winning” only by a plurality. In Plaintiffs’ Proposed District 36, however, the recompiled results show McCall with 58.8% of the vote. Other recompiled election results also tend to support the conclusion that the Proposed District 36 would provide Hispanic voters with, in Dr. Stanley’s words, a “winnable” district.

Under these circumstances, because Hispanics in the proposed Bronx district would constitute a geographically compact majority of the VAP and a near-majority of the CVAP, and because evidence indicates that they would likely be able to elect candidates of choice, we find that the plaintiffs have presented sufficient evidence to pass the first Gingles threshold. As explained with respect to the Nassau County vote-dilution claim, the Gingles preconditions are directed at determining whether an injury occurred, and the focus on the concept of injury explains why, under these circumstances, the electability analyses have differing ramifications for the Nassau County and Bronx County claims.

In Count V, the plaintiffs complained that a substantial black population in Nassau County has been “cracked” by district lines and thus deprived blacks of representation in any Nassau County Senate district. But even though Plaintiffs’ Proposed District 8 encompassed substantially all of the black population in Nassau County, blacks still would have constituted only 34.5% of the VAP and 36.9% of the CVAP in the proposed district. The central question of injury in that claim was thus whether black-preferred candidates were consistently defeated because of a voting practice that affected blacks as a minority race or simply because the black population was relatively small and in the political minority.

The core issue in the Bronx claim, unlike the Nassau-County claim, is not whether any effective minority district can be created; the issue is whether section 2 of the VRA requires that a fourth Hispanic district be created. The central injury issue in the Bronx claim thus has more to do with proportionality than electability. However, to reach the question of proportionality, the plaintiffs must still establish the second and third Gingles factors.

4. The Second Gingles Precondition

The plaintiffs contend that Hispanics vote cohesively as a bloc.(“Hispanic cohesion”); that blacks vote cohesively as a bloc (“black cohesion”); and that Hispanics and blacks vote cohesively together (“Hispanic-black cohesion”). The plaintiffs’ argument is based on racial bloc voting analyses performed by their expert, Dr. McDonald, who focused upon election results from two majority-white districts, SDs 34 (Bronx) and 35 (Westchester). The plaintiffs argued that it is appropriate to do so because the other districts that would comprise part of Proposed District 36, .namely, SDs 32, 33, and 36, “are packed majority-minority districts and thus not relevant to the Cringles analysis.” (McDonald Decl. ¶ 19.) 93 While the defendants criticized the narrow focus on SDs 34 and 35, their expert, Dr. Stanley, did not analyze other districts. Instead, Dr. Stanley focused on Hispanic-black cohesion and argued, based on the regression analyses performed by Dr. McDonald for those districts, that no such cohesion exists. We conclude in the discussion that follows that, while Hispanic cohesion and black cohesion appear to be shown separately, the plaintiffs have not established Hispanic-black cohesion in the Bronx. Therefore, the plaintiffs’ minority-coalition claim must fail. 94

Focusing on contests in SDs 34 and 35, Dr. McDonald performed El analyses for endogenous and exogenous elections in those districts and generated 38 election contest results. 95 The parties, however, dispute whether all elections should be included in the assessment of cohesion. Dr. McDonald excluded various elections for reasons explained below and based his cohesion assessment on 33 election analyses. Dr. Stanley criticized Dr. McDonald’s exclusion of certain elections and based his assessment on all 38 election contest analyses performed. 96 While the parties categorized the election results in various ways, we observe that three kinds of elections define the universe of relevant results: (1) endogenous elections for State Senate Districts 34 and 35; (2) exogenous statewide elections for state and federal offices; and (3) exogenous Democratic primary elections involving minority candidates. We discuss these categories of elections and the corresponding contest results in turn.

Dr. McDonald analyzed seven State Senate (endogenous) general elections in SDs 34 and 35 between 1996-2002. Those elections are as follows:

Table 9: Seven Endogenous General Elections 97

State Senate Election 98 Candidates 99 Hispanic Cohesion Black Cohesion Hispanic-Black Cohesion White Bloc Voting Source Page 100

2002 SD 34 Velella/Mahoney * Yes Yes Yes Yes. 6

2002 SD 35 Spano/Ramos * Yes Yes Yes Yes 7

2000 SD 34 Velella/Koppel * Yes Yes Yes Yes 13

2000 SD 35 Spano/Abinati Yes Yes Yes Yes 14

1998 SD 34 Velella/Spallone * Yes Yes Yes Yes 22

1998 SD 35 Spano/Ploski * Yes Yes Yes Yes 23

1996 SD 35 Spano/Abinati * Yes Yes Yes Yes 30

In all of these elections, Hispanics voted as a bloc, blacks voted as a bloc, [and] black and Hispanic voters cohesively supported the same candidate.” (Id. ¶ 23.) Dr. McDonald concluded that “[t]he evidence from the endogenous elections, in both [SDs 34 and 35], is overwhelming that minority group voters have been politically cohesive.... The 2nd ... Gingles prong[ is] clearly and unequivocally satisfied by the results in these elections.” (Id. ¶ 25.)

Dr. McDonald performed El analyses for 10 exogenous statewide general elections in SDs 34 and 35 between 1996 and 2002, 101 thus generating results for 20 statewide election contests. Dr. McDonald included all the results except those generated out of the 2002 Attorney General election between Eliot Spitzer and Dora Irrizary. (Id. ¶¶ 36-37, 48-51.) 102 The statewide election results are as follows:

Table 10: Twenty Exogenous General Election Results 103

Election

Candidates Hispanic Cohesion Black Cohesion Hispanic-Black Cohesion White Bloc Voting Source Page

18 Election Results Included bvDr. McDonald

2002 Governor (SD 34) Pataki / McCall* Yes Yes Yes Yes 59

2002 Governor (SD 35) No* 104 Yes No* Yes 60

2002 Comptroller (SD 34) Hevesi* / Faso Yes Yes Yes No 90

2002 Comptroller (SD 35) Yes Yes Yes No 91

2000 U.S. President (SD 34) Bush / Gore* Yes Yes Yes No 98

2000 U.S. President (SD 35) Yes Yes Yes No 99

2000 U.S. Senate (SD 34) Clinton* / Lazio Yes Yes Yes Yes 106

2000 U.S. Senate (SD 35) Yes Yes Yes No 107

1998 U.S. Senate (SD 34) Schumer* /D’Amato Yes Yes Yes Yes 113a

1998 U.S. Senate (SD 35) Yes Yes Yes No 114

1998 Governor (SD 34) Pataki / Vallone* Yes Yes Yes Yes 121

1998 Governor (SD 35) Yes Yes Yes Yes to 122

1998 Attorney General (SD 34) Spitzer* / Vacco Yes Yes Yes Yes 129

1998 Attorney General (SD 35) Yes Yes Yes No o 130

1998 Comptroller (SD 34) McCall* I Blakeman Yes Yes Yes No ^ Ui "J 75

1998 Comptroller (SD 35) Yes Yes Yes No w '-4 76

1996 U.S. President (SD 34) Clinton* / Dole Yes Yes Yes No W s) . 137

1996 U.S. President (SD 35) Yes Yes Yes No U-» CO 138

2 Election Results Not Included bv Dr. McDonald

2002 Attorney General (SD 34) Spitzer* / Irrizary Yes Yes Yes No 67

2002 Attorney General (SD 35) , Yes Yes Yes No 68

In the elections with a minority candidate — Carl McCall for comptroller in 1998 and governor in 2002 — Dr. McDonald found Hispanic-black cohesion in SD 34 for both elections. He found Hispanic-black cohesion in support of McCall in the 1998 election in SD 35, but he testified that it “could not be determined, based on the 95% confidence interval, whether Hispanic support for McCall reached 60%.” (See id. ¶ 36 (corrected)). In all fourteen statewide contests without a minority candidate, Dr. McDonald found Hispanic-black cohesion. (Id. ¶ 37.)

Dr. McDonald also presented the results from eight exogenous “local” — -New York City and Bronx County — Democratic primary elections, each of which included a minority candidate and, according to Dr. McDonald, “showfed] a strong pattern of racial polarization.” (Id. ¶ 38.) (For these citywide and countywide elections, he analyzed only contests in SD 34 because SD 35, as noted, is in Westchester County.) He also generated results for three pri mary election contests that he excluded from his cohesion assessment: the 2001 CD 17 results for SDs 34 and 35 and the 1998 Lieutenant Governor results m SD 34. 105 eieven citywide and statewide primary election results are as follows:

Table 11: Eleven Exogenous Democratic Primary Election Results 106

Election Candidates Hispanic Cohesion Black Cohesion Hispanic -Black Cohesion White Bloc Voting Source Page

8 Local-Election Results Included bv Dr. McDonald

2001 NYC Mayor (SD 34) Ferrer* / Green / Vallone Yes Yes Yes Yes 45

2001 NYC Mayor Runoff (SD 34) Green / Ferrer* Yes Yes Yes Yes 46

2001 NYC Comptroller (SD 34) Thompson / Berman No Yes No Yes 47

Colon / Multiple others 2001 NYC Pub. Adv. (SD 34) No No* 107 No No 48 '

2001 Bronx Borough Pres. (SD 34) Carrión* / Eisland / Espada* Yes 108 Yes Yes Yes ' 52

2001 Civil Court (SD 34) Schachner / Jenkins Yes No* No* Yes 49

1997 NYC Mayor (SD 34) Mcssinger / Sharpton / No No No No 50

Albanese

1997 NYC Pub. Adv. (SD 34) M. Green* / R. Green Yes Yes Yes No 51

3 Primary-Ejection Results Hot Included bv Dr. McDonald

2001 U.S. CD 17 (SD 34) Engel/Seabrook/Zayas 109 No Yes No Yes

2001 U.S. CD 17 (SD 35) Engel / Seabrook / Zayas No No No No

1998 Lt. Gov. (SD 34) Frankel/King Yes 110 No No No

Dr. McDonald found Hispanic-black cohesion in four of the eight primaries (50.0%), each of which included one or more minority candidates. (Id. ¶¶ 39-46.) He noted that in the 2001 primary for Bronx Borough president (in SD 34), there were two Hispanic candidates (Espada and Carrión) and one white candidate (Eisland), and he stated that “black and Hispanic voters were estimated to have given Eisland 16.3% and 15.1% of their votes, respectively, ... [and the] two Hispanic candidates proved equally attractive to both black and Hispanic voters.” (Id. ¶ 43.) Combining the vote estimates for the Hispanic candidates, he concluded “this election shows a strong pattern of racial polarization” and Hispanic-black cohesion. (Id.)

In sum, Dr. McDonald performed analy-ses for 38 election contests for SDs 34 and 35 and included in his assessment of cohesion the results for 33 election contests, including 7 endogenous general election results, 18 exogenous statewide general election’ results, and 8 exogenous citywide primary election results. See supra Table 9-Table 11. He found that Hispanic cohesion existed in 28 of 33 (84.8%) election contests; black cohesion existed in 29 of 33 (87.9%) election contests; and Hispanic-black cohesion existed in 28 of 33 (85.7%) election contests. The five contests lacking Hispanic-black cohesion included the 2002 Governor’s election for SD 35 (where Hispanic cohesion was inconclusive) and four Democratic primary elections in SD 34.

The defendants’ expert, Dr. Stanley, assessed Hispanic-black .cohesion based on the results of all 38 election contests for which Dr. McDonald performed El analy-ses. In the elections that Dr. Stanley determined were most probative, namely, the 11 primaries with minority candidates, he found Hispanic-black' cohesion in only three of eleven (27.3%). His analysis, in more detail, is as follows.

Preliminarily, Dr. Stanley raised serious, and we think cogent, challenges to Dr. McDonald’s results, arguing that, “it is not possible to. conclude that blacks and Hispanics are cohesive voting groups,” even using Dr. McDonald’s own election analy-ses. (Stanley Aff. ¶ 46.) “The only similarity between the groups,” Dr. Stanley contended,

is that blacks vote Democratic in overwhelming numbers and Hispanics, to a lesser extent, also tend to support Democrats in general elections. But this phenomenon, typical throughout the United States, is a very weak indication of cohesion since there are many jurisdictions in New York and the United States where whites are predominantly Democratic, but presumably would not, on that basis, be. deemed “cohesive” with blacks or Hispanics. More importantly, in both the minority/white elections Dr. McDonald deems the “best test” of cohesive voting by groups and in Democratic primaries, where partisanship is not a factor, blacks and Hispanics are not cohesive in the vast majority of elections.

(Id. (emphasis added).) Dr. Stanley also criticized Dr. McDonald’s methodology, noting that Dr. McDonald “reports estimates from his three methods for a combined group — blacks and Hispanics — but does not rely on these estimates to draw conclusions about black and Hispanic cohesion. His conclusions about black and Hispanic cohesion rely instead on his separate estimates of each group’s vote.” (Id. ¶ 14.)

Dr. Stanley correctly noted that Dr. McDonald’s “State Senate analyses are restricted to Senate Districts 34 and 35, areas that contain less than a third of the population of the plaintiffs’ proposed State Senate District 36. Findings for particular State Senate elections pertain to only a section of [the] proposed district.” (Id. ¶ 75.) 111 Dr. Stanley further testified:

Assessing voting patterns on the basis of such an extraordinarily small portion of the area alleged to be racially polarized can say very little about whether there are such voting patterns actually exist ing in that area. Indeed, any such analysis would seem to be a facially invalid methodology ....

(Id. ¶ 78 (emphasis added).) Pointing out that the “vast bulk of Plaintiffs’ proposed District 36 is contained by the current District 33 (63.2%),” Dr. Stanley argued, “Dr. McDonald, for some unexplained reason, entirely ignores these districts in assessing cohesion ... even though these districts contain a substantial majority of the population in Plaintiffs’ District 36.” (Id. ¶ 79.)

In the eleven elections which he determined to be most probative — statewide, citywide, countywide, and congressional primaries with minority candidates (see DX 148) — Dr. Stanley concluded that there was Hispanic-black cohesion in only 3 of 11 (27.3%) elections 112 and that “[bjlack and Hispanic voters were cohesive behind a candidate in only 2 of those 11 contests (18%) when white voters were not cohesive behind that candidate.” (Id. ¶ 65.) In the latter two elections (namely, the 2001 Democratic primary for mayor and the subsequent runoff election), the minority-preferred candidate, Fernando Ferrer, “won the primary in District 34 and lost the runoff by only 139 votes.” (Id.) Dr. Stanley concluded that “Hispanic voters were never cohesive for a black candidate [and b]lack voters were never cohesive for a Hispanic candidate, except Ferrer.” (Id. ¶ 66.) He argued that “Dr. McDonald’s evidence shows that when Hispanic voters had the choice of a black candidate in a Democratic primary, they seldom backed the black candidate. Indeed, in the six Democratic primaries with black candidates, Hispanics and whites voted together far more often than did Hispanics and blacks, who were cohesive only once, and in four of the other five such elections, Hispanics “voted cohesively against” the black candidate.” (Id. ¶ 83; DX 167.) And in the seven Democratic primaries with Hispanic candidates, Dr. Stanley found that Hispanic-black cohesion “may have occurred, by the most favorable count, only half the time.” (Stanley Aff. ¶ 52; DX 169.)

Dr. Stanley also analyzed seven general election contests involving minority candidates. The contest results are based on four elections (one endogenous, three statewide exogenous) and are as follows:

Table 12: Seven General Election Contests with Minority Candidates 113

Election Candidates Hispanic Cohesion Black Cohesion Hispanic-Black Cohesion White Bloc Voting Source Page

2002 State Senate (SD 35) Spano / Ramos* Yes Yes Yes Yes 7

2002 Gpvemor (SD 34) Pataki / McCall* Yes Yes Yes Yes 59

2002 Governor (SD 35) Pataki / McCall* No* 114 Yes No* Yes 60

1998 Comptroller (SD 34) McCall* / Blakeman Yes Yes Yes No 75

1998 Comptroller (SD 35) McCall* / Blakeman Yes Yes Yes Yes 76

2002 Attorney General (SD 34) Spitzer* / Irrizary Yes Yes Yes No 67

2002 Attorney General (SD 35) Spitzer* / Irrizary Yes Yes Yes No 68

The sole endogenous contest was the 2002 Senate race in SD 35, in which “black and Hispanic voters were cohesive for a candidate when white voters were cohesive against the candidate.” (Stanley Aff. ¶ 80.) But, as the defendants point out, the plaintiffs’ own expert acknowledged on cross examination that the Hispanic candidate in this election was selected to run “to establish a record of racial bloc[ ] voting” for the litigation. (Tr. at 79-80 (McDonald).) In the other six contests in SDs 34 and 35, Dr. Stanley concluded that black and Hispanic voters “were predictably cohesive in these 6 contests, but were cohesive in only 2 of these 6 contests when white voters were cohesive against the candidate.” (Stanley Aff. ¶ 81.) 115 Thus, overall, in “only 3 of the 7(43%) elections were black and Hispanic voters cohesive for a candidate when white voters were cohesive against the candidate.” (Id. ¶ 82 (misnum-bered).)

Dr. Stanley also analyzed 20 statewide contests without minority candidates, which included 6 endogenous elections and 14 exogenous election contests. See supra Table 9, Table 10. He found: “Black and Hispanic voters were cohesive behind the same candidate, the white Democrat, in all 20 (100%) contests. In 11 of these 20 contests (55%), black and Hispanic voters were cohesive behind a candidate when white voters were cohesive against the candidate. This rate of black-Hispanic cohesion is again nearly double the rate of black-Hispanic cohesion exhibited in those same State Senate Districts when minority candidates ran.” (Id. ¶ 85.)

We find the evidence sufficient to establish Hispanic cohesion separately (which the defendants do not appear to dispute), as reflected in Dr. McDonald’s 33 election results and Dr. Stanley’s 38 election results. 116 The plaintiffs have met the second Gingles precondition for Hispanic cohesion.

At the same time, we find that the plaintiffs’ proof is insufficient to demonstrate Hispanic-black cohesion. See Growe, 507 U.S. at 41, 113 S.Ct. 1075 (citing Badillo v. Stockton, 956 F.2d 884, 891 (9th Cir.1992)); Hardee County, 906 F.2d at 527; Campos v. Baytown, 840 F.2d 1240, 1244 (5th Cir.1988). The plaintiffs have not proven that Hispanics and blacks in the Bronx have “worked together and formed political coalitions.” Hardee County, 906 F.2d at 527. Rather, the evidence is mixed with respect to when and with whom Hispanics, blacks, and whites form political coalitions in the Bronx. Where, as here, the two minority groups are generally affiliated/registered with the same party (Democratic) and vote for that party’s candidates at high rates, primary elections for that party’s candidate are by far the most probative evidence of cohesion. And, the Democratic primary results presented here show that Hispanics and blacks are cohesive barely one third of the time.

The weaknesses in the plaintiffs’ Hispanic-black cohesion claim are as follows:

First, we are not convinced that the existing districts (SDs 32, 33, and 36) that make up almost two-thirds (65.7%) of the VAP in the Bronx and provide over two-thirds (68.1%) of the VAP of Plaintiffs’ Proposed District 36 must be ignored in assessing cohesion, as the plaintiffs argue. See Gingles, 478 U.S. at 60 n. 28, 106 S.Ct. 2752 (citing the need for “district specific” analysis). Such districts are plainly relevant, especially given the importance of primaries in discussing cohesion. The plaintiffs’ own expert was unable to state affirmatively that the evidence of Hispanic-black cohesion was reliable. (See Tr. at 159-60.) 117

Second, we are troubled by the fact that Dr. McDonald, while conceding that statistical procedures work better when more information is available and that “the combined [group] estimates are usually more reliable than estimates for the groups in isolation” (Tr. at 123), presented evidence of Hispanic-black cohesion based on separate estimates of each group’s vote. (See Stanley Aff. ¶¶ 14, 17.) We agree with Dr. Stanley that “[t]he fact that the combined black and Hispanic estimates so frequently contradict the individual estimates casts serious doubt on the reliability of the individual estimates.” (Id. ¶¶ 16,17.) 118

Third, in the primaries analyzed, all of which included at least one minority candidate, we find that there was Hispanic-black cohesion in at most 4 of 11 (36.4%) election contests. 119 This is a critical eohe sion figure. We do not accept the plaintiffs’ suggestion that we should exclude three of these elections either because the candidate was little known, or because there was low turnout, or because controversy touched the election. Most of the primaries, for example, had low turnout. (See DX 146 at 45-52, 83, 152-53 (providing data).) Dr. Stanley was correct in observing that “[y]ou can look at these other contests and there are anomalies in each and every one of them. There is a question whether you should be throwing out elections or you should simply take all the data presented and see what it adds up to.” (Tr. at 715-16.) Based on the evidence presented at trial, we find that Dr. McDonald did not evaluate systematically the contexts for the elections and did not apply any objective principle for excluding elections from his assessment. Moreover, even if we did exclude the three allegedly unreliable election results, there would still be Hispanic-black cohesion in only 4 out of 8 Democratic primary contests with minority candidates.

We find that “[b]lacks and Hispanics do not vote cohesively in primary elections, where their allegiance is free of party affiliation.” Session, 298 F.Supp.2d at 478. Because of this finding and the fact that the plaintiffs have presented cohesion evidence for a very limited portion of their Proposed District 36, the plaintiffs cannot meet their “higher-than-usual” burden to prove black-Hispanic political cohesion. See Growe, 507 U.S. at 41, 113 S.Ct. 1075. The plaintiffs have not presented reliable evidence sufficient to prove that Hispanics and blacks in the Bronx should be treated as a cohesive minority group with unified political interests. Cf. Gingles, 478 U.S. at 51, 106 S.Ct. 2752 (“[I]f the minority group is not politically cohesive, it cannot be said that the [existing plan] thwarts distinctive minority group interests.”).

5. The Third Gingles Precondition

Because no Hispanic-black cohesion has been shown, we continue the Gingles analysis considering the plaintiffs’ proposed (single minority) Hispanic-majority claim. The issue, under the third Gingles requirement, is whether the plaintiffs can prove that white bloc voting usually defeats the Hispanic-preferred candidate.

The plaintiffs focused on the 25 general election contests surveyed for SDs 34 and 35 (including 7 endogenous and 18 exogenous contests out of the 33 total election contests assessed for the cohesion analysis), to show that white voters typically voted as a bloc against the Hispanic-preferred candidate who was usually defeated by the white-preferred candidate. Those elections are as follows:

Table 13: Dr. McDonald’s Twenty-Five General Election Results 120

Election Candidates 121 Hispanic Cohesion White Bloc Voting Result for HPC 122 Source Page

7 Endogenous Elections

2002 SD 34 Velella / Mahoney* Yes Yes Defeated 6

2002 SD 35 Spano / Ramos* Yes Yes Defeated 7

2000 SD 34 Velella/Koppel* Yes Yes Defeated 13

2000 SD 35 Spano / Abinati* Yes Yes Defeated 14

1998 SD 34 Velella / Spallone* Yes Yes Defeated 22

1998 SD 35 Spano / Ploski* Yes Yes Defeated 23

1996 SD 35 Spano / Abinati* Yes Yes Defeated 30

18 Exogenous Election Results

2002 Governor (SD 34) Pataki/McCall* Yes Yes Defeated 59

2002 Governor (SD 35) No* Yes Defeated 60

2002 Comptroller (SD 34) Hevesi* / Faso Yes No Won 90

2002 Comptroller (SD 35) Yes No Won 91

2000 U.S. President (SD 34) Bush / Gore* Yes No Won 98

2000 U.S. President (SD 35) Yes No Won 99

2000 U.S. Senate (SD 34) Clinton* / Lazio Yes Yes Won 123 106

2000 U.S. Senate (SD 35) Yes No Won 107

1998 U.S. Senate (SD 34) Schumer* / D’Amato Yes Yes Defeated 113a

1998 U.S. Senate (SD 35) Yes No Won 114

1998 Governor (SD 34) Pataki / Vallone* Yes Yes Defeated 121

1998 Governor (SD 35) Yes Yes Defeated 122

1998 Attorney General (SD 34) Spitzer* / Vaoco Yes Yes Defeated 129

1998 Attorney General (SD 35) Yes No Won 130

1998 Comptroller (SD 34) McCall* / Blakeman Yes No Won 75

1998 Comptroller (SD 35) Yes No Won 76

1996 U.S. President (SD 34) Clinton* / Dole Yes No Won 137

1996 U.S. President (SD 35) Yes No Won 138

Dr. McDonald’s strongest evidence was that in all seven endogenous general elections, “white voters voted as a bloc against the minority-preferred candidate[ ] and the minority-preferred candidate was defeated by the white-preferred candidate.” (McDonald Decl. ¶ 23.) (In each such case, the Hispanic-preferred candidate and the minority-preferred candidate were the same.) Dr. McDonald concluded, on the basis of these seven elections, that “[t]he evidence from the endogenous elections is overwhelming that ... minority-preferred candidates are usually defeated by racial bloc voting by white voters in the existing Senate districts. The ... 3rd Gingles prong[ is] clearly and unequivocally satisfied by the results in these elections.” (Id. ¶ 25.) Indeed, the plaintiffs need to emphasize these elections because the evidence of racial bloc voting for statewide exogenous elections was mixed, at best. 124

The defendants’ expert, Dr. Stanley, presented results from the 18 election contests with minority candidates, including one endogenous election, six exogenous statewide general election contests, and eleven exogenous Democratic primary contests. Those elections are as follows:

Table 14: Dr. Stanley’s Eighteen Election Results with Minority Candidates

Election Candidates Hisnanic Cohesion White Bloc Voting Result for HPC 125 Source Page

1 Endogenous Election

2002 Senate District 35 Spano/ Ramos* Yes Yes Defeated 7

6 Statewide General Election Results

2002 Governor (SD 34) Pataki / McCall* Yes Yes Defeated 59

2002 Governor (SD 35) Pataki / McCall* No* Yes Defeated 126 60

2002 Attorney General (SD 34) Spitzer* / Irrizary Yes No Won 67

2002 Attorney General (SD 35) Spitzer* / Irrizary Yes No Won 68

1998 Comptroller (SD 34) McCall* / Blakeman Yes No Won 75

1998 Comptroller (SD 35) McCall* / Blakeman Yes No Won 76

8 Local Primary Election Results

2001 NYC Mayor (SD 34) Ferrer* / Green / Vallone Yes Yes Won 45

2001 NYC Mayor Runoff (SD 34) Green / Ferrer* Yes Yes Defeated 127 46

2001 NYC Comptroller (SD 34) Thompson / Berman No Yes No HPC 128 47

2001 NYC Pub. Adv. (SD 34) Colon / Multiple Others No No No HPC 129 48

2001 Bronx Borough Pres. Carrión* / Eisland / Yes Yes Won 52

(SD 34) Espada*

2001 Civil Court (SD 34) Schachner / Jenkins* Yes Yes Won 49

1997 NYC Mayor (SD 34) Messinger/Sharpton/Albanese No No No HPC 130 50

1997 NYC Pub. Adv. (SD 34) M. Green* / R. Green Yes No Won 51

3 Other Primary Results

2001 U.S. CD 17 (SD 34) Engel / Seabrook / Zayas No Yes No HPC 131 152

2001 U.S. CD 17 (SD 35) Engel / Seabrook / Zayas No No No HPC 153

1998 Lt. Gov. (SD 34) Frankel* / King Yes No Won 132 83

'Dr. Stanley’s universe included only five elections that Dr. McDonald had also fo cused on for the racial bloc voting assessment. Dr. Stanley found white bloc voting in 9 of 18 (50.0%) elections, racially polarized voting (as defined by Dr. McDonald) in 6 of 18 (33.3%) elections, and the defeat of the Hispanic-preferred candidate in 4 of 6 (66.7%) racially polarized elections. He concluded that even “[u]nder Dr. McDonald’s own analysis, white bloc voting defeated the minority preferred candidate in only 4 of 18 contests involving minority candidates” in SDs 34 and 35. (Stanley Aff. ¶ 95.) In those 18 elections, according to Dr. Stanley, white bloc voting defeated the minority-preferred candidate in one of 11 primaries and three of seven general election contests. (Id) Dr. Stanley noted that in the 2001 Democratic mayoral primary, Ferrer, who is Hispanic, was backed by Hispanics in the Bronx part of SD 34 and actually won in that district. But in the run-off election that ensued, Ferrer, again backed by Hispanics, narrowly lost in SD 34 by 139 votes. (Id ¶ 96.)

Dr. Stanley also observed “substantial crossover” voting by whites in minority-white elections, “with a 36.1% overall crossover rate and a 41.2% rate when blacks and Hispanics were cohesive, with a range of 5.0% to 90.7%.” (Id ¶¶ 100-01; DX 151.) Thus, he concluded, SDs 34 and 35 “are not jurisdictions that can be fairly characterized as racially polarized.” (Id) The defendants also criticize the plaintiffs’ results because they rely heavily “on contests involving popular Republican incumbents,” for example, Governor Pataki, Sen. Nicholas Spano, and Sen. Velella, who repeatedly “defeated Democratic challengers (the vast majority of whom were white).” (DFF ¶ 471.) As they argued with respect to Nassau County, the defendants claimed that any white bloc voting and defeat of Hispanic-preferred candidates was attributable to partisanship, not race.

“The purpose of inquiring into the existence of racially polarized voting is- twofold: to ascertain whether minority group members constitute a politically cohesive unit and to determine whether whites vote sufficiently as a blo.c usually to defeat the minority’s preferred candidates.” Gingles, 478 U.S. at 56, 106 S.Ct. 2752. We have already determined that Hispanics are cohesive in the Bronx. With respect to white bloc voting, we find, and the parties agreed, that the most probative elections are endogenous elections. See NAACP v. City of Niagara Falls, 65 F.3d 1002, 1015 n. 16 (2d Cir.1995). And, in seven of seyen (100%) endogenous elections white bloc voting defeated the Hispanic-preferred candidate. That is,, there was a Hispanic-preferred candidate in each election; the Hispanic-preferred candidate received between (an estimated) 72.3% and 99.2% of the Hispanic vote in each election; the Hispanic-preferred candidate received between (an estimated), 27.1% and 39.7% of the white vote in each election; and each Hispanic-preferred candidate lost to the white-preferred candidate.

A more mixed .result is shown in exogenous elections in SDs 34 and 35. In 20 exogenous. general election contests, whites voted as a bloc in only seven (of 20) election contests; they did .so in five of the ten election contests in SD 34. In six of the seven elections exhibiting white bloc voting, the Hispanic-preferred candidate was defeated. (And, in each election, whites in SD 34 gave greater support to the white-preferred candidate than they did in SD 35.) See Gingles, 478 U.S. at 55-58, 106 S.Ct. 2752 (“Because .... the extent of bloc voting necessary to demonstrate that a minority’s ability, to elect its preferred representatives is impaired varies according to several factual circumstances, the degree of bloc voting which constitutes the threshold of legal significance will vary from district to district.”).

Our conclusion that there appears to be white bloc voting is tempered by the facts that (i) the plaintiffs’ evidence considers so little of the Bronx population and only about one-third of the Proposed District 36 population, and (ii) partisanship rather than race, which we discuss below, may be the cause. However, as explained in the context of the Nassau County claim, for the purposes of the third Gingles precondition, we do not inquire whether partisanship, rather than race, is the cause of any racial bloc voting. See Goosby, 180 F.3d at 493 (ruling that questions of partisanship should be addressed as part of inquiry into totality of circumstances rather than third Gingles factor). Therefore, we will assume without conclusively deciding that the plaintiffs have met their burden of proving the third Gingles precondition and proceed to an analysis of the “totality of circumstances,” including partisanship. See De Grandy, 512 U.S. at 1013, 114 S.Ct. 2647 (“As facts beyond the ambit of the three Gingles factors loom correspondingly larger, factfinders cannot rest uncritically on assumptions about the force of the Gin-gles factors in pointing to dilution.”). As will be shown, a “totality” analysis defeats the plaintiffs’ claim.

6. Totality of Circumstances

To prove a violation of section 2 of the VRA, the plaintiffs must ultimately establish, “based on the totality of the circumstances,” that because of an unlawful voting practice or procedure, members of a protected class “have less opportunity than other members of the electorate to participate in the political process and to elect representatives of their choice.” VRA § 2(b), 42 U.S.C. § 1973(b). While the “totality of the circumstances” inquiry is guided by the Senate Factors 133 and the “substantial proportionality” analysis articulated by De Grandy, the inquiry must be geared toward the plaintiffs’ theory of how the alleged voting practice dilutes the voting strength of minority voters and limits their opportunity to participate in the political process. A central part of the plaintiffs’ Bronx claim is that the 2002 Senate Plan “packs” minorities into four Bronx districts, avoiding the creation of a fifth majority-minority district and diluting the voting strength of minorities in the Bronx.

The fact that the Bronx currently has three Hispanic-majority districts represented by Hispanic Senators and a majority-black district represented by a black Senator — which provides representation proportionate to the minority populations in the Bronx — is powerful evidence that Hispanics and blacks in the Bronx have equal access to the political process and have the ability to elect candidates of choice. “The extent to which members of a protected class have been elected to office in the State ... is one circumstance that may be considered .... ” 42 U.S.C. § 1973(b). The plaintiffs first presented their claim as a Hispanic-black coalition claim, and by aggregating the minority populations, they asserted that four of the five Bronx districts were loaded with minority VAPs of around 80% to 90%. But those extreme supermajority-minority numbers are produced by treating blacks and Hispanics as a monolithic minority group. Because the plaintiffs have failed to show Hispanic-black cohesion, their claim must be analyzed as an issue of Hispanic vote-dilution. The Hispanic VAPs in the three Hispanic-majority districts range from 54.0% to 58.1%. See supra Table 7. This is clearly not a case where a minority population is “packed” into a certain number of districts denying them the representation that their numbers otherwise would have supported. Under these circumstances, the fact that Hispanics already have representation proportionate to their population in the Bronx is persuasive evidence that the 2002 Senate Plan does not employ invidious voting practices with discriminatory effects.

As we explain below, under the totality of the circumstances, the plaintiffs have failed to establish that the 2002 Senate Plan as it related to the Bronx violates section 2 of the Voting Rights Act.

(a) Substantial Proportionality

The defendants contend that the Legislature adopted a plan in 2002 which achieves- “substantial proportionality” between majority-minority districts and population demographics. See De Grandy, 512 U.S. at 1013-14, 114 S.Ct. 2647. The plaintiffs argue that the defendants cannot prove “statistical proportionality,” and that, under De Grandy, the defendants’ proportionality argument is, at most, a totality factor to be considered and is not a “safe harbor” or affirmative defense. 512 U.S. at 1014, 1021-22, 114 S.Ct. 2647; see also id. at 1025, 114 S.Ct. 2647 (O’Connor, J., concurring) (“The [majority] opinion’s central teaching is that proportionality ... is always relevant evidence in determining vote dilution, but is never itself disposi-tive.”).

The parties presented proportionality calculations based on a claim that aggregated Hispanics and blacks in the Bronx. Because we find no Hispanic-black cohesion and have treated Proposed District 36 as a Hispanic-majority district, the inquiry here will focus on the 'representation of Hispanics. The defendants presented the following proportionality calculations relevant to Hispanics. First, in the Bronx, Hispanics form 42.0% of the CVAP (45.3% VAP) while 60% (three of five) of the districts are majority Hispanic. (DX 57-58; PX 10.) Second, Hispanics are 21.1% of New- York City’s CVAP, which would equate -to 5.5 Senate seats. In fact, Hispanics have six seats (if SD 31 is included) under the 2002 Senate Plan. 134

The plaintiffs respond by redefining the geographic area relevant for the proportionality inquiry. First, the plaintiffs argue that the relevant universe should not be limited to the (five) New York City counties but should also include Westches-ter (“six-county region”) because all of the districts with Hispanic majorities are in the six-county region. In the plaintiffs’ six-county region, 28.7% of the VAP and 19.8% of the CVAP is Hispanic, but, according to the plaintiffs, in only 5 out of 28.6 districts in the six-county region (17.5%) do Hispanic voters have the ability to elect candidates of their choice. (Bever-idge Decl. ¶ 111.) Second, the plaintiffs argue that the three-county region of the Bronx, Manhattan, and Westchester (“three-county region”) is a relevant universe for assessing proportionality. They say that, including the Bronx and the Westchester portion (fraction) of SD 40 and the Manhattan portion (fraction) of SD 25, the three-county region contains 12.2 districts. The Hispanic VAP for the three-county region is 28.5% (CVAP is 23.8%) but Hispanics are able to elect the senator of their choice in three (24.6%) districts (SDs 28, 32, and 33). (Id. ¶ 114.) 135

The Supreme Court has said that the appropriate measure is “substantial proportionality” between the number of majority-minority districts (or effective voting majority districts) and the minority group’s share of the relevant population. De Grandy, 512 U.S. at 1013-15, 114 S.Ct. 2647 (“The court failed to ask whether the totality of facts, including those pointing to proportionality, showed that the new scheme would deny minority voters equal political opportunity.”). We believe that the most appropriate geographical point of reference is Bronx County and the five districts under the Senate Plan that are located there entirely or primarily. See id. at 1022, 114 S.Ct. 2647 (“[W]e have no occasion to decide which frame of reference should have been used if the parties had not apparently agreed in the District Court on the appropriate geographical scope for analyzing the alleged § 2 violation and devising its remedy.”). These five districts in the Senate Plan include 98.0% of Bronx County’s total population, and more than four-fifths or 86.2% of the total population of those five districts are based in the Bronx. Under the Senate Plan, four of the five Bronx districts (80%) are majority minority: three are majority Hispanic (60%); one is majority black (20%); and one is majority white (20%). Thus, the Senate Plan’s Bronx districts are substantially proportional to the Bronx population, which has a CVAP of 42.0% Hispanic, 34.1% black, and 20.6% white, and which has a VAP that is 45.3% Hispanic, 31.0% black, and 17.7% white. 136

“[PJroportionality is significant in evaluating dilution claims and has become a preeminent measure of fairness in redistricting.” Session, 298 F.Supp.2d at 477. 137 The fact that Hispanics, who already control three out of the five districts in the Bronx, enjoy more-than-proportionate representation in that area is strong evidence that the 2002 Senate Plan does not “pack” Hispanics into the current districts so as to dilute their voting strength. 138

(b) Legislative Policies Underlying the Contested Practices .

The plaintiffs contend that each of the “[tjhree discrete discriminatory practices” alleged — namely, packing, gerrymandering, and overpopulating “downstate”' — resulted in the Legislature’s failure to draw a fifth (out of five) Bronx majority-minority district (Compl.lffl 305-08), and that with respect to each practice the underlying State policy is “tenuous,” as defined in Gingles, 478 U.S. at 45, 106 S.Ct. 2752 (citing S.Rep. No. 97-147, at 29, 1982 U.S.C.C.A.N. at 207). The defendants counter that the policies underlying each alleged .practice are sound. “Regarding the idea of a ‘tenuous’ practice, the Senate Report states that a ‘markedf] departure] from past practices or from practices elsewhere in the jurisdiction’ would ‘bear[] on the fairness of [a practice’s] impact.” Niagara Falls, 65 F.3d at 1008 n. 6 (citing S.Rep. No. 97-147, at 29 n. 117, 1982 U.S.C.C.A.N. at pp. 177, 207 n. 117) (emphasis added). We find that the practices reflected in the Senate Plan are sound and that they are consistent with past and present practices of the Legislature. 139

We reject the central premise or thrust of the plaintiffs’ Count III claim, which is that SDs 28, 32, 33, and 36 were “packed” in 2002 with minorities — that is, both Hispanics and blacks — and that it is appropriate to require the State to siphon off minority populations from the majority-Hispanic districts and from the majority-black district in the Bronx to create Plaintiffs’ Proposed District 36. The plaintiffs’ argument is that the four Bronx districts could and should have been drawn with smaller minority VAP and CVAP to facilitate the formation of a fifth majority-minority (Hispanic) district (with a 50.1% VAP and a 45.4% CVAP). That is, the plaintiffs contend that:

• SD 28, with a 54.0% Hispanic VAP and 48.4% Hispanic CVAP, should be redrawn to have a 51.7% Hispanic VAP and 49.7% Hispanic CVAP (Proposed District 32);

• SD 32, with a 58.1% Hispanic VAP and 56.3% Hispanic CVAP should be redrawn so that Hispanics have a 55.1% VAP and 50.8% CVAP (Proposed District 31); and

• SD 33, with a 56.7% Hispanic VAP and 52.5% CVAP, should be redrawn so that Hispanics are 53.4% of the VAP and 47.9% of the CVAP (Proposed District 34).

The plaintiffs thus would reduce both the Hispanic VAP and CVAP of these districts by an average of 2.9 percentage points, allegedly without jeopardizing the ability of Hispanics to elect candidates of choice or reducing their voting strength compared to the current Senate Plan. The plaintiffs also contend that the sole majority-minority black district, SD 36, currently consisting of a 65.2% black VAP and 64.6% black CVAP, could and should be substantially redrawn to have only a 51.2% black VAP and 48.8% black CVAP (Proposed District 35). 140

The defendants correctly contend that the plaintiffs offer no authority for the proposition that section 2 of the Voting-Rights Act compels the Hispanic and black VAP and CVAP reductions the plaintiffs seek — much less a finding that the difference between the VAP and CVAP of those districts in the Senate Plan as compared to the Plaintiffs’ Revised Plan constitutes the difference between “packed” and “unpacked” districts. The defendants contend that the plaintiffs’ proposed dilution of SDs 28, 32, 33, and 36 may threaten the continued viability of those districts for minority voters. (See Burgeson Aff. ¶ 52 (“The difference between the Revised Plaintiffs’ plan and the 2002 Senate plan is especially significant since the 1990’s plan under the 2000 census, which I understand is the legal benchmark for Section 5 of the Voting Rights Act, had an average Hispanic VAP of almost 62% in three comparable districts. My understanding is that minority population reductions such as those in the Revised Plaintiffs’ plan are ... problematic .... ”).) Indeed, the Democratic Senate minority leader and the plaintiffs’ witness in this case, Sen. David Paterson, expressed his doubts that a 55% Hispanic-majority district would be an effective majority-Hispanic district. (See Stip. ¶ 105.) The plaintiffs would lower the CVAP in three of the four Hispanic “majority” districts proposed in their Revised Plan to under 50%. (See DX 60 (noting that Hispanic CVAP is 49.7% in Proposed District 32, 47.9% in Proposed District 34, and 45.4% in Proposed District 36).)

Moreover, the evidence from the redistricting process undermines any contention that the districts as currently drafted further a legislative policy that limits the opportunities for minority voters. Sens. Mendez, Espada, and Gonzalez voted in favor of the Senate Plan, and Sen. Hassell-Thompson “ ‘supported the lines that were drawn that would configure her district.’ ” (Stip. ¶ 107; DX 135, 136.) 141 Far .from reflecting an attempt to pack the Hispanic districts, the 2002 Senate Plan decreased the Hispanic VAPs in the Hispanic-majority districts by 4.2 to 7.4 percentage points. (See PX 27 at 7-11 (LATFOR’s preclearance submission under section 5 of the Voting Rights Act for its 2002 Senate and Assembly Plans).) 142 The 2002 redrawing of the Hispanic-majority districts thus shows an attempt to move away from su-permajority Hispanic districts while maintaining Hispanics’ ability to elect their representatives of choice.

The plaintiffs’ untenable proposition that the State should be required to reduce even further the Hispanic VAP and CVAP in the majority-minority Bronx districts flies in the face of logic and Supreme Court rulings. See Session, 298 F.Supp.2d at 479 (“[S]tates are free to choose the best way to avoid retrogression and ensure equal opportunity to minority voters.” (citing Georgia v. Ashcroft, 539 U.S. 461, 123 S.Ct. 2498, 156 L.Ed.2d 428 (2003))); see also Miller v. Johnson, 515 U.S. 900, 915, 115 S.Ct. 2475, 132 L.Ed.2d 762 (1995). Georgia v. Ashcroft, discussed earlier in the context of the Nassau County crossover claim, involved a “retrogression” claim under section 5 of the Voting Rights Act where the Democrat-controlled legislature developed a redistricting plan that “unpacked the most heavily concentrated majority-minority [State Senate] districts ... and created a number of new influence districts.” Ashcroft, 539 U.S. at —, 123 S.Ct. at 2506. The plan in Ashcroft reduced by five the number of districts with á black VAP in excess of 60%, increased the number of districts with a black VAP of between 25% and 50% by four, and increased.the number of 50% black VAP districts by one. Id. While the United States argued against preclearance because the plan eliminated safe black districts, the Supreme Court recognized that states must have “flexibility” in redistricting, including the latitude to choose one theory of effective representation over the other:

The ability of minority voters to elect a candidate of their choice is important but often complex in practice to determine. In order to maximize the electoral success of a minority group, a State may choose to create a certain number of ‘safe’ districts, in which it is highly likely that minority voters will be able to elect the candidate of their choice. Alternatively, a State may choose to create a greater number of districts in which it is likely — although perhaps not quite as likely as under the benchmark plan— that minority voters will be able to elect candidates of their choice. Section 5 does not dictate that a State must pick one of these methods of redistricting over another.

Id. at 2511-12 (citations omitted) (emphasis added).

While we recognize the difference between the section 5 analysis under Ashcroft and the section 2 analysis at issue here, “some parts of the § 2 analysis may overlap with the § 5 inquiry.” Id. at 2510 (citation omitted). It was reasonable and appropriate for the Senate to opt to maintain the cores of, and to support the incumbents in, its four majority-minority Bronx districts and one majority-white Bronx district given current Bronx demographics. We find no evidence in the record that the New York State Legislature in 2002 went outside permissible bounds in selecting its method of redistricting. Id. at 2512.

With respect to upstate/downstate population differentials, we have found elsewhere in this decision that there was no invidious discrimination and the Legislature pursued legitimate, non-discriminatory redistricting goals. Nor is the “effect” of these population differentials problematic. In fact, when CVAP or registered or actual voters are taken into account, the purported disproportion between “upstate” and “downstate” (on the basis of total population) is reversed. See supra Part III (one-person, one-vote claim); (Defs. SOF ¶¶ 16, 18, 22). And, the parties have stipulated that the 2002 Assembly plan has a total population deviation of 9.43% (Stip-¶ 46), which is almost precisely the deviation in the Senate Plan (9.78%) (id. ¶45). The average deviation of each is also consistent, namely, 2.67% (Assembly) and 2.22% (Senate). (Id. ¶¶ 47-48.) The Senate Plan is plainly not a marked departure from “practices elsewhere in the jurisdiction.” Niagara Falls, 65 F.3d at 1008 n. 6 (citation omitted). 143

With respect to alleged gerrymandering of SD 34, we find that traditional district-ing principles' — including the desire to preserve the cores of existing districts and to protect incumbents — are clearly reflected in the Senate Plan. (See, e.g., Burgeson Aff. ¶¶ 32-35, 38.) As we discuss in detail infra, the core of SD 34 was first established decades ago, and its preservation as a “Republican seat” in 2002 is not a departure but a continuance of past practices. See Niagara Falls, 65 F.3d at 1008 n. 6 (citation omitted).

We find on the facts presented here that the Legislature was “free to choose the best way to ... ensure equal opportunity to minority voters.” Session, 298 F.Supp.2d at 479.

(c) The Degree of Racially Polarized Voting and Political Partisanship

The plaintiffs maintain that the evidence of racially polarized voting was adduced conclusively by Dr. McDonald in seven endogenous elections he analyzed. (See McDonald Decl. ¶ 23.) The defendants conténd that the plaintiffs have failed to show polarization that is attributable to anything other than ordinary partisan politics, which we discuss in the next section of this decision. We have already described the relatively limited nature of the evidence about racial polarization in the Bronx. The evidence is certainly insufficient for us to conclude here that there is “a persistent and significant degree of racial polarization in” ■ Bronx State Senate elections.

The defendants’ expert, Dr. Stanley, argues that any patterns of racial polarization are attributable primarily to partisanship. He contends that “there is more disagreement between white and minority voters in white/white elections than in elections involving minority candidates. If race was a factor in voting, one would expect to see just the opposite.” (Id. ¶ 40.) In all general (endogenous and exogenous) elections in the Bronx, he .observed that white support for Democrats averages 42.9% in minority-white .elections and 39.3% in white-white elections. (Id. ¶ 41; DX 170.) Among Hispanics in the same elections, support for Democrats averages 87.7% in minority-white elections and 91.6% in.white-white elections. (Id. ¶ 42; DX 170.) Among blacks in the same elections, “support for Democrats averages 90.9% in minority-white elections and 91.9% in white-white elections.” (Id.) And in Democratic primaries, “where partisanship is not a factor ...[,] there is no pattern of agreement between blacks and Hispanics that exceeds the extent to which whites and Hispanics agree.” (Id.; DX 154.) “In short,” Dr. Stanley concluded,

Dr. McDonald’s “evidence” of racial bloc voting consists of the fact that whites in Senate elections supported well-entrenched- Republican incumbents who defeated all challengers, white or minority, and the fact that Governor Pataki, another popular incumbent who was easily reelected, defeated Carl McCall in the 2002 gubernatorial election. The Senate election contests reveal no race-based disadvantage since the incumbents defeated all opponents of any race or ethnicity. =

(Id. ¶ 45 (summarizing racial bloc voting analysis with respect to both Nassau and Bronx).) Dr. Stanley’s conclusion with respect to the voting patterns in Nassau County and the Bronx was the same: “[A]ny divergence between white and minority voters, as well as any cohesion between black and Hispanic voters, is better correlated with partisanship than with race.” (Id. ¶39.)

The plaintiffs say that the fact that divergent voting patterns may logically be explained by a factor other than race does not end the inquiry because racially polarized voting and partisanship are not mutually exclusive. See Goosby v. Town Bd., 956 F.Supp. 326, 355 (E.D.N.Y.1997), aff'd, 180 F.3d 476 (2d Cir.1999). They argue that partisanship, does not explain racial polarization better than race and that the evidence establishes that white voters in the endogenous elections analyzed support the minority-preferred candidate when that candidate is white at an average rate of 33.3%, but when the minority-preferred candidate is a minority (in only one contest), the support level drops to 27.0%. (PX 73.) With respect to primaries, the plaintiffs say that “the 2001 Democratic primaries ... show[ ] a strong pattern of racial polarization, and primary results cannot be explained by partisanship.” (McDonald Decl. ¶ 54.)

Neither party to this litigation has, in addition to analyzing racial polarization, employed regression analysis to determine the relationship between voter support and party registration in election districts, and the correlations offered by both sides’ experts based on El results are inconclusive. While the plaintiffs’ evidence is restricted to the more probative endogenous elections, it is also based on only one election (of seven) that had a minority candidate. At the same time, we cannot say that the defendants have established conclusively that partisanship is a better predictor of voting patterns than is race.

(d) Official Discrimination

The plaintiffs argue that “[t]here is a history of official discrimination” particularly in the Yonkers portion of the plaintiffs’ proposed District 36. (PFF at 173.) In support of their position, the plaintiffs urge us to take judicial notice of the United States v. Yonkers public school litigation and state that the defendants should be “collaterally estopped from challenging the detailed findings of fact made” in that case. See United States v. City of Yonkers, 96 F.3d 600, 616 (2d Cir.1996); see also Yonkers, 123 F.Supp.2d 694, 706 n. 20 (S.D.N.Y.2000) (“[T]here are policies and practices in Yonkers that disproportionately impact minority students and are traceable to the prior segregation, [and therefore] we conclude that those policies and practices are vestiges of discrimination.”). The plaintiffs contend that the findings in Yonkers are relevant because the “area served by the Yonkers school district ... is completely contained in” Plaintiffs’ Proposed District 36 and is “demographically very similar to the rest of’ that district. (See Beveridge Deck ¶ 154).

The defendants assert that the record confirms that “New York has ensured African-Americans the right to vote since 1874” and “has taken steps to encourage minority voting .... ” France, 71 F.Supp.2d at 330. “[T]he state of New York has long been in the vanguard of the struggle for African American civil rights.” Reed v. Town of Babylon, 914 F.Supp. 843, 886 (E.D.N.Y.1996). The defendants cite thirteen stipulations agreed to by the parties in this case to demonstrate the ways in which New York State and New York City have sought to ensure equality of opportunity. (See Stip. ¶¶ 118-30 (noting, among other things, the enactment of laws forbidding discrimination in state employment (1918) (¶ 119) and creation of the New York City Commission on Human Rights (1955) (¶ 125)).) The defendants also argue that the most recent findings by the district court in Yonkers in 2000 were never reduced to final judgment because the case was settled prior to the issuance of a final judgment (or appeal), and that those findings are not the proper subject for collateral estoppel effect. 144

We are familiar with the Yonkers decision and keenly aware of the troubling history and the practices it has sought to ameliorate. At the same time, over two-thirds of Proposed District 36 comes from the Bronx, and we are not persuaded that the plaintiff has shown official discrimination in the Bronx that touches on or infringes the opportunities of minority voters. Indeed, the evidence presented at trial suggests that Hispanics^ — and the various ethnic groups that fall under that label — have very actively participated in local Bronx politics. (See, e.g., Tr. at 559 (Paterson) (noting that Bronx has had Puerto Rican borough presidents going back to the 1980s); see also id. at 509-14).

Therefore, while we find the history reflected in Yonkers troubling and recognize the need for continued vigilance and resolve to ensure equality of opportunity, we do not find in the record a “history of voting-related discrimination in the State.” Gingles, 478 U.S. at 44, 106 S.Ct. 2752 (citing S.Rep. No. 97-147, at 28, 1982 U.S.C.C.A.N. at 205-06).

(e) Socioeconomic Disparities

The plaintiffs assert that “[t]he social condition of African American and Hispanic residents with respect to income, jobs, education, housing, and in many other ways, is substantially deprived compared to that of the white population in ... Bronx County, in Westchester County, and in the City of Yonkers.” (Beveridge Decl. ¶ 123.) They contend that these disparities “manifest [themselves] in every aspect of life, plainly affecting their ability to participate in the political process.” (Id. ¶ 123.) The plaintiffs assert that “the political process itself continues to deny [Hispanics and blacks] their ‘fair share’ of funding for schools and other programs that might serve to mitigate this deprivation.” (Beveridge Decl. ¶ 123.) In addition to the Yonkers litigation, the plaintiffs cite Campaign for Fiscal Equity, Inc. v. State (“CFE”) to demonstrate that New York City’s public school system does not provide students with a “sound basic education” that “prepares them to function productively as civic participants.” See CFE, 100 N.Y.2d 893, 769 N.Y.S.2d 106, 801 N.E.2d 326, 332 (2003) (Kaye, C.J.). 145

The defendants contend that the plaintiffs have not shown that socioeconomic disparities between whites , and minorities are significant or that they are the result of official discrimination. They assert that “the record shows that socioeconomic disT parities exist throughout the country and also that the status of minorities in the areas at issue ... is well ahead of minorities elsewhere ...: ” (DCL ¶ 81.) They cite to Goosby, 956 F.Supp. at 342, for the proposition that “differences in the socioeconomic status of blacks ... do not significantly impair their relative ability to participate in the political process,” and to Niagara Falls, 65 F.3d at 1021, for the conclusion that “socioeconomic disadvantages have not hindered blacks’ ability to participate in the political process.”'

It is certainly reasonable to observe that to some extent minorities in New York “bear- the effects of past discrimination,” Gingles, 478 U.S. at 44-45, 106 S.Ct. 2752 (citing S.Rep. No. 97-147, at 28-29, 1982 U.S.C.C.A.N. at 205-207), as reflected in the CFE decision with which we are familiar. At the same time, there is not a substantial or adequate showing that the socioeconomic status of minorities significantly impairs their ability to participate in the political process in the relevant geographical area.

(f) Electoral Mechanisms

The plaintiffs argue that several New York State and City electoral mechanisms have “disproportionately affected” minority voters. (Hayduk Decl. ¶ 4.) They cite, among other things, to funding cuts for the State agency charged with implementing the National Voter Registration Act (“motor voter law”) (id. ¶ 6); the alleged failure of New York City public assistance agencies adequately to implement the motor voter law (id. ¶ 8); polling problems such as sites opening late, a shortage of properly trained poll workers, and voting machine “breakdowns” (id. ¶ 10); the (past) practice of purging non-voters from voter rolls (id. ¶ 11); “lost votes” due to the decision of the City Board of Elections to disable sensor latches on voting machines (id. ¶ 20); the City’s failure to provide sufficient numbers of language interpreters at polling stations (id. ¶ 22); and problems affecting particularly Asian voters, such as “[p]oll workers demand[ing] identification from ... voters” (id.).

The defendants contend that the plaintiffs have not produced evidence indicating that “either New York State or the particular parts of it at issue here have ‘unusually large-election districts, a majority vote requirement ... or other voting practices or procedures that might enhance the opportunity for discrimination against the minority group.’ ” (Thernstrom Aff. ¶ 41 (citing Gingles, 478 U.S. at 45, 106 S.Ct. 2752 (citing S.Rep. No. 97-147, at 29, 1982 U.S.C.C.A.N. at 206-05)).) They cite to the measures taken by New York State and its localities in recent decades to expand the electorate (Thernstrom ¶ Aff. 16), including revamping the State’s voter registration forms (id.), the City’s expansion of voter registration sites (id.), and measures undertaken to implement the motor voter law (Tr. at 463). The defendants argue and we agree that the plaintiffs have failed to establish voting irregularities, such as those enumerated in the Senate Report and Gingles, that “enhance the opportunity for discrimination against the minority group.” Gingles, 478 U.S. at 45, 106 S.Ct. 2752 (citing S.Rep. No. 97-147, at 29, 1982 U.S.C.C.A.N. at 206-07).

(g) Other Totality Factors

The plaintiffs argued that, because of the nature of them packing claim, four of the Senate Factors are irrelevant, including: access to the candidate slating process, racial appeals in political campaigns; the number of minorities elected to public office; and the responsiveness of elected officials to minority needs. We believe these factors are relevant and that our analysis suggests that Hispanics in the districts at issue in this claim and the surrounding areas are not denied equal opportunity to participate in the political process. As the defendants rightly point out, there is little or no evidence that the more overtly invidious factors, such as denying minorities access to the slating process and employing racial appeals in campaigns, are present. More importantly, the number of Hispanics elected to state and local office has risen substantially over the past decade (see Stip. ¶¶ 141, 143-45), 146 and Sen. Paterson conceded that Sen. Velella has been responsive, to the needs of black and Hispanic citizens in his district (SD 34) and is well liked in the minority community (Tr. at 519-20).

We therefore conclude, based on the totality of the circumstances, that there is no evidence that Hispanics in the relevant area are prevented from participating in the political process. Moreover, there is evidence that Hispanics have been able to elect candidates of choice and that their political involvement is growing. And even where they have not been able to elect a candidate of choice — such as in SD 34 — the evidence shows that their elected representatives have been responsive. ■

7. Conclusion

For the reasons stated, the plaintiffs have not established by a preponderance of the evidence, under Cingles, that Hispanics in the Bronx “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).

D. The Statewide Vote-Dilution Claim (Count II)

The second count of the Complaint was a “statewide” section 2 vote-dilution claim. The Complaint alleged that the 2002 Senate Plan was a voting practice or procedure that diluted minority voting strength by drawing fourteen majority-minority districts when additional majority-minority districts could have been created. (See Compl. ¶¶ 294-95, 299.) The plaintiffs’ theory was that the 2002 Senate Plan sys-tematieally underpopulated the upstate districts and overpopulated the downstate districts where the vast majority of blacks and Hispanics in the state reside. (See Summ. J. Tr. at 54-55); cf. infra Part III (Count I: one-person, one-vote claim). The remedy offered was the creation of an additional majority-minority district in the Bronx-Westchester region — namely, SD 36. (See Summ. J. Tr. at 55.) 147

We granted summary judgment dismissing this claim because, as the plaintiffs conceded at the summary judgment argument, Count II is duplicative of Count III. (See id. at 54-57.) As explained above, Count III also alleged that the upstate-downstate population distribution was a voting practice that lead to the packing of minorities in the Bronx and the failure to create an additional majority-minority district. Both Counts II and III offered the creation of Proposed District 36 in the Bronx as a remedy. The plaintiffs agreed that the . statewide section 2 claim was duplicative and that their case would not be harmed by dismissing Count II. (Id. at 57.)

E. The LVRC Intervenors: Senate District 31 (Count IV)

In Count IV of the Complaint, the plaintiff-intervenors Latino Voting Rights Committee (“LVRC intervenors”) challenged SD 31, as redistricted following the 2000 census. SD 31 is comprised of sections of the Bronx and Manhattan’s Upper West Side. SD 31’s VAP is approximately 53.2% Latino, 8.0% black, and 33.8% white. (See Braatz Decl. Ex. 24 at 5 (Report of Dr. Rodolfo O. de la Garza, dated July 31, 2003 (“De la Garza Report”).)) 148 SD 31’s CVAP is approximately 40.6% Latino, 10.5% black, and 44.9% white. (See Net-burn Decl. Ex. 4 tbl. 27 (revised).) The district that LVRC intervenors argue should have been created instead, Proposed District 32, substantially overlaps geographically with SD 31 and would have a 61.2% Latino VAP. (See de la Garza Report at 8.) Under Proposed District 32, the black VAP would also be increased to 16.3%. (See id.) 149 The LVRC intervenors claim that, in choosing SD 31 over Proposed District 32, the defendants violated section 2 of the Voting Rights Act.

While the defendants effectively conceded for the purposes of this claim the substance of the first two Gingles preconditions — the majority-in-a-district and cohesion requirements — the LVRC inter-venors have failed to adduce sufficient evidence to meet the third Gingles precondition. They have not presented evidence that would enable them at trial to show that “the white majority votes sufficiently as a bloc to enable it ... usually to defeat the minority’s preferred candidate” in SD 31. Gingles, 478 U.S. at 50-51, 106 S.Ct. 2752. In addition, they have failed to offer a reasonable standard against which we could usefully measure any dilutive effect of SD 31.

The LVRC intervenors’ claim cannot survive summary judgment because the information provided by their expert is incomplete and insufficient to address specific aspects of the third Gingles factor. The racial bloc voting report prepared by the plaintiffs’ expert provides the share of votes given by Latinos and “non-Latinos” to the candidates in each election analyzed. (See generally de la Garza Report at 11-13.) However, the report does not give the total vote shares for the candidates. That is, the report does not explain who won in SD 31 for each election. It does not state whether the Latino-preferred candidates were “defeated” within the districts. Even if there were racially polarized voting, the report does not speak— one way or the other — to the effects of the polarized voting.

This omission in the analysis is compounded by others. For example, the LVRC intervenors provided no estimates of the turnout rates of the various groups of voters. Without turnout rates, it is impossible to deduce the results in SD 31 for the various elections. More generally, turnout-rate information is relevant to whether whites' — despite not being a majority of the VAP or CVAP in SD 31 — can “outvote” Latinos with the “usual” regularity required to satisfy third Gingles precondition. The LVRC intervenors also failed to provide estimates of how blacks vote in SD 31 (or would vote in Proposed District 32). Instead they simply grouped blacks and whites as “non-Latinos.” Knowing how blacks vote separately from whites is also relevant to determining whether white bloc voting, as opposed to black and white combined voting, usually defeats the Latino candidate in SD 31. In sum, the LVRC intervenors’ analysis examines racially polarized voting without addressing the specifics of the third Gin-gles factor, which requires white majority bloc voting that usually defeats the Hispanic-preferred candidate. See Gingles, 478 U.S. at 49, 106 S.Ct. 2752. The failure to provide the necessary information warrants dismissing the LVRC intervenors’ claim on summary judgment. It is their burden to adduce evidence sufficient to satisfy the third Gingles precondition, and they plainly would not be able to prove their claim at trial based on the evidence provided. See Gingles, 478 U.S. at 46, 106 S.Ct. 2752.

The LVRC intervenors’ theory is that Latino candidates are consistently defeated in the primary stages of elections, thus forcing Latinos to vote for their second preference in the general elections. Having framed the inquiry as such, Dr. de la Garza based his analysis primarily on local exogenous primary elections and one endogenous primary election — all involving a Latino candidate. (See id. at 11-14.) Although endogenous elections are the generally considered most probative for racial bloc voting analysis, the evidence is largely limited to exogenous elections. See, e.g., NAACP v. City of Niagara Falls, 65 F.3d 1002, 1015 n. 16 (2d Cir.1995); Clay v. Bd. of Educ., 896 F.Supp. 929, 940 (E.D.Mo.1995). 150

To show racial bloc voting against Hispanic-preferred candidates, Dr. de la Garza’s report first cites the 1993 Democratic primary for Public Advocate, which involved five candidates. (See de la Garza Report at 11, 13 tbl. 6.) For that election, the purportedly “Latino-preferred” candidate, Puerto Rican Roberto Ramirez, commanded the plurality support of 42% of the Latino vote in SD 31, and Mark Green was the “non-Latino preferred candidate,” receiving 63% of the non-Latino vote. (Id.) While the election shows some signs of racially polarized voting, the LVRC inter-venors’ labeling of Ramirez as the “Latino candidate of choice” is highly questionable. Not only did 24% of Latinos in SD 31 support Green, but also a full 58% of Latinos in SD 31 supported candidates other than Ramirez. Therefore, it is difficult to say that Ramirez’s defeat in SD31 — assuming he was defeated 151 — is attributable to white bloc voting rather than the lack of cohesive support from Latinos. The ambiguity of this election is only increased by the failure to provide information such as the ultimate results in the district, the race/ethnicities of the other candidates, and the rates at which blacks and whites separately supported the different candidates.

The second election that the LVRC in-tervenors offer in support of their claim is the 1993 Democratic primary for City Comptroller. (See id. at 11, 13 tbl.7.) In that primary, Herman Badillo, a Puerto Rican who, at various times during his political career, has been both a Republican and a Democrat, received -55% of the Latino vote. (Id.) The non-Latinos split their vote; 43% favored Alan Hevesi, who was ultimately successful citywide in the primary and general election, 28% favored Elizabeth Holtzman, and 27% favored Ba-dillo. (See id. at 11,13 tbl.7.) We note that 31% of Latinos favored Holtzman in that race as well. In sum, this election also does not show any clear evidence of racial bloc voting defeating the Latino-candidate of choice in SD31. As was true in the Public Advocate primary for SD 31, Latinos did not coalesce around their “candi- daté of choice” in overwhelming numbers, and non-Latinos supported Badillo at a rate equal to another white candidate. Moreover, we do not even know that Badil-lo was “defeated” within SD 31. Dr. de la Garza’s report does not provide the results for SD 31 nor does he provide turnout rates that would allow us to deduce the results. This is a glaring example of how incomplete information provided by the LVRC intervenors undermines their..ability to establish their claim.

The same omission vitiates the probative value' of the 2001 mayoral primary and mayoral runoff-elections involving Ferrer, a Hispanic candidate, and Green. (See id. at 11-12, 13 tbl.9, 14 tbl.10.) The LVRC intervenors suggest that voting within SD 31 was racially polarized for those elections because Latinos overwhelmingly favored Ferrer and non-Latinos favored Green. But again, Dr. de la Garza fails to report whether Ferrer was “defeated” by Green in SD 31. Therefore, these elections, on their face, are not evidence that white bloc voting usually defeats Latino candidates of choice in SD 31.

The only probative evidence of racially polarized voting defeating the Latino candidate of choice comes from the 2002 Deim ocratic primary race for State Senator for SD 31. In that election, Guillermo Li-nares, a Dominican Latino and former City Councilman who, had represented the Washington Heights area for over ten years (see Chin Aff. ¶ 29), received the support of 81% of the Látino vote. (De la Garza Report at 12, 14 tbl. 11.) But Eric Schneiderman, a white incumbent State Senator, 152 won the election-with 88% of the non-Latino vote. (Id.) Putting aside Schneiderman’s incumbency advantage and the fact that this election took place post-litigation, to the extent that this single election manifests racially polarized voting defeating the Latino-preferred candidate, it is insufficient to raise a genuine issue of material fact that would enable the LVRC intervenors to satisfy the third Gingles precondition.

Moreover, the LVRC intervenors sought in the first instance to prove a proposition that is insufficient to show that the third Gingles factor had been met: the LVRC intervenors’ expert, Dr. de la Garza prepared his report to answer the question, “[C]an Latinos elect their candidates of choice when their most preferred candidate is a Hispanic candidate? ” (Id. at 5 (emphasis added).) The answer to the question posed by the report is only part of the picture. It is also important to know whether Latinos can elect their candidates of choice when their preferred candidate is not Hispanic. This is so because in New York Latinos do not always prefer Hispanic candidates, even when Hispanics are in the race. For example, Dr. de la Garza’s report mentions but discounts the Democratic mayoral primaries of 1993 and 1997, where, in each, there was a Latino candidate,.Eric Ruane-Melendez, who was not preferred by the Latino electorate. (See de la Garza Report at 11, 12 tbl. 5, 13 tbl. 8.) In 1993, the Latino-preferred candidate was a black candidate, David Dinkins; in 1997, the Latino-preferred candidate was a white candidate, Ruth Messinger. Both Dinkins and Messinger received higher percentages of the Latino vote than the presumptive “Latino-preferred” candidates in the Public Advocate and the City Comptroller races, which were offered to show racial bloc voting. While the LVRC intervenors assert that in “a few anomalous] races ... both the Latino VAP and non-Latino VAP majorities supported white candidate[s]” (see Chin Aff. ¶ 30), they do not explain why these elections are anomalous.

In sum, the LVRC intervenors asked the wrong question and provided insufficient and incomplete evidence in answering it. The question is not whether Latinos can elect a preferred candidate who is Hispanic, but a preferred candidate period. If Latinos often prefer non-Latino candidates, LVRC intervenors cannot meet the third prong of Gingles by showing only that racial bloc voting defeated a Latino candidate of choice in one instance when that candidate of choice was Latino. Even being as generous as possible to LVRC intervenors, despite the fact that it is their burden to show the third Gingles factor, no reasonable person could conclude from these analyses that a white majority usually defeats the minority’s preferred candidate as required by Gingles. See Gingles, 478 U.S. at 49, 106 S.Ct. 2752.

Finally, regardless of the insufficiency of their racial bloc voting analysis, the LVRC intervenors’ claim must be dismissed because the LVRC intervenors have failed to provide “a hypothetical alternative” that would enable us to assess whether the 2002 Senate redistricting plan “results in a denial or abridgement of the right of any citizen ... to vote on account of race or color.” Reno v. Bossier Parish Sch. Bd., 528 U.S. 320, 334, 120 S.Ct. 866, 145 L.Ed.2d 845 (2000). LVRC intervenors propose only a single district that they claim the Senate should have drawn as part of a 61-seat plan. They do not propose an alternative redistricting plan for the entire state or propose an alternative with a 62-seat plan in mind. Here, it was incumbent upon them to have done so.

To be sure, the plaintiffs in a lawsuit such as this need not perfectly design an airtight alternative that is beyond reproach; they need only present a “reasonable” idea as a baseline. See, e.g., Holder v. Hall, 512 U.S. 874, 880, 114 S.Ct. 2581, 129 L.Ed.2d 687 (1994). And generally, when the challengers, as here, propose a single alternative district, it enables a court to assess whether the first two Gin-gles preconditions have been met; whether the new district would contain a majority of minority members; and whether there would be cohesion among the minority voters in the newly proposed district. See, e.g., Sanchez v. State of Colorado, 97 F.3d 1303, 1314 (10th Cir.1996). However, we agree with the defendants that in this case, whére the lead plaintiffs challenge many different aspects of the Senate Plan and provide an alternative of their own — one that does not embrace the intervenors’ alternative — the hypothetical alternative offered by the intervenors must be more complete than the offering of a single district to serve as the relevant baseline for their Section 2 claim. The main plaintiffs’ Revised Plan in this case does not adopt Proposed District 32 as part of their hypothetical alternative, and their main expert, Dr. McDonald, does not himself include SD 31 in his list of districts that reflect racially polarized bloc voting. Thus, here, where LVRC intervenors proposed only a single alternative district and their expert performed his analyses by comparing only the current SD 31 and Proposed District 32 without attention to how other adjacent districts and the 62-seat scheme would be affected, we cannot entertain the claim.

For all the foregoing reasons, we granted summary judgment dismissing Count IV.

F. The Rivers/Barbour Intervenors: Congressional District 17 (Count VIII)

In Count VIII of the Complaint, plaintiffs-intervenors Rivers/Barbour (“Rivers/Barbour intervenors”), along with the Council of Black Elected Democrats (“COBED”), challenge New York State’s redistricting of Congressional District 17 (“CD 17”) following the 2000 census. The district is composed of portions of the Bronx, Westchester, and Rockland Counties. The Rivers/Barbour intervenors propose an alternative Congressional District 17 (“COBED-CD 17”).

In the current CD 17, the black VAP is approximately 29.32%; the Hispanic VAP is 18.74%; the Asian VAP is 4.64%; and the white VAP is 44.06%. (See Hedges Aff. ¶ 8.) In COBED-CD 17, the Rivers/Barbour intervenors argue that the black VAP would be 37.2%; the Latino VAP would be 30.2%; the Asian VAP would be 4.3%; and the white VAP would be 28.0%. 153 (COBED’s Rule 56.1 Counterstatement ¶ 360.) The Rivers/Barbour intervenors urge that COBED-CD 17 would provide blacks and Hispanics with more of an opportunity' to elect a candidate of their choice and that in choosing CD 17 over COBED-CD 17 the defendants violated section 2 of the Voting Rights Act.

The defendants moved for summary judgment on this claim, arguing, among other things, as follows: (1) The law of the case requires our rejection of this count because we previously approved a very similar CD' 17 developed by our Special Master, Fredrick Lacey, before his plan was superceded by the legislature’s enactment. The law of the case doctrine “posits that when a court decides upon a rule of law, that decision should continue to govern the same issues in subsequent stages in the same case.” Arizona v. California, 460 U.S. 605, 618, 103 S.Ct. 1382, 75 L.Ed.2d 318 (1983).(2) The COBED-CD 17 remedy would violate Shaw v. Reno, 509 U.S. 630, 113 S.Ct. 2816, 125 L.Ed.2d 511 (1993) because the shape of the proposed district would be irregular and drawn for predominantly racial reasons. (3) The Rivers/Barbour intervenors cannot meet the second Gingles precondition for their claim because (a) neither blacks nor Hispanics would constitute a majority unto themselves and (b) they may not be grouped together for Section 2 analysis because the two separate minorities cannot be shown to vote cohesively as one.

We decline to rule on the first two of the defendants’ arguments because we find the defendants’ third argument to be disposi-tive and to require that this count be dismissed on summary judgment. It is undisputed that under COBED-CD 17 no single minority group would constitute a majority of the population. Instead, the Rivers/Barbour intervenors propose to meet the first Gingles requirement by aggregating the VAPs of blacks and Hispanics within COBED-CD 17. This claim thus raises the same minority-coalition claim issues raised by the Bronx vote-dilution claim. See Bridgeport Coalition for Fair Representation v. City of Bridgeport, 26 F.3d 271, 279 (2d Cir.) (accepting evidence of black and Hispanic cohesion but noting that “[i]t is preferable to create majority districts for each [minority] group”), vacated on other grounds, 512 U.S. 1283, 115 S.Ct. 35, 129 L.Ed.2d 931 (1994); Brewer v. Ham, 876 F.2d 448 (5th Cir.1989). As we did in Count III, while we assume that minority groups may be aggregated to meet the Gingles preconditions, the plaintiffs-intervenors must show that the minority groups are politically cohesive for voting purposes. See Growe v. Emison, 507 U.S. 25, 41, 113 S.Ct. 1075, 122 L.Ed.2d 388 (1993). Indeed, the Supreme Court has instructed that, when voting rights claims are based on a combination of distinct ethnic and language minority groups, “proof of minority political cohesion is all the more essential” and must be held to a “higher-than-usual” standard. Id; see also Page v. Bartels, 248 F.3d 175, 197 (3d Cir.2001). However, the Rivers/Barbour intervenors, like the plaintiffs in other claims, have failed to meet their burden in showing that blacks and Hispanics are politically cohesive in CD 17 or COBED-CD 17. They have thus failed to adduce evidence sufficient to withstand summary judgment. 154

In this case, the Rivers/Barbour interve-nors sought to make their prima facie case through their expert, Frank Lewis, who analyzed four interracial Democratic primaries: the 1997 New York City primary for Mayor, the 2000 Congressional primary for CD 17, the 2001 primary for Mayor, and the 2001 primary for City Comptroller. (See Chill Aff. Ex. O (“Lewis Report”).) The Rivers/Barbour intervenors summarize their results as follows: “In each of these elections, Blacks and Hispanics in statistically material numbers voted for the same or another minority preferred candidate, rather than ... the candidates preferred by the White voting bloc within Congressional District 17.” (See Gibbs Aff. ¶ 59.)

Even accepting this conclusion, we find that Rivers/Barbour intervenors still fail to demonstrate cohesion, much less the high standard of political cohesion required for minority-coalition districts. To be cohesive, blacks and Latinos must have the same preferred candidate; it is insufficient that they, “in statistically material numbers” (whatever that may mean), “vote[] for the same or another minority preferred candidate,” (See id (emphasis added).) Plainly, there is no black-Hispanic cohesion if blacks tend to vote for black candidates and against Hispanic candidates, and if Hispanics vote for Hispanic candidates and against black candidates. It cannot be sufficient for the purposes of cohesion merely that both groups vote against whichever candidate is preferred by whites if the two groups do not generally vote cohesively for the same candidate. 155

The Rivers/Barbour intervenors’ insufficient conclusions were the only ones supported by the data produced by their own expert, Frank Lewis. In the 1997 mayoral primary, 62.7% of blacks and 30.3% of Hispanics supported the black candidate, Reverend Al Sharpton. (See Lewis Report at 5.) However, the crucial fact that the Rivers/Barbour intervenors omit (see Gibbs Aff. ¶ 60) is that 48.5% of Hispanic voters supported the white candidate, Ruth Messinger, who was also supported by 52.5% of white voters. (Lewis Report at 5.) In the CD 17 Democratic primary in 2000, 48.3% of blacks and 30.3% of Hispan- íes supported black candidate Larry Sea-brook. (Id.) However, 49.5% of Hispanics supported the white candidate, Eliot En-gel, who was supported by 86.4% of white voters. Moreover, while 24.8% of Hispanics supported Zayas, the Hispanic candidate, only 10.8% of blacks supported Zayas compared to 40.9% of blacks who supported Engel. In the 2001 City Comptroller race, 81.7% of black voters and 38.7% of Hispanic voters supported a black candidate, William Thompson. (Id.) Again, the Rivers/Barbour intervenors neglect to mention that 61.3% of the Hispanic voters supported the white candidate, Herbert Berman, who was supported by 69.6% of white voters. (Id.) 156

The only election analysis to show some level of black-Hispanic cohesion is the 2001 mayoral primary in which the Hispanic candidate, Fernando Ferrer, received the support of 62.7% of black voters and 74.1% of Hispanic voters. (Id.) One election, however, cannot establish cohesion among ethnic blocs; the possibility that Ferrer was simply the better candidate looms too large and the strong support from both of the minority groups forced a run-off primary between Ferrer and the ultimate winner, Mark Green, who commanded only 31.4% of white support in the five-contestant race. (Id.) We do not consider this limited evidence sufficient to allow the Rivers/Barbour intervenors to survive summary judgment. Three out of four elections show Hispanics preferring white candidates over black candidates. Larry Seabrook — a member of COBED’s executive committee, who since 1984 has represented part of CD 17 as an Assemblyman, a State Senator, and a City Councilman, and who lost the 2000 primary race to incumbent Engel — has explicitly conceded that there is no cohesion between blacks and Latinos in CD 17, stating at his deposition that “traditionally Hispanics have basically voted for white candidates, traditionally.” (See Chill Aff. Ex. P at 80-81 (Tr. of Dep. of Larry B. Seabrook, dated Sept. 11, 2003).) The absence of cohesion sufficient to satisfy Gingles’s second factor is fatal to the Rivers/Barbour intervenors’ minority-coalition claim regarding CD 17. Without aggregating the black and Hispanic populations, the intervenors cannot satisfy the first Gingles precondition.

Accordingly, for the foregoing reasons, Count VIII. was dismissed on summary judgment.

V. EQUAL PROTECTION CLAIM: SENATE DISTRICT 34 (COUNT VII)

The plaintiffs allege that, in contravention of Shaw v. Reno, 509 U.S. 630, 113 S.Ct. 2816, 125 L.Ed.2d 511 (1993) (“Shaw I”), SD 34 was drawn in 2002 “predominantly on the basis of race,” and that the Legislature intended to create “an overwhelming non-Hispanic white majority” district in violation of the Fourteenth Amendment. (Compl.lffl 330-31, 334.) The plaintiffs further allege that in drawing SD 34 in 2002 the Legislature subordinated or disregarded traditional districting principles. (Id. ¶ 333.) Alternatively, throughout their submissions, the plaintiffs suggest that the alleged racial gerrymander of SD 34 first occurred in 1992 and was continued or preserved in 2002, although they also acknowledge it has never before legally been challenged on Shaw grounds. The plaintiffs point to SD 34’s “non-compact” shape and to several documents, including, among others, two 1992 court filings in Wolpoff v. Cuomo, 80 N.Y.2d 70, 587 N.Y.S.2d 560, 600 N.E.2d 191 (1992), as evidence of a racial gerrymander. (ComplJ 332.) As set forth in Count III of their Complaint, the plaintiffs’ proposed solution is to dismantle majority-white SD 34 in order to create a fifth (out of five) majority-minority coalition district located primarily in the Bronx whose voting age population would be 50.1% Hispanic (45.4% CVAP), 23.0% black, and 19.1% white. (See PX 6 tbl. 16.)

The defendants respond persuasively that the plaintiffs have not met the heavy burden of establishing that the New York State Legislature drew SD 34 predominantly on the basis of race. Rather, the record clearly shows that (i) relatively minor changes were made to SD 34 in 2002, principally the addition of the contiguous town of Eastehester in Westchester County; (ii) the configuration of SD 34 in 2002 is largely the same as it was in the 1990s (and the 1980s), with 87.7% of the 2002 SD 34, 641 N.W.2d 656 population coming from 1992 SD 34 (DX 106); and (iii) those changes that were made in 2002 enabled the Legislature to achieve substantial population equality, maintain the core of SD 34’s predecessor district, protect the position of SD 34’s 16-year incumbent state senator, avoid incumbent pairing, comply with the Voting Rights Act, and help Republicans retain their nearly 40-year-long Senate majority. 157

A. Applicable Law

“[A] plaintiff can challeng[e] a reapportionment statute ... by alleging that the legislation, though race neutral on its face, rationally cannot be understood as anything other than an effort to separate voters into different districts on the basis of race, and that the separation lacks sufficient justification.” Shaw, 509 U.S. at 649, 113 S.Ct. 2816. A Shaw plaintiff bears the burden of proving that racial considerations were “the ‘predominant factor’ motivating the legislature’s districting decision,” Hunt v. Cromartie, 526 U.S. 541, 551, 119 S.Ct. 1545, 143 L.Ed.2d 731 (1999) (“Cromartie I”) (citations omitted), not simply “a motivation for the drawing of a majority-minority district,” Bush v. Vera, 517 U.S. 952, 959, 116 S.Ct. 1941, 135 L.Ed.2d 248 (1996). The plaintiffs’ burden is a “demanding” one. Miller v. Johnson, 515 U.S. 900, 928, 115 S.Ct. 2475, 132 L.Ed.2d 762 (1995) (O’Connor, J., concurring) (“I understand the threshold standard the Court adopts ... to be a demanding one.”). Indeed, “[t]o invoke strict scrutiny, a plaintiff must show that the State has relied on race in substantial disregard of customary and traditional districting practices.” Chen v. City of Houston, 206 F.3d 502, 506 (5th Cir.2000) (citing Miller, 515 U.S. at 928, 115 S.Ct. 2475 (O’Connor, J., concurring)). The court in Chen also observed that “the Supreme Court does not believe that the mere presence of race in the mix of decision making factors, and even the desire to craft majority-minority districts, ... alone automatically trigger[s] strict scrutiny.” Id. Where race and politics overlap, a plaintiff has an additional burden of proof. That is,

[i]n a case ... where majority-minority districts (or the approximate equivalent) are at issue and- where racial identifica tion correlates highly with political affiliation, the party attacking the legislatively drawn boundaries must show at the least that the legislature could have achieved its legitimate political objectives in alternative ways that are comparably consistent with traditional dis-tricting principles. That party must also show that those districting alternatives would have brought about significantly greater racial balance.

Easley v. Cromartie, 532 U.S. 234, 258, 121 S.Ct. 1452, 149 L.Ed.2d 430 (2001) (“Cromartie II ”). A plaintiffs heavy burden of establishing the predominance of race can be met either through direct evidence of the Legislature’s purpose or through circumstantial evidence, including, among other things, a district’s demographics or shape. See Shaw v. Hunt, 517 U.S. 899, 905, 116 S.Ct. 1894, 135 L.Ed.2d 207 (1996) (“Shaw II”); Miller, 515 U.S. at 916, 115 S.Ct. 2475.

If a plaintiff meets its heavy burden of proving racial “predominance,” the challenged district is subject to “strict scrutiny,” which means that the district must be shown to be “narrowly tailored to further a compelling state interest.” Vera, 517 U.S. at 976, 116 S.Ct. 1941; see id. at 977, 116 S.Ct. 1941 (“A § 2 district that is reasonably compact and regular, taking into account traditional districting principles ... may pass strict scrutiny without having to defeat rival compact districts designed by plaintiffs’ experts in endless ‘beauty contests.’ ”).

Legislatures are afforded the benefit of a presumption of good faith when they conduct redistricting. See Chen, 206 F.3d at 505 (“[Our] review must be understood in the context of the courts’ traditional reluctance to interfere with the delicate and politically charged area of legislative redistricting. The [Supreme] Court has clearly indicated that th[e] presumption [of good faith] may impact the assessment in a suit challenging districts as racial gerrymanders.” (citations omitted)); see also Miller, 515 U.S. at 916, 115 S.Ct. 2475 (“The distinction between being aware of racial considerations and being motivated by them may be difficult to make. This evidentiary difficulty, together with the sensitive nature of redistricting and the presumption of good faith that must be accorded legislative enactments, requires courts to exercise extraordinary caution in adjudicating claims that a State has drawn district lines on the basis of race.”).

Preliminarily, we note that this is an unusual claim because Shaw challenges have almost always been directed against majority-minority districts. 158 Indeed, the parties here have not analyzed cases challenging majority-white districts, although there appears to be at least one such case on point. In Chen, 9 F.Supp.2d 745 (S.D.Tex.1998), aff'd, 206 F.3d at 521-22, the court dismissed on summary judgment a claim challenging a majority-white Houston city council district which had been adopted in 1997 and which allegedly “perpetuated” a racial districting structure that was originally enacted in 1991. See Chen, 206 F.3d at 514-15, 521-22 (“[W]e cannot say that plaintiffs have carried their burden and made a sufficient showing, either circumstantially or directly, to sustain a finding that race predominated in the districting.”). We certainly have found no successful Shaw challenge to a district, such as SD 34 here, which has been a majority-white district for over twenty years and which has been served continuously by the same Republican incumbent for the past sixteen years. 159

1. Senate District 34

SD 34 is the only majority-white district of the five districts located entirely or predominantly in the Bronx. Its total population is 311,260. Its VAP is 59.5% white; 20.4% Hispanic; and 13.3% black. SD 34 has been represented by the same (white) Republican senator, Guy Velella, for the past 16 years — interesting, perhaps, because SD 34 is a predominantly Democratic district in terms of voter enrollment: 50.2% Democrat to 25.2% Republican. 160 Sen. Velella is a senior member of the State Senate and holds the positions of Senior Assistant Majority Leader and Liaison to the Executive Branch, and Chairman of the Senate Standing Committee on Labor. (Burgeson Aff. ¶ 38.)

The geographic core of SD 34 is located in the eastern portion of the Bronx where 70.1% of the voting-age population resides, the balance being in southern Westchester County and parts of western Bronx. SD 34 extends from the southeastern portion of the Bronx to Throgs Neck (where Sen. Velella lives) (Tr. at 919), Country Club, City Island, and Pelham Bay Park, and into Westchester County, where it picks up the towns of Pelham and Eastchester and parts of the cities of New Rochelle, Mount Vernon, and Yonkers. It re-enters the Bronx from the north around Van Cortlandt Park, picking up Bronx areas such as (parts of) Riverdale, Woodlawn, and Belmont. See N.Y. State Law § 124 (McKinney 2002); (Beveridge Decl. ¶ 61). 161

During the 1980s and into the 1990s, another of the five Bronx districts had a predominantly white voting age population. It was represented by a white (Democratic) senator, Abraham Bernstein, who held office from 1961 until his death in 1990. 162 Jeffrey Korman, another white Democrat, succeeded Sen. Bernstein for the next two-and-one-half years. In the redistricting following the 1990 census, this district was eliminated in order to create a third Bronx Hispanic-majority district (Tr. at 936 (Burgeson)), which left SD 34 as the sole Bronx majority-white district. In brief, the history of the other Bronx districts is as follows. Old SD 30 was represented by a Hispanic (Democratic) senator, Olga Mendez, who was first elected in 1978 and remains in office today. Old SD 32 was represented by a Hispanic (Democratic) senator, Israel Ruiz, from 1975 until 1989. He was succeeded by a Hispanic (Democratic) senator, Efrain Gonzalez, who remains in office today. Old SD 31 (thereafter called SD 33) was represented by a black (Democratic) senator, Joseph Galiber, from 1969 until his death in 1995. He was succeeded by black (Democratic) senators Larry Seabrook and Ruth Has-sell-Thompson,-who holds office today.

The shape of SD 34, certainly not perfectly compact, appears to have evolved relatively slowly since the 1980s. 163 Following the 1982 rédistricting, SD 34 remained predominantly in the Bronx, but, for the first time, crossed the Bronx border into Westchester County to include the Westchester towns of Pelham and portions of New Rochelle, Mount Vernon, and Yonkers. The move of SD 34 into West-chester in 1982 contributed to the current curved (or circular) appearance of the northern portion of the district. Following the 1992 redistricting — when the majority-minority black district (old SD 33 and currently SD 36) also moved north of the Bronx border into Westchester County' — portions of SD 34, including predominantly black southern Mount Vernon, were added to SD 36. As a result, in 1992, SD 34 made a second cróssing into the Bronx from Westchester County (to the west) to include Bronx sections of Van Cortlandt Park and parts of Riverdale, Woodlawn, Wakefield, Norwood, and Belmont, although, in fact, SD 34 at the time had already included sections of the west Bronx along with the Westchester towns of Pelham and portions of New Rochelle, Mount Vernon, and Yonkers. This change appears further to have accentuated the curved shape of the district.

2. 2002 Redistricting

The principal population and geographic changes that were made to SD 34 in 2002 are as follows: (1) A net total of 13,889 people were removed from SD 34 to achieve its target redistricting population of 311,260, with the general effect of increasing the white VAP. Approximately 29,528 (nonwhite) persons were moved into other districts in the Bronx and Manhattan (SDs 31, 32, 33, and 36) and into Westchester (SDs 35 and 37), and approximately 15,639 (white) persons were added to SD 34 from other districts in the Bronx and Manhattan (SDs 3(3, 31, 32, 33, and 36) and Westchester (SD 35), including, particularly, the entire Town of Eastchester (population 31,318). 164 During the 1990s, the white VAP of SD 34 dropped from approximately 69% to 53%. (Burgeson Aff. ¶ 45) SD 34 was redrawn. with a 59.5% white VAP. (See id.) (2) As noted, in 2002, the Westchester Town of Eastchester became part of SD 34. Other geographic changes included the addition of neighborhoods in Yonkers and. the removal of parts of New Rochelle and Yonkers in Westchester, parts of Castle Hill in the southeastern Bronx, and parts of Wakefield, Riverdale, and Norwood, which is in the northwestern Bronx. And (3) the additions of Eastches-ter and parts of Yonkers comprised the vast majority (nearly 98%) of the additions to SD 34 VAP in 2002 — other additions were de minimis. (See DX 106.) Stated in other terms, 87.7% of the population of SD 34 in 1992 remained in SD 34 in 2002. (Burgeson Aff. ¶ 32.)

The 2002 redistricting also impacted the four Bronx majority-minority districts:

a. SD 36, the only Bronx majority-black district, which is adjacent to SD' 34 and which is sometimes referred to as the “hole” in the SD 34 doughnufi’ became in VAP terms: 65.2% black, 25.8% Hispanic, and 6.5% white. Compared to its 1992 predecessor (SD 33) under the 2000 census, the black VAP increased by 3.9 percentage points. 165 “Eighty-eight percent of the new District 36 comes from the old District 33, the majority black district that is and was the companion to District 34.” (Burgeson Aff. ¶ 34.)

b. SD 28, a Bronx/Manhattan district, became in VAP terms: 54.0% Hispanic, 28.8% black, and 12.9% white. Compared to its 1992 predecessor (also SD 28) under the 2000 census, the Hispanic VAP decreased by 7.4 percentage points.

c. SD 32, a south-central Bronx district, became in VAP terms 58.1% Hispanic, 30.0% black, and 7.2% white. Compared to its 1992 predecessor (also SD 32) under the 2000 census, the Hispanic VAP decreased by 4.2 percentage points.

d. SD 33, a west Bronx district, became in VAP terms 56.7% Hispanic, 23.3% black, and 13.0% white. Compared to its 1992 predecessor (SD 31) under the 2000 census, the Hispanic VAP decreased by 5.3 percentage points.

(See PX 27 at 7-11 (LATFOR’s section 5 preclearance submission regarding change in demographics for SDs 28, 32, and 33).)

It is, we believe, especially noteworthy in analyzing this Shaw claim that all three Bronx Hispanic senators — Sens. Mendez, Espada, and Gonzalez — voted in favor of the Senate Plan as, of course, did Sen. Velella. While Sen. Hassell-Thompson did not vote for the Plan, she testified that she supported the lines that were drawn that would configure her district. (Stip-¶ 107.)

3. The Plaintiffs’ Case

The plaintiffs point to circumstantial and direct evidence that they contend supports their claim that SD 34 was drawn in 2002 predominantly on the basis of race in disregard of traditional districting principles, including: (a) the shape (non-compactness) and the boundaries of SD 34; (b) the information available to the Legislature before it voted on the Senate Plan; (c) certain written submissions made by (then) Senate Majority Leader Ralph Mar-ino and Mr. Burgeson in 1992 in Wolpoff v. Cuomo; and (d) memoranda written by Mr. Burgeson in 2001. We conclude that “[o]n this record, we cannot say that plaintiffs have carried their burden and made a sufficient showing, either circumstantially or directly, to sustain a finding that race predominated in the districting.” Chen, 206 F.3d at 521.

(a) Shape and Boundary Lines

The plaintiffs observe that SD 34 in 2002 “looks something like a doughnut that has had a bite taken out of it and then been run over by a car.” (Beveridge Decl. ¶ 50.) While the district contours may be ungainly, it is not the least compact Senate district and, in any case, its shape is relevant only for any light it may shed on the claim here that “race for its own sake, and not other districting principles, was the legislature’s dominant and controlling rationale in drawing its district lines.” Miller, 515 U.S. at 915, 115 S.Ct. 2475. We do not find SD 34’s shape to be proof of racial districting. For one thing, the shape of SD 34 changed relatively little from 1992 to 2002 — or from 1982 to 1992. Its geographic “core” has been in the east Bronx during that entire period. It crossed the Bronx/Westchester border in 1982 when the Westchester town of .Pelham and parts of New Rochelle, Mount Vernon, and Yonkers were added to it. The changes to SD 34’s contours that have occurred — including what the plaintiffs refer to as the “second” Bronx/Westchester border crossing — reflect changes principally in the adjacent majority-minority black district. So, for example, when the Bronx majority-minority black district first crossed into Westchester in 1992, it did so by picking up a portion of SD 34’s Westchester population. The creation of a third Bronx majority-minority Hispanic district in 1992 also impacted the contours of SD 34, extending its western portion into Van Cortlandt Park and Riverdale.

Nor do the compactness measures the plaintiffs have relied upon (see PX 6 tbl.12) demonstrate racial predominance. Two compactness measures presented by the plaintiffs (Reock and Ehrenburg) reflect that SD 34 is close to the average in compactness of all districts in the State. SD 34 is in a tie with SD 16 as the least compact New York district according to only one measure (Polsby-Popper); it is the “second worst” according to two measures (Population Polygon and Sehwartz-berg tests), where it is behind SDs 14 and 16, respectively; and it is the “fourth worst” behind SDs 14, 51, and 59 according to another measure (Population Circle). 166 Compared to its neighbors in the Bronx, SD 34 is only slightly less compact than SDs 28 and 36. And, SD 34 is hardly less compact than the proposed new majority-minority .Hispanic district under the Plaintiffs’ Revised Plan based, for example, upon the Reock measure.

The plaintiffs also contend that the addition in 2002 of the (predominantly white) Town of Eastchester to SD 34 and an analysis of the borders and border crossings of SD 34 indicate that the boundaries of the district were changed to track racial residential patterns. Dr. Beveridge, the plaintiffs’ expert, testified that “[flour-fifths (31,318) of the people added [on a gross basis] to District 34, and 99.66% of those added from Westchester, were added as a single indivisible unit, the Town of Eastchester” where the VAP is 85.7% white, 2.7% black, and 4.1% Hispanic. (Beveridge Decl. ¶ 53.) The plaintiffs’ position is not compelling with regard to Eastchester because, among other things, that community was already contiguous to the Mt. Vernon portion of SD 34 at the time it was added and because Eastchester accounts for a relatively small portion (9.8%) of the SD 34 VAP. (See DX 73.) And, as the plaintiffs themselves acknowledge, under the New York State Constitution, Eastchester could only have been added as an entire indivisible unit. See N.Y. Const, art. III, § 4 (“No town, except a town having more than a full ratio of apportionment ... shall be divided in the formation of senate districts.”). 167 The fact of a double border crossing by District 34 is, as noted, attributable largely to changes in 1992 in the contours of SD 36, the adjacent Bronx majority-minority district. 168 With respect to block by block comparisons, the plaintiffs present maps showing the white VAP by block in SD 34 in 1992 (PX 4) and 2002 (PX 3), but they do not provide a complete analysis of particular block-by-block changes. A (visual) comparison of the two maps presented appears to show both changes that may relate to the number of whites in the district and also changes that do not appear to reflect such correlation. See Chen, 206 F.3d at 509. The defendants point out, moreover, that most of the population that was moved out of SD 34 in 2002 was moved into three of the Bronx majority-minority districts, namely, SDs 32, 33, and 36 (see DX 106), which presumably enhanced the ability of minorities to elect representatives of their choice. “[W]e simply cannot find that the challenged shape of the districts here offers adequate evidence that race predominated in the districting.” See Chen, 206 F.3d at 509.

The plaintiffs also contend that the shape and boundaries of SD 34 violate provisions of the New York State Constitution relating to compactness and county border crossings. N.Y. Const. art. III, § 4 (“[Districts shall be so readjusted or altered that each senate district shall ... be in as compact form as practicable ... and no county shall be divided in the formation of a senate district except to make two or more senate districts wholly in such county.”). We have described above the plaintiffs’ compactness arguments. As to border crossings, the plaintiffs point out that there are five county crossings in the Bronx under the Senate Plan. SD 34, as noted, crosses the Bronx/Westchester border twice; SD 36 crosses the Bronx/Westchester border once; and SDs 28 and 31 both cross the Bronx/Manhattan border once. (DX 2.) 169

Neither county-border crossings nor compactness, under New York law, are absolutes. We find that the district is not out of step with New York jurisprudence and that these very issues were considered in 1992 by the New York State Court of Appeals in Wolpoff, which upheld the 1992 Senate plan. Wolpoff involved State constitutional challenges by Bronx voters to the 1992 State Senate redistricting plan, and included challenges to the Bronx County Senate districts at issue in this case. Wolpoff, 80 N.Y.2d at 75-76, 587 N.Y.S.2d 560, 600 N.E.2d 191. The plaintiffs in Wolpoff alleged, among other things, that “the Senate redistricting plan unconstitutionally fragments Bronx County into six separate Senate districts, only two of which are contained wholly within Bronx County, despite the fact that by virtue of population, Bronx County could support only four wholly self-contained Senate districts.” Wolpoff, 80 N.Y.2d at 75, 587 N.Y.S.2d 560, 600 N.E.2d 191. The Wolpoff plaintiffs also alleged that the plan “creates districts that are neither compact nor contiguous,” id. at 76, 587 N.Y.S.2d 560, 600 N.E.2d 191, and that under the 1992 Senate plan, border crossings “compromise[d] the integrity of 23 counties.” Id. at 80, 587 N.Y.S.2d 560, 600 N.E.2d 191.

The trial judge in Wolpoff was persuaded by the plaintiffs’ arguments and held that the Senate plan “ ‘excessively, gratuitously and without supervening need dictated by federal law, disregards the integrity of county boundaries in the creation of Senatorial districts.’ ” Wolpoff, 80 N.Y.2d at 76, 587 N.Y.S.2d 560, 600 N.E.2d 191. The Court of Appeals reversed. It reviewed the New York Constitution provision at issue which states that “ ‘each senate district shall contain as nearly as may be an equal number of inhabitants, excluding aliens, and be in as compact form as practicable, and ... shall at all times con sist of contiguous territory, and no county shall be divided in the formation of a senate district except to make two or more senate districts wholly in such county.’” Id. at 77, 587 N.Y.S.2d 560, 600 N.E.2d 191 (quoting N.Y. Const. art. III, § 4). And, it framed the issue as follows:

[W]e examine the balance struck by the Legislature in its effort to harmonize competing Federal and State requirements. The test is whether the Legislature has “unduly departed” from the State Constitution’s requirements regarding contiguity, compactness and integrity of counties (Matter of Schneider v. Rockefeller, 31 N.Y.2d 420, 429, 340 N.Y.S.2d 889, 293 N.E.2d 67 [ (1972) ]) in its compliance with Federal mandates .... “Our duty is ... to determine whether the legislative plan substantially complies with the Federal and State Constitutions” (id., at 427, 340 N.Y.S.2d 889, 293 N.E.2d 67). A strong presumption of constitutionality attaches to the redistricting plan and we will upset the balance struck by the Legislature and declare the plan unconstitutional “ ‘only when it can be shown beyond reasonable doubt that it conflicts with the fundamental law, and that until every reasonable mode of reconciliation of the statute with the Constitution has been resorted to, and reconciliation has been found impossible’ ” (Matter of Fay, 291 N.Y. 198, 207, 52 N.E.2d 97 [ (1943) ]).

Wolpoff, 80 N.Y.2d at 78, 587 N.Y.S.2d 560, 600 N.E.2d 191. The Court of Appeals determined that the defendants had “put forth more than enough evidence to support [their] argument that any such [county crossings and compactness] violation was minimized and that the district lines were drawn as they were in order to comply with Federal statutory and constitutional requirements.” Id. at 79-80, 587 N.Y.S.2d 560, 600 N.E.2d 191. 170 It concluded that “[although we are troubled by the number of divided counties ..., [w]e are satisfied that in balancing State and Federal requirements, the [State] has complied with the State Constitution as far as practicable .... ” Id. at 80, 587 N.Y.S.2d 560, 600 N.E.2d 191. The. Court also considered the plaintiffs’ compactness and contiguity claims and found them “to be without merit.” Id.

Wolpoff appears to us to have confirmed the appropriateness of border crossings in the 1992 Senate plan and, to some extent at least, also the “compactness” of SD 34 under New York law. The holding in Wolpoff also negates the plaintiffs’ argument that the defendants in this case have disregarded traditional State districting principles in fashioning SD 34. 171

(b) Information Available to the Legislators in 2002

The plaintiffs contend that before the 2002 Senate Plan came up for a vote, only limited information (presumably “racial, but not political or socio-economic data”) had been supplied to the legislators. (See PX 48.) Sen. Paterson, himself a named Defendant in this case, testified for the plaintiffs that

[i]n early April 2002, a booklet was delivered to the offices of every member of the Senate and Assembly titled, “Proposed Senate and Assembly District Lines 2002, Revised Effective April 7.” The booklet contained a map of every proposed Senate and Assembly district in the redistricting bill that was endorsed by [LATFOR] and approved by the Legislature that week. It also contained census data for each proposed district: total population, deviation from the ideal population, and both population and voting-age population broken down by race and Hispanic origin. The booklet provided no census data about any other population characteristics than race, and no data other than the census data. The booklet and the bill setting forth the metes and bounds of the districts ... were the only information about the proposed districts distributed to the members of the Legislature before we voted.

(Paterson Decl. ¶¶ 12-13.) The defendants respond by pointing out, among other things, that extensive “political information ... was available to Senators and their aides on the [LATFOR] computer system.” (Burgeson Aff. ¶ 40.) The transcript of the Senate floor debate prior to passage of the Senate Plan also indicates that individual legislators were informed about the redistricting process:

As ten years ago, we did try to meet with individual senators] and work with them so that we could make sure the districts again best represented your constituents.... In this redistricting, [LATFOR] had the opportunity on numerous occasions to meet with individual members at their request, and certainly no individual, whether a member of the Majority or the Minority, was denied their opportunity to speak either with me, ... with Senator Bruno, with Mark [Burgeson].... We made all our resources available. And a number of members attended [the public] meetings and met with us: Senator Mendez, Senator Gonzalez, Senator Sampson, Senator Carl Kruger, Senator Gentile, Senator Santiago, Senator Malcolm Smith, Senator Toby Stavisky, among others. So I thank you for your participation.

(DX 134 at S 1350-51 (testimony of Sen. Skelos).) Specifically, “[a]t the Task Force public hearing in Bronx County, extensive testimony was given ... that, regardless of the unusual shapes of the two districts [SDs 34 and 36], the area had received superior representation by the two senators representing Bronx/West-chester and the community endorsed that continuation.” (Burgeson Aff. ¶ 36.) 172

The “booklet” referred to by Sen. Paterson hardly establishes the plaintiffs’ heavy Shaw burden of proving racial predominance. For one thing, it includes information relating to district population and population equality including deviation from the ideal. Equality of population is a constitutional requirement and, in and of itself, a fundamental districting principle. See Marylanders for Fair Representation, Inc. v. Schaefer, 849 F.Supp. 1022, 1054 (D.Md.1994) (three-judge court). The booklet also describes each district’s metes and bounds and VAP, including demographic breakdowns. This information is unobjectionable and the presence of racial information does not render the decision-making suspect. See Chen, 206 F.3d at 502 (“[T]he mere presence of race in the mix of decision making factors ... does not alone automatically trigger strict scrutiny.”). In view of the series of hearings, meetings and public comment (and website material) that preceded the LATFOR proposals- — and the normal dialogue expected to occur between and among politicians whose districts were up for renewal — it would be naive to suggest that all of New York’s assemblypersons and senators were unaware of the political import and stakes, and details, of the 2002 redistricting. 173 And, in light of the fact that most New York senators and assemblypersons are incumbents, it is also difficult to imagine, let alone conclude, that they were unaware of any other data and characteristics about their existing and proposed constituents. 174

(c) Wolpoff Case Submissions

The plaintiffs also cite to excerpts from (then) Republican Senate Majority Leader Marino’s June 22, 1992 portion of a brief submitted in Wolpoff (PX 31) and an affidavit of Mr. Burgeson, dated June 8, 1992, also submitted in the Wolpoff case (PX 30), as evidence of racial predominance. 175 The plaintiffs offer Sen. Marino’s section of the 1992 Wolpoff brief to suggest that SD 34 in 1992 (and presumably also in 2002) was configured with the purpose of including as much as possible of the white population of the Bronx without exceeding permissible population limits. (Beveridge Deck ¶ 39.) The brief states, in part, that: “The other county crossing in the Bronx, District 30, united the white concentrations in the Northwest Bronx with the west side of Manhattan. This crossing is necessary to avoid diluting minority voting strength and to maintain population equality. Not all of the white residents in the Bronx can reside in District 34 because the district would then exceed permissible population limits.” (PX 31 at 61 n. 19.) We do not conclude from this that SD 34 was a racial gerrymander in 1992. That is, we read the Marino portion of the State’s 1992 brief as demonstrating that Sen. Marino and the Legislature in 1992 sought to enhance minority voting strength in accordance with the Voting Rights Act and to achieve population equality in accordance with Reynolds v. Sims, 377 U.S. 533, 84 S.Ct. 1362, 12 L.Ed.2d 506 (1964), which are lawful districting objectives. See Marylanders, 849 F.Supp. at 1035 n. 11, 1054; cf. 42 U.S.C. § 1973.

The plaintiffs cite the 1992 affidavit of Mr. Burgeson also to suggest that SD 34 in 1992 (and presumably also in 2002) was a racial gerrymander. In 1992, Mr. Burgeson stated: “It is impossible to create four districts wholly within Bronx county and still comply with the Voting Rights Act. In order to create districts in which minority voters comprise a majority of the [VAP], it is necessary to both remove white voters that would dilute minority voting strength, and place- together concentrations of minorities.” (PX 30 ¶ 11(a).) 176 And, after describing the creation of the four Bronx majority-minority districts in 1992, Mr. Burgeson explained:

These districts left areas of white concentrations in the Bronx.... District 34 unites the white concentration in the northwest Riverdale section of the Bronx with the white section on the east side of the Bronx. It was necessary for District 34 to unite the white concentrations on the east and west sides of the Bronx by going north into Westchester County, to avoid cutting off the black communities in the North Bronx and in Mount Vernon from the rest of District 33.

(Id. ¶ 11(d).) “It is my understanding that the Voting Rights Act does not permit respect for county lines to dilute minority voting strength in this way.” (Id.) Again, we do not read racial predominance into Mr. Burgeson’s 1992 affidavit. Rather, we infer a good faith attempt to comply with the Voting Rights Act and to respect, as much as possible, state constitutional principles. relating to border crossings. See Chen, 206 F.3d at 520 (“[Accepting the plaintiffs’ minimal showing here would ... in practice narrow what is already a difficult passage through the Scylla of the Voting Rights Act and the Charybdis of Shaw nearly to the vanishing point — virtually every decision the jurisdiction made would run aground on the rocks of litigation.”).

The plaintiffs argue that “neither the [1992] Burgeson affidavit nor the [1992] Marino brief offered any non-racial explanation for the configuration of District 34, such as the preservation of the cores of existing districts, partisan advantage, or avoiding the pairing of Republican incumbents.” (Beveridge Decl. ¶ 41.) The defendants counter that there is no evidence that SD 34 in the 1990s reflected a purpose to create a white district, rather than to accommodate the neighboring majority-black district. (See Tr. at 278-88 (Bever-idge)). During a colloquy with the Court at trial, defense counsel further explained the 1992 Burgeson affidavit and the Wol-poff brief (which counsel helped prepare) this way: “Everything Mr. Burgeson and I said in 1992 had nothing to do with the purpose behind District 34. It was explaining why, after you have created the districts mandated by the Voting Rights Act, this left over district has to be constructed in this way- ....” (Tr. at 917-18.) 177

(d) Burgeson 2001 Memoranda

The plaintiffs also cite three memoranda written by Burgeson in 2001, dated May 4, 2001 (PX 18), June 21, 2001 (PX 19), and December 18, 2001 (PX 21), which they claim “document a clear intent to pack the [Westchester County] area north of the Bronx border with population in order to make that white population available to the redrawn District 34.” (PFF ¶ 711.)

The key passages of the memoranda that the plaintiffs cite come from a December 18, 2001 memorandum from Burgeson to Sen. Skelos and copied to Vinny Bruy (a Republican appointee to LATFOR), entitled “The 135,” which states in pertinent part:

There seems to be a bit of confusion over the provenance of the population numbers in Westchester which is to be attached to Bronx County. There has been some speculation that this number has been arbitrarily chosen out of thin air. It has not....

Through the examination of various combinations of counties, the following combination (which I’ll call the Hudson R/A [Reapportionment Area]) maximizes the Westchester portion attached to Bronx.

Thus, there is method (maximizing the Westchester total), not arbitrariness in selecting which R/A to use in calculating the portion of Westchester connected to Bronx.

The only consideration now, is how (or if) it is divided between Senators Velella and Hassell-Thompson. If it is not divided, and Senator Velella receives the entire “135”, I would suggest that the U.S. Justice Department would look unkindly on eliminating a minority legislator.

(PX 21 at 1-2.) 178

Dr. Beveridge, the plaintiffs’ expert, construes these memos to mean that “Districts 40 and 41 [in the three northern counties of Putnam, Dutchess, and Columbia] were underpopulated, and the Districts 17-37 [in the more southern counties of Westchester, Bronx, New York, and Richmond] were further overpopulated, in order to increase the amount of Westches-ter population in the bi-county districts. The decisions to keep Queens County intact, and not to keep Kings County intact, were made in order to produce the ‘reapportionment area’ that would maximize the amount of Westchester population in the bi-county districts.” (Beveridge Decl. ¶ 43; PX 17 at 1-2; PX 21 at 2.) He, thus, concludes that approximately 135,000 peo- pie were included in the Bronx/Westchester districts for the “specific goal” of “maximizing] the amount of Westchester population, which is overwhelmingly white, in District 34” (id. ¶ 49), and that the 135,000 people were divided between SDs (old) 33 and 34 to avoid pairing Sen. Has-sell-Thompson (who lives in Mount Vernon, in Westchester) with Sen. Velella (PX 21 at 2).

These memos hardly prove racial predominance. Quite the contrary. First, as the plaintiffs themselves concede, one objective for splitting the Westchester population between Sens. Velella and Hassell-Thompson was to avoid harming the majority-black district, which is plainly a legitimate districting consideration. (See id. (“If it [the Westchester portion containing 135,000 people] is not divided, and Senator Velella receives the entire ’135’, I would suggest that the U.S. Justice Department would look unkindly on eliminating a minority legislator.”).) Second, the memo-randa reflect a political intention to support Sens. Velella and Hassell-Thompson, which is what the Senate Plan did, and to avoid pairing incumbents, both of which are appropriate districting considerations. (See id.) Third, there is no indication in the memoranda that the Senate majority’s goal was to create a white district (SD 34) as opposed to a safe Republican district, also an appropriate (political) districting objective. See Gaffney v. Cummings, 412 U.S. 735, 753, 93 S.Ct. 2321, 37 L.Ed.2d 298 (1973) (“Politics and political considerations are inseparable from districting and apportionment_The reality is that dis-tricting inevitably has and is intended to have substantial political consequences.”). Fourth, the options and scenarios reflected in the memoranda appear to be well within the purview of legislators who seek to balance multiple and sometimes competing interests in the course of developing redistricting plans. See Miller, 515 U.S. at 915, 115 S.Ct. 2475 (“Electoral districting is a most difficult subject for legislatures, and so the States must have discretion to exercise the political judgment necessary to balance competing interests.... [T]he good faith of a state legislature must be presumed.”).

4. Traditional Districting Principles

For the reasons stated, the evidence presented by the plaintiffs does not reflect that race predominated over traditional districting principles. Indeed, the record reflects unequivocally that SD 34 comports with several traditional (non-racial) dis-tricting principles, as follows:

First, SD 34 is of substantially equal population with the Senate Plan’s 61 other districts. Its total population of 311,260 falls within 1.7% of the ideal district population. (PX 6 tbl.10.) See Marylanders, 849 F.Supp. at 1054.

Second, 2002 SD 34 preserves the geographic and population cores of its 1992 predecessor district. In so doing, we think it a fair inference that SD 34 also maintains established links between its constituents and their incumbent senator, Sen. Velella, while also maintaining corresponding links for the constituencies of neighboring districts with their Bronx senators. The configuration of the - district has changed little since 1992, with 87.7% of the 2002 SD 34 population coming from the 1990s district. (Burgeson Aff. ¶ 32.) And, if we focus upon the Bronx base of the district, fully 96.8% of SD 34 was retained in 2002. (Id.) The preservation of the core of this district is an important, plainly observable, traditional districting principle. (See, e.g., DX 141 ¶ 39 (in which districting expert Bernard Grofman included “preserving the cores of existing districts and, to the extent feasible, ... the configurations of existing districts” within the category of “traditional district- ing criteria”).) 179 The plaintiffs’ own senior research analyst, Mr. Breitbart, “agreed that the 2002 Senate Plan did a much better job of preserving the cores and configurations of existing districts than the Revised Plaintiffs’ plan.” (StipJ 64.) Indeed, the parties have stipulated that: “The 2002 Senate Plan preserves, on average, more of the cores of existing districts than the Revised Plaintiffs’Plan.” {Id. ¶ 235.)

Third, SD 34’s contours protect an incumbent (Republican) senator and avoid pairing incumbents, particularly Sen. Vel-ella and Sen. Gonzalez. (Burgeson Aff. ¶ 39 (“88% of Senator Velella’s constituents under the 2002 Senate plan come from the district that he easily won in the 1990s.”).) This contrasts with the Plaintiffs’ Revised Plan, which pairs Sen. Velella with (Democratic) Sen. Gonzalez in a matchup favoring Sen. Gonzalez. (Id. ¶ 39-40.) “It is a practical flaw for a plan to have excessive incumbent pairings.” {Id. ¶ 14.) We agree with the defendants that it is plain that the Senate majority had a powerful political incentive to avoid eliminating the district of a senior Republican incumbent or pairing him with a Democratic incumbent in a majority-Hispanic district that the Republican could not expect to win. 180 See Vera, 517 U.S. at 965, 116 S.Ct. 1941 (“[W]e have recognized incumbency protection, at least in the limited form of ‘avoiding contests between incumbent[s],’ as a legitimate state goal.” (citations omitted)); see also Chen, 206 F.3d at 507 (“The objective case against a district may also be affected by an analysis of whether the district adheres to traditional districting principles other than compactness, such as political considerations .... ” (citing Vera, 517 U.S. at 969-71, 116 S.Ct. 1941)). All of the Bronx 2002 redistrieting favored incumbents, and it is not at all surprising that the Senate Plan received political support from Bronx Democrats and Republicans alike, including the votes of four of the five senators and the approval of the fifth, Sen. Hassell-Thompson.

Fourth, SD 34 has evolved as the Bronx population has changed and as majority-minority districts have been added to reflect Voting Rights Act requirements. See Robertson v. Bartels, 148 F.Supp.2d 443, 458 (D.N.J.2001) (finding that strict scrutiny did not apply because the districting plan “carefully was drawn utilizing traditional redistrieting principles while seeking to comply with the Voting Rights Act by giving minority candidates the opportunity to be elected to political office”), aff'd mem., 534 U.S. 1110, 122 S.Ct. 914, 151 L.Ed.2d 881 (2002). In 1992, as we have shown, the majority-minority black district — then SD 33 — moved into Westchester County and annexed a portion of SD 34. Changes to SD 34 were also occasioned, in part, by the addition of a third majority-Hispanic district in the Bronx; a part of the district represented by Sens. Bernstein and Korman was added to SD 34. These changes, in turn, facilitated the establishment of three Hispanic districts, SDs 28, 31, and 32, in the southern and western portions of the Bronx. 181 There is no persuasive evidence that SD 34 in 2002 or in the 1990s reflected a legislative purpose to create a white district, rather than, for example, drawing a district accommodating incumbent senators and majority-minority districts. 182

5. Politics/Race: Cromartie II

The Supreme Court has observed that “[politics and political considerations are inseparable from districting and apportionment .... It is not only obvious, but absolutely unavoidable, that the location and shape of districts may well determine the political complexion of the area.” Gaffney, 412 U.S. at 753, 93 S.Ct. 2321 (emphasis added). It is certainly true that politics played a role in the 2002 redistricting of SD 34. Among the documents which Magistrate Judge Frank Maas reviewed in camera in this case were political requests by Republican legislators to adjust the lines of their districts in ways they believed would “enhance their chances (or those of their successors) to prevail in future elections.” (Rodriguez, 2003 WL 22109902, at *2, dated Sept. 10, 2003, (Maas, M.J.).) 183 The plaintiffs’ own senior research analyst with the Office of Minority Counsel for the State Senate, Mr. Breitbart, “ ‘assume[d]’ that the Senate Republican majority had a political motiva-' tion for drawing Senate District 34 the way it did.” (StipJ 176.) And, it is quite obvious, as the defendants maintain, that the Senate majority had a strong motive to help Sen. Velella, a Republican, hold onto his seat — perhaps especially because it is in a Democratic district. We have already noted that Sen. Vellela has prevailed in SD 34 for sixteen years, and, Sen. Paterson testified that the Democrats failed in their efforts to field a serious challenger against Sen. Velella in 2002 because “there really wasn’t a realistic chance of beating him.” (Tr. at 519.) The Plaintiffs’ Revised Plan would spell political trouble for Sen. Velel-la by pairing him against Sen. Gonzalez.

We agree with the defendants that the redistricting of ’SD 34 is properly justified by adherence to neutral districting principles, such as preserving cores and avoiding incumbent pairing, and by avoiding the partisan reconfiguration against Sen. Velella “that characterizes all of the alternatives.” (DFF Response ¶ 32.) Indeed, the partisan political impact in the Bronx and elsewhere of adopting the 2002 Senate and Assembly Plans, rather than the alternative plans, had the anticipated consequence: virtually all Republican and Democratic incumbents were reelected. (See Burgeson Aff. ¶ 38.)

We also believe it has been shown here that racial identification correlates with political affiliation in the Bronx. Dr. Stanley described the data in the following way:

Analyzing voting patterns in general elections, the bulk of the contests Professor McDonald and Dr. Handley analyze, confirms that black and Hispanic voters back Democratic candidates. This is no surprise, since the Democratic inclinations of black and Hispanic voters are well-known facts in the political science literature. In contrast, whites are much more split in their partisan preference, sometimes supporting Republicans, sometimes Democrats. With respect to the elections analyzed, whites certainly seem to lean Republican in the Nassau districts (although the pattern is far more mixed in the Bronx/Westches-ter).

(Stanley Aff. ¶ 37.) While the evidence of white voting behavior in the Bronx/West-chester may be somewhat “mixed,” there is evidence that in SD 34 white voters lean Republican at least in State Senate elections and have done so for at least sixteen years. 184

The plaintiffs have not shown, as they must, that considerations of race rather than politics predominated in the configuration of SD 34. The plaintiffs cite the defendants’ assertion of legislative privilege and the absence of political (presumably written) data before the Legislature when it voted on the Senate Plan, arguing that the defendants cannot now rely on the argument that politics predominated “when the uneontradicted testimony is that members of the legislature were supplied with racial data, not political data, on the districts before they voted on the 2002 Senate Plan.” (PFF ¶ 721.) The plaintiffs also urge us to draw the adverse inference that the defendants’ invocation of legislative privilege regarding how SD 34 was drawn indicates that race must have predominated. The defendants counter persuasively that such an inference is unwarranted because “it was not necessary for [Senate Majority Leader] Bruno to take the stand and make the astonishing revelation that politics may have seeped into the redistricting process, or that Senator Vel-ella is better off in current District 34 than in a majority-Hispanic district with a Hispanic incumbent.” (DFF Response ¶ 31.) We find that no adverse inference is warranted.

“In a case ... where majority-minority districts (or the approximate equivalent) are at issue and where racial identification correlates highly with political affiliation, the party attacking the legislatively drawn boundaries must show at the least that the legislature could have achieved its legitimate political objectives in alternative ways that are comparably consistent with traditional districting principles. That party must also show that those districting alternatives would have brought about significantly greater racial balance.” Cromartie II, 532 U.S. at 258, 121 S.Ct. 1452. The plaintiffs have made no such showing here. We find that “where, as here, there is a strong correlation between race or ethnicity and party, drawing district lines along political lines that may coincide with racial or ethnic lines does not evidence intentional racial discrimination.” Session, 298 F.Supp.2d at 508 n. 201. The plaintiffs have therefore failed to prove that SD 34 was drawn predominantly on the basis of race. Accordingly, we find no equal protection violation under Shaw.

VI. CONCLUSION

For the reasons explained above, each of the claims by the plaintiffs and the plaintiffs-intervenors should be dismissed. The Clerk is therefore directed to enter judg ment dismissing the Complaint and closing this case.

SO ORDERED.

ORDER

Upon the record and prior proceedings herein, including Plaintiffs’ Objections to the Direct' Testimony of Dr. Stephan Thernstrom, dated November 17, 2003; Senator Bruno’s Objections to Plaintiffs’ Direct Testimony and Exhibits, dated November 17, 2003; Plaintiffs’ Responses to Defendants’ Objections to Plaintiffs’ Direct Testimony and Exhibits, dated November 18, 2003; and Senator Bruno’s Response to Plaintiffs’ Objections to the Testimony of Stephan Thernstrom, dated November 18, 2003, and the transcript of proceedings held November 19, 2003, the Court makes the following evidentiary rulings.

1. Defendants’ objection on relevance grounds to the admission of Plaintiffs’ exhibits (“PX”) 18-21, and the testimony of Andrew Beveridge with respect to those exhibits, is denied.

2. Defendants’ objection on hearsay grounds to the admission of PX 29, 36-38, and 41 is granted, see Blue Cross and Blue Shield of New Jersey, Inc. v. Philip Morris, Inc., 141 F.Supp.2d 320, 323 (E.D.N.Y.2001)(“Judicial findings in other cases proffered as evidence are generally characterized as inadmissible hearsay.”)(citations omitted), provided that the Court may take judicial notice of the reported decisions contained in those exhibits, see United States v. Jones, 29 F.3d 1549, 1553 (11th Cir.1994)(citations omitted).

3. Defendants’ objection on relevance grounds to the admission of PX 1 and 6 is denied.

4. Defendants’ objection on relevance and hearsay grounds to the admission of PX 5 is denied.

5. Defendants’ objection .on hearsay grounds to the admission of PX 14, 28, 42, 45, 49, and 62 is denied.

6. Defendants’ objection on hearsay grounds to the admission of PX 61 is granted as Plaintiffs have not established its admissibility finder Fed. R. . Evid. 803(8) or any other hearsay exception.

The weight the Court may give to exhibits and testimony will be reflected in the Court’s decision.

1 . The plaintiffs' definitions of "upstate” and "downstate” are unclear according to Todd A. Breitbart ("Breitbart”), one of the plaintiffs' own proposed fact witnesses and a senior research analyst with the Office of Minority Counsel for the State Senate. (See Stip. of Undisputed Facts ("Stip.”) It 51 ("In Todd A. Breitbart’s August 13, 2003 deposition, he stated 'I am not aware of any generally accepted practice for defining upstate and downstate.' ”).) The plaintiffs' definition of downstate appears to exclude Nassau and Suffolk Counties and includes Rockland County and part of Orange County.

2 . The term "black” as used here denotes non-Hispanic census respondents who indicated "black" only or "black” and any other race; "Hispanic” denotes all census respondents indicating "Hispanic”; and “white” denotes non-Hispanic census respondents who indicate “white” and no other race. The term "proportionality,” as articulated in De Gran-dy,

links the number of majority-minority voting districts to minority members' share of the relevant population. The concept is distinct from the subject of the proportional representation clause of § 2, which provides 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). And the proviso also confirms what is otherwise clear from the text of the statute, namely, that the ultimate right of § 2 is equality of opportunity, not a guarantee of electoral success for minority-preferred candidates of whatever race.

De Grandy, 512 U.S. at 1014 n. 11, 114 S.Ct. 2647.

3 . Blacks would constitute 17.5% of the VAP and 18.1% of the CVAP, and Hispanics would constitute 22.7% of the VAP and 15.6% of the CVAP.

4 . LATFOR consists of six members: two co-chairmen, including a member of the Senate (Sen. Dean Skelos) appointed by the President of the Senate and a member of the Assembly (Assemblyman Sam Hoyt) appointed by the Speaker of the Assembly; two other members of the Legislature, including one (currently unfilled) appointed by the minority leader of the Senate and one (Assemblyman Chris Ort-loff) appointed by the Minority Leader of the Assembly; and two individuals who are not members of the Legislature, including one (Mark Bonilla, Esq.) appointed by the President of the Senate and one (Dr. Roman Hedges) appointed by the Speaker of the Assembly. (See Stip. ¶ 6).

Section 83-m of New York's Legislative Law directs LATFOR to "engage in such research studies and other activities as its co-chairmen may deem necessary or appropriate in the preparation and formulation of a reapportionment plan for the next ensuing reapportionment of senate and assembly districts and congressional districts of the state.” (StipA 5.)

5 . The plaintiffs include Eric Rodriguez et al., and Howard T. Allen et al. and four groups of plaintiffs-intervenors (collectively, "the plaintiffs”). See infra note 14. The defendants include Governor Pataki, Lieutenant Governor Mary O. Donohue, Attorney General Eliot Spitzer, Senate Majority Leader Joseph L. Bruno, Assembly Speaker Sheldon Silver, Assembly Minority Leader Charles Nesbitt, Senate Minority Leader David A. Paterson, the Commissioners of the New York State Board of Elections, and defendant-intervenors Anthony Creaney and Thomas Terry (collectively, "the defendants”).

6 . The defendants have relied, in part, upon the direct testimony of Mark Burgeson, assistant to Sen. Sítelos (R), Co-Chairman of LAT-FOR, who worked on New York's redistricting plans during the 1980, 1990, and 2000 districting cycles. Mr. Burgeson testified as a redistricting expert and was "asked to provide an objective assessment of the 2002 Senate plan's compliance with various redistricting criteria.” (Burgeson Aff. ¶ 2.)

7 . The plaintiffs have also proposed several Senate plans, including a 62-seat plan incorporated in the Amended Complaint ("Plaintiffs' Original Plan”); a 62-seat plan incorporated in the Joint and Consolidated Amended Complaint ("Plaintiffs’ Revised Plan”); a 61-seat plan which had been submitted by the plaintiffs to Special Master Frederick B. Lacey ("Plaintiffs' 61-seat plan”); and a 62-seat plan submitted by the plaintiffs to Special Master Lacey ("Plaintiffs' 62-seat plan”). The plaintiffs have also developed an alternative plan called the "Reverse Maximization Plan.” The Senate Democrats prepared all «districting plans presented by the plaintiffs in this litigation. (Stip.1I 84.)

8 .See Letter, from David M. Gouldin of Le-vene, Gouldin & Thompson LLP to Richard M. Berman, U.S.D.J. (Nov. 3, 2003) ("The issues involving [Doty's] application were resolved even before removal and a discontinuance to that effect was signed by our office and most, but not all, of the opposing counsel. No reason was offered by those declining to sign.”); see also Memo Endorsement, dated Nov. 6, 2003 (“Any remaining claims by or against Thomas Doty are hereby dismissed on consent.”).

9 . On April 30, 2002, Justice Cahn appointed Kenneth J. Bialkin, Esq., as Referee “to hear and report, and to submit to [the Supreme Court] a valid reapportionment plan for the Congressional districts of the State of New York ....” Alien, Order at 2-3 (May 3, 2002).

10 . The configuration of CD 17, which is challenged in this litigation, is similar to the CD 17 proposed in the Lacey Plan. (See Aff. of C. Daniel Chill in Supp. of Defs.’ Mot. for Summ. J. with Regard to the Eighth Cause of Action, dated Oct. 2, 2003, ¶¶ 18-19 (“In light of the Court's determination that the configuration of the ‘voting rights’ districts within the City of New York in the Lacey plan complied with the Voting Rights Act and the United States Constitution, the Legislature, on June 5, 2002, passed districting legislation which adopted, without change, the district lines in the Court's plan within the City of New York.... [T]he Legislature (as was Special Master Lacey in crafting his plan) was compelled to reach up into the heavily white area of Westchester to find the additional population necessary to satisfy the one/person, one/ vote Constitutional mandate in respect to District 17.”).)

11 . New York may not change a "voting qualification or prerequisite to voting, or standard, practice, or procedure with respect to voting” in covered jurisdictions unless the Attorney General of the United States or the United States District Court for the District of Columbia has precleared the change. 42 U.S.C. § 1973c; Jud. Admin., Procedures for the Admin. of Section 5 of the Voting Rights Act of 1965, as amended, 28 C.F.R. Part 51 app. (2003). To obtain preclearance, the State must demonstrate that proposed changes "do[ ] not have the purpose and will not have the effect of denying or abridging the right to vote on account of race or color.” 42 U.S.C. § 1973c.

12 .On July 25, 2002, this Court denied the plaintiffs’ motion to reinstate the Lacey Plan and to reduce the number of signatures required for primary ballot access for the 2002 congressional elections. Rodriguez v. Pataki, Nos. 02-618 & 02-3239, 2002 WL 1733676, at *7 (S.D.N.Y. July 25, 2002) (three-judge court).

13 . With the gain of one seat in the November 2002 elections, the Democratic majority in the congressional delegation increased from 19/12 to 19/10.

14 . References to "the plaintiffs” generally include the plaintiffs in the Allen and Rodriguez cases and the plaintiffs-intervenors. The plaintiffs-intervenors are sometimes referred to separately as follows:

(i)The Latino Voting Rights Committee of Metro New York et al. ("LVRC Interve-nors”);

(ii) Peter K. Lau et al. ("Lau Interve-nors”);

(iii) Sandra Rivers, Leslie Barbour, et al. ("Rivers/Barbour Intervenors”); and

(iv) Assemblyman David F. Gantt et al. ("Gantt Intervenors”).

On April 23, 2003, the Gantt Intervenors asked the Court for permission to withdraw Count IX, which challenged CD 28 (covering western upstate New York, including Rochester and Buffalo). The Court granted the request. (See Memo Endorsement, dated Apr, 30, 2003.)

15 . The Rivers/Barbour intervenors urge us to exercise our discretion to sever this claim because it does not clearly come within the jurisdictional authority afforded this court under 28 U.S.C. § 2284. While the challenge to the Congressional Plan remaining in the Complaint is distinguishable from other counts that challenge the Senate Plan, in the course of this litigation, this three-judge court has heard many claims and arguments within the § 2284 parameters challenging the Congressional Plan. Accordingly, we decline to sever this claim. See Page v. Bartels, 248 F.3d 175, 190 (3d Cir.2001); Armour v. Ohio, 925 F.2d 987 (6th Cir.1991).

16 . The summary judgment submissions were voluminous and comprehensive. The defendants' papers included: (i) Memorandum of Law in Support of Defendants' Motion for Summary Judgment ("Defs. Mem.”); (ii) Affidavit of C. Daniel Chill in Support of Defendants' Motion for Summary Judgment With Regard to the Eighth Cause of Action, dated Oct. 2, 2003 ("Chill Aff.”), with exhibits; (iii) Declaration of John R. Braatz, dated Oct. 3, 2003 ("Braatz Decl.”), with exhibits; (iv) Defendants’ Statement of Material Facts Pursuant to Local Civil Rule 56.1, dated Oct. 3, 2003 ("Defs. SOF”); and (v) Reply Memorandum of Law in Further Support of Defendants' Motion for Summary Judgment and in Opposition to Plaintiffs' Cross Motion, dated Oct. 31, 2003 ("Defs. Reply”).

The plaintiffs' papers included: (i) Plaintiffs’ Memorandum of Law in Opposition to Defendants' Motion for Summary Judgment and in Support of Their Cross-Motion for Summary Judgment [corrected version] ("Pis. Mem.”); (ii) Declaration of Sarah Netburn, dated Oct. 23, 2003 ("Netburn Decl.”), with exhibits; (iii) Plaintiffs’ Revised Response to Defendants’ Rule 56.1 Statement, attached as Ex. A to Netburn Decl. ("Pis. SOF Response”); (iv) Plaintiffs’ Rule 56.1 Counter-statement of Material Facts for Trial, attached as Ex. B to Netburn Decl. ("Pis. SOF”); (v) Affidavit of Joan P. Gibbs, Esq. in Opposition to Defendants' Motion for Summary Judgment as to the Eighth Cause of Action, dated Oct. 23, 2003 ("Gibbs Aff.”), with exhibits; (vi) Plaintiffs-Intervenors Council of Black Elected Democrats et al and Sandra Rivers et al Response to Defendants' 56.1 Statement ("Rivers/Barbour Pis. SOF Response”); (vii) Attorney Declaration in Support of Plaintiff-Intervenors Latino Voting Rights Committee of Metro New York et. al. (LVRC) Reply to Defendants' Summary Judgment Motion as to Fourth Cause of Action, dated Oct. 23, 2003 ("Chin Aff.”); (viii) Affidavit of Angelo Falcon in Support of Plaintiff-Intervenors’ Reply to Defendants’ Motion for Summary Judgment on the Fourth Cause of Action, dated Oct. 23, 2003, attached as Ex. 2 to Chin Aff. ("Falcon Aff.”), with exhibits; and (ix) Plaintiff-Inter-venors Latino Voting Rights Committee of Metro New York et. al. (LVRC) Local Rule 56.1(b) Statement of Disputed Facts in Reply to Defendants' Motion for Summary Judgment on the Fourth Cause of Action, with exhibits ("LVRC SOF Response”).

17 .At the pre-trial conference held on November 19, 2003, the vast majority of exhibits and proposed testimony were admitted without objection. (See Nov. 19, 2003 Tr. at 10.) The Court reserved decision with respect to several evidentiary issues, which have been resolved in a Order issued contemporaneously with this decision.

18 . The witnesses’ direct testimony submitted in the form of a declaration or affidavit is referenced below as, e.g., "McDonald Deck ¶ _.”

19 . In particular, the letters addressed Session v. Perry, 298 F.Supp.2d 451 (E.D.Tex.2004) (three-judge court), and Larios v. Cox, 300 F.Supp.2d 1320 (N.D.Ga.2004) (three-judge court).

20 . The deviation percentages are calculated using the following methodology: The deviation of a district is the percent a district’s population is above or below the ideal population, where the ideal number is determined by assuming that each district should have perfect population equality. "Maximum deviation” or "total deviation” is determined by adding the deviation of the district with, the largest population to the deviation of the district with the smallest population. "Average deviation” is determined by averaging the deviations of all the districts. See, e.g., Farnum v. Burns, 561 F.Supp. 83, 87 n. 5 (D.R.I.1983) (three-judge court).

21 . The plaintiffs’ mathematical argument is as follows: The average overpopulation of the 29 "downstate” districts in the 2002 Senate Plan is +2.37%. The number of people the 2.37% represents would total 68.80% of a single district while the average underpopulation of the 24 "upstate” districts is -2.86%, totaling 68.57% of a district. (See Pis. Mem. at 3.) Thus the "downstate” region, the plaintiffs argue, is underrepresented by approximately two-thirds of a Senate seat. The plaintiffs contend that the redistricting plan should have rounded off by eliminating an "upstate” district and creating a "downstate” one.

22 . When pressed at oral argument, counsel for the plaintiffs conceded that this count was not intended to charge defendants with intentional racial discrimination, although they suggested that the population deviation has a discriminatory effect. (See Summ. J. Tr. at 65-66.)

23 . The three-judge court in Larios v. Cox, 300 F.Supp.2d 1320 (N.D.Ga.2004), recently rejected Georgia's redistricting plans for its state legislature. In the course of that decision, Larios rejected the State’s rationale of incumbency protection, which it found to be "a permissible cause of population deviations only when it is limited to the avoidance of contests between incumbents and is applied in a consistent and nondiscriminatory manner.” Id. at 1338. The Georgia legislature's House and Senate plans violated those principles. The House plan, which included some multimember districts, paired 47 incumbents, 37 of whom were Republicans (50% of the Republican caucus as a whole). See id. at 1328-29. The effect was to ensure that 19 incumbents would not be reelected, and, in the 2002 general elections, 18 Republican incumbents lost their seats. Id. Similarly, under Georgia's 56-seat Senate Plan, see id. at 1326-27, there were 6 incumbent pairings involving 10 Republicans. Id. at 1328-29. Four of the paired Republican incumbents lost their seats. Id.

In contrast, the 2002 New York State Senate Plan has only two incumbent pairings while the Plaintiffs' Revised Plan produces four incumbent pairings. (See Braatz Deck Ex. 14.) The 2002 Senate Plan thus reflects a legitimate effort to avoid incumbent pairs. While the two pairings both involve a downstate Democrat-Democrat contest, there is no pernicious pattern of Senate Republicans manipulating population deviations in order to manufacture the pairs or oust large numbers of Democrat incumbents. Cf. Larios, 300 F.Supp.2d at 1330-31.

24 . At no time have plaintiffs asserted a claim of "partisan gerrymandering,” as contemplated by Davis v. Bandemer, 478 U.S. 109, 106 S.Ct. 2797, 92 L.Ed.2d 85 (1986).

25 . For this claim, we need not settle the dispute in the parties’ briefs about whether the New York Constitution, Article III, §§ 4, 5-a, takes a clear position on whether apportionment should proceed strictly on total population measures. The state constitution has no bearing on whether we may analyze the CVAP numbers in assessing the viability of a one-person, one-vote cause of action under the United States Constitution. In any case, here we refer to the CVAP numbers to ascertain the practical effect of the Senate Plan on actual voters. There is no finding here that the districts were apportioned on that basis.

26 . See Garza v. County of Los Angeles, 918 F.2d 763, 780-85 (9th Cir.1990) (Kozinski, J., concurring and dissenting in part) (distinguishing theory of representational equality based on total population — which protects interests such as access to government — with electoral equality based on CVAP, for example — which protects power of voters in choosing their representatives); see also Daly, 93 F.3d at 1222-28 (discussing competing theories and criticizing Judge Kozinski's promotion of one theory — electoral equality— over the other).

27 .The Larios court found the regional disparities in the Georgia plans unconstitutional particularly because there was "no evidence that the population deviations in the plans were driven by the neutral and consistent application of any traditional redistricting principles.” Larios, 300 F.Supp.2d at 1349; see id. at 1349*53. For example, the court expressly found that "core retention was not a concern in the redistricting process.” Id. at 1334. Unable to connect the population deviations to a permissible rationale, the court found that "the goal of rural district underpopulation is plainly to overweigh the votes of the region's residents and thereby increase the power of rural voters at the demonstrable expense of voters in other parts of the state.” Id. at 1346.

In this case, the Senate Plan did apply the accepted principle of maintaining the core of districts in both the upstate and downstate regions. Moreover, Larios involved the wholesale distortion of district lines throughout the state in order to target and oust members of the minority political party. In this case, the result of the alleged "upstate” favoritism is the loss of just two-thirds of a district for the downstate region. In addition, the practical effect is thát downstate districts remain underpopulated in terms of CVAP. Maintaining the cores of the "upstate" districts has not been done “at the demonstrable expense of voters” in the downstate region.

28 . Section 1973(a) refers to § 1973b(f)(2), which provides that "[n]o voting qualification or prerequisite, to voting, or standard, practice, or procedure shall be imposed or applied by any State ... to deny or abridge the right of any citizen of the United States to vote because he is a member of a language minority group.” Section 1973Z (c)(3) defines "language minority group” to mean “persons who are American Indian, Asian American, Alaskan Natives or of Spanish heritage.”

29 . The Senate Factors include:

the history of voting-related discrimination in the State ...; the extent to which voting in the elections of the State ... is racially polarized; the extent to which the State ... has used voting practices or procedures that tend to enhance the opportunity for discrimination against the minority group, such as unusually large election districts, majority vote requirements, and prohibitions against bullet voting;. the exclusion of members of the minority .group from candidate slating processes; the extent to which minority group members bear the effects of past discrimination in areas such as education, employment, and health, which hinder their ability to participate effectively in the political process; the use of overt or subtle racial appeals in political campaigns; and the extent to which members of the minority group have been elected to public office in the jurisdiction.... The [Senate] Report notes also that evidence demonstrating that elected officials are unresponsive to the particularized needs of the members of the minority group and that the policy underlying the State's ... use of the contested practice or structure is tenuous may have probative value.

Gingles, 478 U.S. at 44-45, 106 S.Ct. 2752 (citing S.Rep. No. 97-147, at 28-29, 1982 U.S.C.C.A.N. at 206-07).

30 . Senate Districts 1 through 9 are located on Long Island. The current SDs 1, 2, 3, and 4 are wholly within Suffolk County, while SDs 5 and 8 are "bridge” districts. SD 5 includes substantial portions of both Nassau and Suffolk Counties, but SD 8 is primarily located in Nassau County and captures only a portion of the Town of Babylon in Suffolk County. SDs 6, 7, and 9 are wholly within Nassau County.

31 . See PX 6 tbl. 11 (VAP data for the 2002 Senate Plan), tbl. 16 (VAP data for the Revised Plaintiffs’ Plan), tbl. 28 (CVAP data for the Revised Plaintiffs' Plan). For CVAP data for the 2002 Senate Plan, see id. tbl. 27.

32 . (S) indicates a Suffolk County district and (N) indicates a Nassau County district.

33 . The Census collects information for people who identify as non-Hispanic black and Hispanic black. The experts in this case generally referred to the non-Hispanic black population statistics when discussing African-Americans or blacks as a distinct minority group.

34 . Comprised of portions of SDs 3, 4, and 8.

35 . Comprised of portions of SDs 6, 7, 8 and 9.

36 . In Growe, the Supreme Court applied the Gingles preconditions to a single-member districting scheme but, like Gingles, did not determine whether plaintiffs can allege "that a voting practice or procedure impairs a minority's ability to influence, rather than alter, election results.” Growe, 507 U.S. at 41 n. 5, 113 S.Ct. 1075. In Voinovich, the Court assumed, arguendo, that "influence-dilution” claims are actionable, but it found no racial bloc voting in a challenge to a plan that allegedly "packed” minorities into majority-minority districts. Voinovich, 507 U.S. at 158, 113 S.Ct. 1149. In De Grandy, the Court found no section 2 violation because the minority group was an effective majority in a number of state legislative districts roughly proportional to the minority groups' respective shares in the voting age population. De Grandy, 512 U.S. at 1000, 114 S.Ct. 2647. The Court did not reach the question of "whether the first Gingles condition can be satisfied by proof that a so-called influence district may be created” where members of the minority group are "a minority of the voters, but a potentially influential one.” Id. at 1008-09, 114 S.Ct. 2647.

37 . The parties often used the term "coalition” claim or "coalition district” to refer to a claim where plaintiffs attempt to combine minority groups — for example, blacks and Hispanics — to satisfy the majority-in-a-district requirement of the first Gingles precondition. Such claims raise an interpretive issue of whether the VRA was only intended to protect individual minority groups, as well as an evi-dentiary issue of whether the minority groups are politically cohesive (the second Gingles factor).

A "coalition” district, as defined in the text above in contrast to “influence” and majority-minority districts, is a distinct concept that raises distinct issues. It raises the interpretive issue of whether a single minority group — blacks, for example — can satisfy the first Gingles precondition without being a numerical majority in the district. The core evidentiary issue is whether the minority group can show that, although not a majority, it would have the ability to elect candidates of choice with a limited amount of crossover votes from whites (or voters from other minority groups — Hispanics, for example — that are not cohesive with the relevant minority).

To avoid confusion, we will refer to claims involving minority-white coalitions as "crossover” or "ability to elect” claims. At trial, Plaintiffs’ Proposed District 8 in Nassau County (Count V) was presented as a "crossover” district that allegedly could provide blacks with the ability to elect candidates of their choice. Other claims in this case involve alleged black-Hispanic coalitions, and we will refer to those as "minority-coalition” claims. See infra note 72; infra Parts IV.C (Count III, The Bronx), IV.F (Count VIII, The Rivers/Barbour Intervenors).

38 . Although the Supreme Court has left the issue open, VAP and, in particular, CVAP, have been considered the most important statistics in defining the “relevant population” for the majority-in-a-district requirement. See De Grandy, 512 U.S. at 1008-09, 114 S.Ct. 2647 (leaving open question of which characteristic of minority population should be touchstone for vote-dilution inquiry); Valdespino v. Alamo Heights Indep. Sch. Dist., 168 F.3d 848, 851-53 (5th Cir.1999) (stating that CVAP and VAP define "relevant population” and emphasizing reliance on CVAP); Negron v. City of Miami Beach, 113 F.3d 1563, 1569 (11th Cir.1997) (requiring minority group to be majority of CVAP but noting that requirement may not apply where statistics on citizenship rates are not reliable); France v. Pataki, 71 F.Supp.2d 317, 326 (S.D.N.Y.1999) ("[T]he proper measure of population is the total citizens eligible to register and vote, not the total population.”).

39 . For example, blacks and Hispanics were not cohesive in the 2000 Democratic primaiy for CD 2, which was in western Suffolk County and included all of SD 4, all of the Suffolk County portion of SD 8, and most of the minority population in SD 3. (See Netburn Decl. Ex. 5 at 44 (McDonald Report).) In that election, the overwhelming majority of Hispanic voters supported two white candidates and a majority of blacks supported an African-American candidate. (See id.) Primaries are particularly probative because they eliminate partisanship as a potential reason for blacks and Hispanics voting cohesively. (See Aff. of Dr. Harold Stanley (“Stanley Aff.”) ¶ 38.)

The overall pattern of black and Hispanic cohesion was mixed in the local elections analyzed by Dr. McDonald, the plaintiffs' expert. Intergroup cohesion was particularly absent with respect to SD 8, which bridges Suffolk and Nassau Counties. (See generally id. at 38-57 (analysis of Suffolk County elections).) The evidence is consistent with the plaintiffs' concession at trial, discussed below, that they could not prove that blacks and Hispanics voted cohesively in neighboring Nassau County elections. With no proven cohesion, Plaintiffs’ Proposed District 4 would contain only a 17.5% black VAP and an 18.1% black CVAP. See supra Table 1.

40 . (See PX 6 this. 16, 28.)

41 . The first Gingles requirement is often said to involve "numerousness” and "compactness.” See, e.g., De Grandy, 512 U.S. at 1011, 114 S.Ct. 2647. Numerousness involves the majority-in-a-districVpotential-to-elect requirement and is the primary focus- of this analysis. The "geographic compactness” requirement generally requires plaintiffs to propose a minority district that is sufficiently compact to comply with traditional districting principles. The failure to gerrymander district lines to create a majority-minority district is not an unlawful voting practice, and thus the plaintiffs must show that compact minority district could exist in order to demonstrate a prerequisite for injury. Although Plaintiffs’ Proposed District 8 is not an extremely compact district, it is more compact than many districts in the 2002 Senate Plan and the defendants have not challenged the plaintiffs’ proposed district on these grounds. (See Tr. at 417 (Beveridge).)

42 . Because Armour based its decision, at least in part, on a finding that Gingles did not apply to single-member-district claims, the validity of Armour was put in doubt by Growe v. Emison, 507 U.S. 25, 113 S.Ct. 1075, 122 L.Ed.2d 388. Moreover, Armour has been heavily criticized by other courts. See, e.g., DeBaca, 794 F.Supp. at 996-97; Hastert, 111 F.Supp. at 652-53.

43 . In Martinez v. Bush, the legislature had created a number of "performing minority districts.” Martinez, 234 F.Supp.2d at 1322. The parties did not dispute that the Gingles conditions were satisfied because there were "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.” Id. at 1321. The dispute was simply over whether “the ostensibly performing districts drawn by the legislature [were] in fact likely to perform.” Id. at 1323. Because the court found that the district would enable black candidates of choice to be elected, it found no vote dilution. Id.

44 . The plaintiffs' redistricting expert, Dr. Andrew Beveridge testified that the intent of Plaintiffs’ Revised Plan, including Proposed District 8, was to produce the maximum minority representation possible while complying with minimum redistricting principles under Shaw v. Reno. (Tr. at 346 (Beveridge).) While the plaintiffs argued that Proposed District 8 is a sufficiently compact district that unites communities of interest, Dr. Beveridge testified that only around twelve out of the sixty-two current Senate districts would be less compact than Proposed District 8. (Tr. at 417.) Dr. Beveridge further testified that under colorblind redistricting principles — that is, acting as if everyone in Nassau were the same race — the current Senate districts would not have to be redrawn. (See Tr. at 329; see also Burgeson Aff. ¶¶ 6-28 (describing neutral districting principles underlying drawing of lines in 2002 Senate Plan).)

It is plain that the plaintiffs' argument is based on the 2002 Senate Plan's failure to change district lines to accommodate a growing minority population. The plaintiffs contend that the legislature diluted minority voting strength because it did not draw a district predominantly along racial lines in order to create a maximum black VAP of 34.5%. (See Tr. at 309 (Beveridge) (admitting that plaintiffs’ purpose was to create effective black district and that to do so, plaintiffs violated traditional districting principles to include Westbury in Proposed District 8)); cf. Goosby, 180 F.3d at 501 n. 3 (Leval, J., concurring) (questioning courts' ability to invalidate district where lines were set long ago with no racial motivation and claim has arisen merely because cluster of members of protected class developed around dividing line). Issues about compliance with neutral districting principles and the maximization of minority representation are usually relevant to the totality of circumstances inquiry. In this case, however, those issues reflect a problem with deviating from the strict majority-in-a-district Gingles factor. Without the bright-line rule, plaintiffs can argue that minority votes have been diluted merely because it would be possible to patch together disparate minority populations that, if combined, could create a vague potential to elect candidates of choice.

45 . For the purposes of this claim regarding Nassau County, the black-preferred candidate is always a cohesively supported black-preferred candidate. The distinction between a preferred candidate and a cohesively preferred candidate arises with respect to Count III involving the Bronx, and Count IV involving the LVRC intervenors.

46 . In the 1998 Democratic primary for Lieutenant Governor, a black candidate, Charles King, was not cohesively supported by black voters in Nassau County and lost the nomination. The defendants attempted to use this primary to argue that white voters in Nassau County do not vote as a bloc against minority candidates. For the reasons explained infra note 59, this primary does not support the defendants' claims about racially polarized voting in Nassau County.

The presidential primary with Jesse Jackson occurred in 1988, which is outside of the 1996-2002 range analyzed by Dr. McDonald. Dr. McDonald acknowledged that the primary is too distant to reflect current polarization in Nassau County because, among other things, the demographics and number of voters in the area has changed. Moreover, Dr. McDonald did not analyze other out-of-range primaries with black candidates, such as a 1994 primary with Al Sharpton. (See Tr. at 113-14.) And he agreed that it is difficult to extrapolate much from this one election that took place fifteen years ago. (See Tr. at 121.)

47 . All of Dr. McDonald's election analyses can be found in DX 146. This Table is a simplified and reorganized version of the Master Table of Racial Bloc Voting Analyses, which is in the beginning of that report. A "corrected” version of that table is produced at DX 148, but none of the corrections are relevant for this analysis of Nassau County.

48 . Elections in bold include a minority candidate. St. George, Dash, and McNight are all black while Santos is Hispanic.

49 . An asterisk (*) reflects the black-preferred candidate.

50 . ''BPC'' refers to the black-preferred candidate.

51 . The source page refers to the page for the date for the individual election analyses as numbered in DX 146.

52 .Three of the four minority-white elections involved black candidates while one involved a Hispanic candidate. Although the plaintiffs no longer base their claim on black and Hispanic cohesion, evidence of white bloc voting against minorities in general is still probative. At any rate, the parties have included the white-Hispanic election in the count of endogenous elections with minority candidates.

53 . This table is based on the Master Table and the individual election analyses in DX 146. A corrected version of the Master Table is produced at DX 148.

54 . Dr. McDonald analyzed the statewide elections separately for each Senate district, generating four contest results in Nassau County for each election. For each election, the result as to black cohesion was the same in all four districts, and the result as to white bloc voting was usually the same across the districts. Thus a “Yes” or “No” reflects the result for all four Senate districts unless otherwise specified.

55 . “Defeated” or “won” refers to whether the BPC received fewer or more votes than the white-preferred candidate within the districts. For each election, the result for the BPC was the same across all four districts.

56 . An asterisk (*) reflects the black-preferred candidate. Elections with a minority candidate are in bold.

57 . “2/4 Yes (SD 6, 8)” reflects that white bloc voting was found in SDs 6 and 8 but not SDs 7 and 9.

For the 2002 Comptroller election, the Master Table reports no racial polarization in all four Senate districts. However, the individual contest results show that Hevesi’s rate of support from white voters was: 35.0% for SD 6; 42.% for SD 7; 35.7 % for SD 8, and 44.7% for SD 9. (DX 146 at 86-89.) Thus in SDs 6 and 8 there was over 60% white bloc voting against the black-preferred candidate. The standard error was not significant enough in any district to alter the conclusions about polarization.

58 . The Master Table (DX 148) indicates that there was racially polarized voting in SD 9 for the 1998 Attorney General election, but the white crossover rate for that district was 42.1%. (See DX 146 at 128.)

59 . A third statewide exogenous election with a minority candidate was the Democratic primary for Lieutenant Governor in 1998, which included a black candidate, Charles King. King was not cohesively supported by black voters, and the primary did not show signs of racial polarization. (See DX 148; see also DX 146 at 79-82.) Turnout was very low, however, and the primary was the only Nassau County election without black cohesion. (See DX 150 (showing turnout to be under 4% for both whites and blacks).) Compared to the clear results found in the endogenous elections' both with and without minority candidates, this primary election is not probative for evaluating racial polarization in Nassau County.

60 . The defendants contend that "Plaintiffs do not cite any cases where a recompilation of votes is relied upon to assess electability.” (Defendant Bruno’s Post-Trial Proposed Findings of Fact and Conclusions of Law ("DFF” or "DCL”), DFF ¶¶ 529-31.) The plaintiffs maintain that “[r]ecompiled election results have been used by a number of voting rights experts to assist in gauging the effectiveness of proposed minority districts.” (Handley Decl. ¶ 18.)

61 . In SD 6, Pataki garnered 45,699 votes to McCall's 22,746, giving him, 22,953 more votes. In SD 7, Pataki won by 50,184 to 22,521, for a difference of 27,663 votes. Pa-taki won SD 8 by 46,796 to 22,841, for a difference of 23,955 votes. And in SD 9, Pataki won 55,877 to 20,492, for a difference of 35,385 votes. (PX 82.) .

62 . In addition, Dr. McDonald looked at the 1988 Democratic presidential primary between Michael Dukakis and Jesse Jackson. Whereas Dukakis obtained substantially more votes in each Nassau County Senate district, Jackson would have received more votes than Dukakis in Plaintiffs' Proposed District 8.

63 . This table is based on DX 149, Dr. McDonald’s table: Electability Analysis — State Senate. The data for the table can be found in DX 146. See also supra Table 3. Dr. McDonald’s table also included results for Hispanics. Those data have been omitted because of the plaintiffs' concessions at trial.

To estimate blacks' ability to elect candidates of choice in Proposed District 8, Dr. McDonald projected an “average” scenario using the average rates for turnout and support for the black-preferred candidate (cohesion and crossover) for each racial group. He also projected a "best case” scenario using the maximum values found for turnout and support. The maximum values are italicized in this Table and are all found in the 2000 elections.

64 . Dr. McDonald's analysis also included Hispanic voters, whose average turnout rate was 1.0% and who supported the minority-preferred candidate at a rate of 63.0%. (DX 149.) The inclusion or exclusion of Hispanic votes does not materially change the analysis.

65 . The defendants, through Dr. Stanley, offered results for a “worst case scenario,” but the results presented at trial were drastically and inexplicably different from those reported for the motions for summary judgment. (Compare Stanley Aff. ¶ 120, with Defs. SOF ¶ 173.) We thus disregard Dr. Stanley’s "worst case scenario” projections, although his general point is well taken: The plaintiffs did not provide a worst case scenario estimate and have not provided any evidence that a "best case scenario” is likely.

66 .Although blacks would be 36.9% of the CVAP, all of Dr. McDonald's statistics are generated using VAP. Moreover, Dr. McDonald testified that because of his method, using CVAP would not produce different numbers. (Tr. at 134-35.)

67 . Nassau County Legislature elections occur in odd-numbered years, in contrast to the State Senate elections, which along with most federal and statewide elections, take place in even-numbered years. While there was testimony that the electorates for the State Senate and County legislature elections are different, neither party elaborated on the issue. (See Tr. at 147 (McDonald).)

68 . Dr. McDonald admitted that he did not employ Dr. Handley’s leading methodology for his district-wide analysis. (Tr. at 124-25.) The reason, he explained, was that Dr. Hand-ley's method was based on analyzing two races (white and black), and when Dr. McDonald first set out to run his analysis, he was analyzing whites, blacks, and Hispanics. (Id.)

69 . Dr. McDonald reported recompiled turnout rates for seven of the ten statewide elections. Of the three elections he did not analyze, two were presidential elections that had relatively high turnout rates and would have slightly inflated the average. (See McDonald Decl. ¶ 80; DX 188.) The other election not reported was the 2002 Attorney General race between Spitzer and Irrizary. When the plaintiffs averaged the recompiled turnout estimates for the six racially polarized statewide elections, they found a turnout rate for blacks averaging 28.6% and for whites it was 42.7%. (PX 77.)

70 . There is also reason to question the reliability of the recompiled statistics. The recompiled statistics for the areas encompassing Proposed District 8 show that black turnout for statewide elections jumped to around 28% from 16% in State Senate elections — a 12% increase. But the difference in turnout is not nearly as dramatic when the statistics include all black voters in SDs 6, 7, 8, and 9. According to Dr. McDonald, the average estimated turnout for all black voters in SDs 6-9 was 17.3% in the State Senate elections. (DX 149.) For the ten statewide elections (including the presidential elections), black turnout in SDs 6-9 averaged an estimated 21.4%, which is only 4.1% higher than the turnout for the State Senate elections. (DX 188.) Neither party explained why the recompiled results showed a 12% increase in turnout rates for statewide elections while the results for all blacks in SDs 6-9 showed only a 4% increase.

71 . We also do not reach the question of whether in the plaintiffs’ proposed district the amount of white crossover votes needed to elect a black-preferred candidate is so large that it is inconsistent with a finding that there is white bloc voting sufficient to satisfy the Gingles preconditions. Cf. Abrams v. Johnson, 521 U.S. 74, 92-93, 117 S.Ct. 1925, 138 L.Ed.2d 285 (1997) (ruling that trial court did not clearly err in finding that the third Gingles precondition was not met where there was white voting for black candidates statewide that ranged from 22% to 38%, and electoral success for black-preferred candidates was demonstrated).

72 . The parties refer to the attempt to combine blacks and Hispanics for the purposes of the Gingles preconditions as a "coalition” claim. To avoid confusion with the distinct concept of a minority-white "coalition” or "crossover” claim, we refer to this count as presenting a black-Hispanic "minority-coalition claim.” See supra note 37.

73 . The creation of Proposed District 36 would eliminate the only majority-white district in the Bronx and would also entail (i) redrawing the entire Senate map to "equalize” population statewide and (ii) "unpacking” majority-minority districts in the Bronx. (See Compl. ¶¶ 306-08; Burgeson Aff. ¶¶ 49-50.)

74 . The Senate Plan has five districts based entirely or primarily in the Bronx: SDs 28, 32, 33, 34, and 36. SDs 34 and 36 have extensions into Westchester and SD 28 extends into Manhattan. We refer to these districts as the five Bronx districts.

75 . The plaintiffs also maintain that "[r]ather than drawing a majority-minority district, the Legislature ... dispersed some minority voters in southern Westchester County into the 35th Senate District.” (Comply 164.)

76 . The plaintiffs respond that "statewide population disparities must be scrutinized solely within the context of whether their 'effect' as one of the voting practices at work here ... denies minorities the equal access to which they are entitled under Section 2.” (Pls. Post-Trial Mem., dated Dec. 11, 2003, at 11-12.)

77 . Indeed, the parties have stipulated: "Blacks and Hispanics cannot be assumed to be a single racial group with common political aims that are usually frustrated because of bloc voting by a hostile non-Hispanic white majority.” (Stip^ 133.)

78 . See The New York Red Book 93 (Mary Ann Ryan-Germain ed., 97th ed.2003) (containing dates of service of current and past senators). Information about current incumbents is also available at http://www.senate.state.ny.us.

79 .The 1990s Senate districting plan also included five Bronx districts. Three were majority Hispanic (former SDs 28, 31, and 32), one was majority black (former SD 33), and one was majority white (former SD 34). The 1990s districts were substantially similar geographically to their current form. Former SD 28 was represented by Sen. Mendez. Former SD 31 was represented by Sen. Gonzalez. Former SD 32 was represented by Pedro Es-pada and David Rosado. Former SD 33 was represented first by Sen. Joseph Galiber, then by Sen. Larry Seabrook, and then by Sen. Hassell-Thompson. Former SD 34 was represented by Sen. Velella.

80 . For data for Table 7 and Table 8, see PX 6 tbls. 11, 16, 27, 28. The population statistics in these tables are rounded off to the nearest tenth of a percent.

81 . The districts as numbered under the Revised Plaintiffs' Plan do not correspond to the districts- as numbered under the 2002 Senate Plan. This column provides the corresponding 2002 Senate Plan District based on the geographic area covered and who the incumbent candidate would be. See DX 189 for a chart providing incumbent information and demographic data for districts under the 1992 Senate Plan, 2002 Senate Plan, and Revised Plaintiffs' Plan.

82 . As explained below, Plaintiffs’ Proposed District 36, which is the focus of the plaintiffs’ Bronx/Westchester vote-dilution claim, would draw its population from five existing districts and would provide an "open” seat.

83 . Plaintiffs' Proposed District 34 would pair two incumbents — Sen. Velella, a white Republican, and Sen. Gonzalez, a Hispanic Democrat — in this majority-Hispanic district.

84 . The first Gingles precondition also requires the relevant minority population to be sufficiently geographically compact. The-plaintiffs have offered six relevant measurements to indicate that Proposed District 36 is relatively compact. (See PX 6 tbls. 12, 17). These include four measurements (the Reock, Schwartzberg, Ehrenburg, and Polsby-Pop-per tests) that show Proposed District 36 is slightly "less compact” relative to the average Senate Plan district; and two (the Population Polygon and Population Circle tests) that show that Proposed District 36 is slightly more compact relative to the average Senate Plan district. We find overall that the plaintiffs have demonstrated, and the defendants do not appear to dispute, that the proposed Bronx district is geographically compact” under Gingles. See Gingles, 478 U.S. at 50-51, 106 S.Ct. 2752.

85 .With respect to the Nassau County claim, we first made findings about racial polarization, the plaintiffs’ methodology, and the elections analyzed before proceeding to the electability analysis. Because we have already provided a foundation for those concepts and because extensive discussion of electability is not necessary for the Bronx claim, we will proceed with the electability analysis here as a subset of the inquiry into the first Gingles precondition.

86 . See supra note 60 (noting parties' dispute over validity of recompilation method).

87 . Dr. McDonald analyzed only majority-white SDs 34 and 35 for racial polarization on the theory that the "packed” minority districts in the Bronx have relatively few whites and are, therefore, "not relevant to the Gingles analysis.” (McDonald Decl. ¶ 19.)

88 . Dr. McDonald also analyzed three racially polarized primary elections (i.e., the 1988 Democratic presidential primary, the 2001 Democratic mayoral runoff, and the 2001 Democratic city comptroller primary) to assess the ability of minorities (but not particularly Hispanics) to nominate their preferred candidate in the face of white bloc voting. He concluded that each election demonstrated that minorities would have the ability to elect the candidates of their choice in Proposed District 36. (McDonald Decl. ¶¶ 67, 69-72.)

In addition to Dr. McDonald’s results, the plaintiffs have provided recompilation results from two elections which they argue also show that the Hispanic-preferred candidate would have won by large margins in Proposed District 36 but which were not racially polarized. (See PX 11, 13.) In the 2002 State comptroller race, the Hispanic-preferred candidate won in Proposed District 36 by 21,379 to 6305. In the 1998 State comptroller race, the Hispanic-preferred candidate won in Proposed District 36 by 30,149 to 4319. (Id.)

89 . Dr. Stanley pointed to three elections that were excluded by Dr. McDonald — the 2000 and 1996 presidential elections and the 2002 attorney general election — where the minority-preferred candidate, in fact, “carried all of the existing districts.’’ (Stanley Aff. ¶ 109.) After conducting his own election analysis, he concluded that minorities already have an opportunity to elect candidates of their choice in SDs 34 and 35. (Id. ¶ 113; DX 157.) Dr. Stanley also criticized the inclusion of the 2002 gubernatorial election because, he ar gued, it was more racially polarized than endogenous elections. (Stanley Aff. V 109.) These points may be probative for the third Gingles precondition and the issue of whether racial bloc voting usually defeats Hispanic-preferred candidates, but they are not relevant to the issue of electability. The recompilation method presumes racial bloc voting; it does not purport to be evidence of it.

90 .Dr. McDonald used the following method in performing his supplemental electability analysis:- (1) calculate the average estimated turnout for each minority group; (2) calculate the average estimated vote share that the minority group's voters gave to the minority-preferred State Senate candidate; (3) multiply the two figures; (4) multiply the product of (3) by the minority group's percentage of the VAP in Plaintiffs' Proposed District 36; and (5) estimate how the total vote would be divided between the minority-preferred candidate and another candidate. (McDonald Decl. ¶ 74.)

91 . He also noted that in the one election (of the seven) in which there actually was a Hispanic candidate, that candidate received 45.5% of the vote. (McDonald Decl. ¶ 78.)

92 . Dr. McDonald also described a “warming” effect that would boost the electability of the district: “When minorities are given a chance of electing a-candidate of choice, it is likely that they will turn out to vote at higher rates.” (Id. ¶ 79.). As we did in the Nassau County electability analysis, we find that any predictions based upon an alleged warming effect are speculative.

93 . Together, SDs 34 and 35 comprise less than one-third of the plaintiffs' proposed majority-minority SD 36. The Bronx districts Dr. McDonald excludes from his analysis comprise more than two-thirds of Proposed District 36. (See DX 172.)

94 . Indeed, the parties have stipulated that: "Blacks and Hispanics cannot be assumed to be a single racial group with common political aims that are usually frustrated because of bloc voting by a hostile non-Hispanic white majority.” (Stip.i 133.)

95 . Dr. McDonald performed separate El anal-yses for elections in each district. Each statewide election — for example, the 2002 election for Governor — would generate results for two election contests: one in SD 34 and one in SD 35. On the other hand, an endogenous election for State Senate only involves one election contest. Thus if Dr. McDonald were analyzing the 2002 State Senate elections in SDs 34 and 35 and the 2002 election for Governor in those districts, there would be four "election contests” or four "election results” to be analyzed. We use these terms to distinguish between statewide and citywide elections and the multiple election contests or election results that they produce.

96 . In addition to the elections discussed below, Dr. McDonald mentioned in his declaration for trial two endogenous primary elections: the 2000 Democratic primaries for SDs 34 and 35. (See McDonald Decl. ¶ 24.) It is not clear, however, whether Dr. McDonald performed El and regression analyses for these primaries. Data for these primaries is not included with the data for all other election contests. (See DX 146.) There is no reference to these primaries in Dr. McDonald's Master Table, which includes results even for elections that he later excluded for the purpose of assessing cohesion. (Id.) Moreover, Dr. Stanley did not include these elections in his cohesion and racial bloc voting assessments. We therefore do not include them for the purpose of our analysis.

97 . All of Dr. McDonald's election analyses can be found in DX 146, and this table is a simplified and reorganized version of the Master Table of Racial Bloc Voting Analyses in the beginning of that report.

98 . Elections in bold include a minority candidate.

99 . An asterisk (*) reflects the minority-preferred (i.e., Hispanic and black) candidate, if any. The candidate who actually won the election is listed first.

100 . The source page refers to the page for the data for the individual election analyses as numbered in DX 146.

101 . The exogenous statewide elections analyzed for the Bronx claim are the same statewide elections analyzed for the Nassau County claim. See supra Table 5.

102 .Dr. McDonald explained that he excluded the results in SDs 34 and 35 for the 2002 Attorney General election because Irizarry, a Hispanic Republican, was "a relative unknown running against a popular incumbent," Eliot Spitzer, a white Democrat (id. ¶ 29). Dr. Stanley criticized Dr. McDonald's discounting of this election for the purposes of examining racial bloc voting. He. observed that “we can start taking elections out and putting them in, and recalculating averages and doing things of that sort and present other numbers, and they will be what they are, but they won't have this ... contest. Similarly, Governor Pataki was very popular in 2002[;] taking his contest out would ... affect results as well.” (Tr. at 668.)

103 . This table is based on the Master Table and the individual election analyses in DX 146. Elections refer to elections for New York State office unless otherwise specified.

104 . The Master Table at 5 (DX 148) reports that there was Hispanic and Hispanic-black cohesion in support of McCall, but the cohesion results for Hispanics were not statistically conclusive. There is statistical uncertainty as to whether Hispanics in SD 35 voted for McCall at a rate of 60% cohesion. (See McDonald Decl. ¶ 36 (corrected).)

Because it is the plaintiffs' burden to establish cohesion, this Table and the Tables to follow will categorize statistically inconclusive results as not showing cohesion. However, we will distinguish such results by marking them with an asterisk (No *).

105 . Dr. McDonald excluded the results because circumstances surrounding the elections purportedly diminished their probative value. The 1998 Democratic primary for lieutenant governor, Dr. McDonald claimed, included "candidates ... largely unknown in ... the Bronx[] and Westchester" and had low turnout. (Id. ¶ 28). He excluded the 2000 Democratic primary for CD 17 because of a "late-breaking” scandal involving the black candidate. (Id. ¶ 30.)

106 . This Table is based on the Master Table (Corrected) at DX 148. The data on which the Table is based is found in DX 146.

107 . "No *" indicates that Dr. McDonald did not find cohesion with statistical certainty. See supra note 104.

108 . The defendants dispute the cohesion findings for this election for the reasons explained in the discussion accompanying this Table and infra note 112.

109 . Engel is white, Seabrook is black, and Zayas is Hispanic. An estimated 51.5% of Hispanics supported Engel, a white candidate. Because of the standard errors in estimating Hispanic support of all candidates for this election, it cannot be statistically determined whether Engel received 60%, a majority, or even a plurality of the Hispanic vote.

110 . Dr. McDonald only generated estimates for King, the black candidate. He estimated that 30.1% of Hispanics supported King, and he thus deduced that Hispanics cohesively supported King's only major challenger, Frankel, a white candidate.

111 . SDs 34 and 35 comprise only 31.9% of the population of Proposed District 36 (with SD 34 constituting only 3.9%). (See Stanley Aff. ¶ 76; DX 172.) SDs 34 and 35 (under the 2000 census) comprise only 39.7% of the population of Proposed District 36 (with SD 34 constituting only 11.7%). (See Stanley Aff. ¶ 77; DX 171.)

112 . While the plaintiffs argued that the 2001 Bronx borough president race exhibited Hispanic-black cohesion, the defendants argued that it did not. (See Stip. ¶ 251; Joint Ex. 5 (noting disagreement); DX 146 at 52 (providing data for contest).) The dispute is over the fact, explained above, that Hispanics and blacks each split their vote evenly between two Hispanic candidates. Blacks and Hispanics thus cohesively voted for the Hispanic candidates against the white candidate, but there was no single cohesively supported minority-preferred candidate. The three primaries exhibiting Hispanic-black cohesion according to Dr. Stanley were thus the two 2001 mayoral primaries involving Ferrer and Green, and the 1997 Public Advocate primary.

113 . See supra Table 9, Table 10; (DX 146, 148).

114 . See supra note 104 and accompanying text.

115 . In stating that there was Hispanic-black cohesion in all six. exogenous contests, Dr. Stanley presumed that Hispanic cohesion was demonstrated in SD 35 for the 2002 Pa-taki/McCall election.

116 .Hispanics were cohesive in 28 of 33 (84.8%) of the elections included by Dr. McDonald in his cohesion assessment and in 31 of 38 (81.6%) of the elections included by Dr. Stanley. (See DX 148.)

117 .Dr. McDonald testified to the sufficiency of the cohesion evidence as follows:

Q [The defendants]: Well, do you think you have got enough information to make a professional judgment about the extent to which blacks and Hispanics are cohesive in elections generally?

A [Dr. McDonald]: We have got as much information as we have.

JUDGE WALKER: Is there a point where it’s reliable and a point where it’s specu-lativef?]

A: There is not a good, you know, definition of when it would and would not be.

JUDGE BERMAN: Would this data that you have here be closer to the speculative range or the reliable?

A: It’s borderline.

(Tr. at 159-60.)

118 . For example, Dr. Stanley cites 107 occasions out of 373 El estimates in which "the cohesion estimate for vote share for a candidate for the combined group is outside the range for the groups estimated separately.” (Stanley Aff. ¶ 16.)

119 . We agree with the plaintiffs that the 2001 Bronx borough president race exhibited Hispanic-black cohesion (see DX 146 at 52), although the defendants dispute this point (see Stip. ¶ 251; Joint Ex. 5 (noting disagreement)).

120 . All of Dr. McDonald's individual election analyses can be found in DX 146. This Table is a simplified and reorganized version of the Master Table of Racial Bloc Voting Analyses found at DX 148.

121

122 . ''HPC” refers to the Hispanic-preferred candidate, and the result reflects how that candidate fared in SD 34 or 35.

123 . Won by a plurality of 255 votes.

124 . The only exogenous elections with minority candidates were the 2002 Governor election and the 1998 comptroller election, both featuring McCall. For the exogenous elections in general, white bloc voting occurred in seven contests and the Hispanic-preferred candidate was defeated in six of those seven (85.7%) contests. In sum, for the seven endogenous elections and 18 exogenous general election contests. Dr. McDonald found white bloc voting (and what he termed "racially polarized voting”) in 14 of 25 elections (56.0%) and he concluded the Hispanic-preferred candidate was defeated in 13 of 14 racially polarized elections (92.9%).

125 . Refers to the result for the Hispanic-pre- ■ ferred candidate within the district.

126 . McCall was the Hispanic-preferred candidate, although Dr. McDonald could not conclude with statistical certainty that he was cohesively supported by at least 60% of Hispanic voters.

127 . Lost by a plurality of 139 votes.

128 . There was no statistically determinable ' HPC.

129 . Colon, a Hispanic candidate, received an estimated 51.4% of the Hispanic vote but it is inconclusive whether he received a majority of the Hispanic vote.

130 . There was no cohesively supported HPC. Messinger, a white candidate, received a plurality of the Hispanic vote and she "won” within the district.

131 . Because of large standard errors, no candidate can be said to be the HPC.

132 . Dr. McDonald estimated that 30.1% of Hispanics supported King, a black candidate. He thus considered Frankel, who is white, to have been the Hispanic-preferred candidate.

133 . See supra note 29 (listing Senate Factors); see also Gingles, 478 U.S. at 44-45, 106 S.Ct. 2752.

134 . The defendants' other proportionality calculations, based on a minority-coalition claim involving blacks and Hispanics, are as follows:

First, Hispanics and blacks combined are 76.1% of the Bronx CVAP (77.5% of the VAP), and 80% of. the districts are majority-minority (four of five: three Hispanic, one black).

Second, throughout New York City, Hispanics and blacks together constitute 46.9% of the CVAP (48.7% VAP), which would equate to 12.2 seats.of the 26 seats falling wholly or primarily in New York City. In fact, blacks and Hispanics control 14 of 26 (53.8%) seats under the Senate Plan.

Third, blacks are 25.8% of New York City's CVAP (24.0% VAP), which would equate to 6.7 of the 26 Senate seats. In fact, blacks have eight seats under the Senate Plan.

Fourth, in the 35 Senate districts (SDs 1-34, and 36) that fall entirely or primarily in New York City and Long Island, the combined Hispanic and black CVAP is 37.4%. Under the Senate Plan, blacks and Hispanics control 14 of 35 or 40% of those districts.

Fifth, throughout New York State, the (combined) black and Hispanic CVAP is 24.6% (28.3% VAP), which would equate to 15.3 Senate seats. In fact, blacks and Hispanics currently have the ability to elect candidates of choice or have a VAP majority in 15 districts.

Sixth, there are four districts in New York State — SDs 12, 16, 23, and 26- — in which black and Hispanic voters do not (together) constitute a VAP majority but do constitute a VAP majority when combined with other minority groups such as Asians — i.e., all voters other than non-Hispanic whites. (DX 62.) Minority groups thus combined constitute 29.4% of the CVAP, which would equate to 18.3 of 62 Senate seats. In fact, all groups thus combined have a VAP majority in 19 Senate seats. Stated in other terms, non-Hispanic whites constitute 70.6% of the State’s CVAP and 69.4% of the Senate seats are (VAP) majority white.

135 . The plaintiffs presented the following proportionality measures relevant to a Hispanic-black coalition claim:

First, in the six-county region, in addition to the five districts in which Hispanic voters can elect their candidate of choice, there are eight more districts in which black voters can elect candidates of their choice. The six-county region has a combined black and Hispanic VAP of 46.6% (43.2% CVAP), which is estimated to grow in 2004. (Beveridge Decl. ¶ 120.) There are 13 Senate districts in this region in which black and/or Hispanic voters currently can elect their candidate of choice, or 45.5% of 28.6 districts.

Second, the three-county region had a combined 48.6% black and Hispanic VAP (44.3% CVAP), which is estimated to grow during 2004. Under the Senate Plan, there are only five districts in this region in which black and/or Hispanic voters can elect their candidate of choice or 41.0% of 12.2 districts.

Third, the combined black and Hispanic statewide VAP is 28.3% (24.6% CVAP), but black and/or Hispanic voters can elect the representative of their choice in only 14 districts, or 22.6% of the Senate districts. (Id.)

136 . The Court in De Grandy posited a hypothetical jurisdiction in which a minority group comprising "40 percent of the [VAP] determines electoral outcomes in 7 out of 10 districts,” thus "enjoy[ing] effective political power 75 percent above, its' numerical strength.” Id., 512 U.S. at 1017 & n. 13, 114 S.Ct. 2647. It cautioned that "reading § 2 to define dilution as any failure to maximize tends to obscure the very object of the statute and to run counter to its textually stated purpose.” Id. at 1016-17, 114 S.Ct. 2647. Plaintiffs' Proposed District 36 goes against the grain of De Grandy. For example, in the Bronx, where the Hispanic CVAP is 42.0% and the Hispanic VAP is 45.2%, The plaintiffs propose that 80% of the districts be majority-Hispanic- — and 100% be majority-minority. See Barnett v. City of Chicago, 141 F.3d 699, 704-05 (7th Cir.1998) (Posner, C.J.) ("[T]he proper benchmark for measuring proportionality is citizen voting-age population .... ”); see also De Grandy, 512 U.S. at 1017, 114 S.Ct. 2647 ("Failure to maximize cannot be the measure of § 2.”).

137 .The Senate Plan includes three majority-Hispanic districts among the five Bronx-based districts (i.e., 60%). Under the Senate Plan (and based on CVAP), Hispanics currently "can be said to enjoy effective political power” 42.9% "above [their] numerical strength.” De Grandy, 512 U.S. at 1017 n. 13, 114 S.Ct. 2647 (providing formula for calculation, which in this case is the difference between 60% (percent of Bronx districts that are majority Hispanic) and 42.0% (Hispanic CVAP), divided by 42.0% (Hispanic CVAP)); see also Baines v. Masiello, 288 F.Supp.2d 376, 391-92 (W.D.N.Y.2003).

138 . We would also find proportionality considering Hispanics and blacks together. That is, blacks and Hispanics combined in the Bronx constitute 76.1% CVAP and have actual or effective voting majorities in 80.0% (or four of five) of the Bronx-based districts. Statewide, blacks and Hispanics combined constitute 24.6% of the CVAP and have actual or effective voting majorities in 24.2% (15 of 62) of the Senate districts. Blacks and Hispanics combined in New York City and Long Island constitute 37.4% of the CVAP and have actual or effective voting majorities in 40.0% (14 of 35) of the Senate districts. In New York City, Hispanics are 21.1% of the’CVAP and have actual or effective voting majorities in 23.1% (six of 26) of the Senate districts. Blacks and Hispanics combined in New York City constitute 46.9% CVAP and have actual or effective voting majorities in 53.8% (14 of 26) of the Senate districts.

139 . The gerrymandering claim is discussed infra in Part V, and the "upstate-downstate” claim is discussed supra in Part III.

140 . The plaintiffs cite unpersuasively to New York’s 2002 section 5 submission to support their proposed Hispanic and black CVAP and VAP reductions in the Bronx districts. In SD 28, “Hispanic voters in the new district will clearly remain able to elect their preferred candidates”; in SD 32, “Hispanic voters will continue to comfortably elect their preferred representative at this VAP level”; in SD 33, "Hispanic voters in the new district will retain their clear ability to elect candidates of choice”; and in SD 36, "[the] increase in African-American voting strength does not negatively impact electability in other districts because it is not possible to shift voters out of this district and create a new majority-African-American district.” (PX 27, at 8, 10-11.) While the submission speaks to the acceptability of the districts as included in the 2002 Senate Plan, it does not support further dilution of minority voting strength in those districts.

141 . The defendants also point out that, on the day before the redistricting vote was scheduled, when a fourth Hispanic district with a Hispanic VAP of only 44.8% (DX 62), was offered for the first time, no data was presented to suggest that a 44.8% Hispanic VAP district would have provided Hispanic voters with an opportunity to elect a candidate of choice. Nor was data presented to establish the electability of minority districts with the demographics proposed by the plaintiffs. (Burgeson Aff. ¶¶ 50, 52.)

142 . According to the statistics provided by LATFOR to the Department of Justice, the three Hispanic-majority districts were all drawn in 1992 to have Hispanic VAPs slightly over 60%. In 2002, the Hispanic VAPs in the three Hispanic-majority districts were drawn to be in the mid-50% range. The Hispanic VAP in the 2002 SD 28 was 7.4% less than in 1992 SD 28. The Hispanic VAP is 2002 SD 32 was 4.2% less than in 1992 SD 32. The Hispanic VAP in 2002 SD 33 was 5.3% less than in the predecessor district, 1992 SD 31. (See PX 27 at 7-11.)

143 . The underpopulation of "upstate” in the Senate Plan actually "prevents the creation of an additional Bronx-Westchester Hispanic-majority district such as Plaintiffs’ Proposed District 36,” as the plaintiffs themselves point out. (See Plaintiffs' Proposed Post-Tried Findings of Fact and Conclusions of Law ("PFF”) ¶512.) And because nothing resembling the Proposed District 36 was suggested until the day before the Legislature’s redistricting vote (see Tr. at 376 (Beveridge)), the population deviations could not have been “employed by the Legislature to deprive minority voters of the opportunity to elect candidates of choice in the Bronx” (PFF ¶ 504).

144 . See Allen v. McCurry, 449 U.S. 90, 94, 101 S.Ct. 411, 66 L.Ed.2d 308 (1980) ("[C]ollateral estoppel relieve[s] parties of the cost and vexation of multiple lawsuits, conserve[s] judicial resources, and, by preventing inconsistent decisions, encourage[s] reliance on adjudication.''); Fuchsberg & Fuchsberg v. Galizia, 300 F.3d 105, 109 (2d Cir.2002). We do not believe collateral estoppel applies here because, among other things, the timing, claims, parties and issues in this case are not the same as in the case cited by the plaintiffs.

145 . "The issue to be resolved ... is whether the State affords New York City schoolchildren the opportunity for a meaningful high school education, one which prepares them to function productively as civic participants. This is essentially the question the trial court addressed, and we conclude that the Appellate Division erred to the extent that it founded a judgment for defendants upon a much lower, grade-specific level of skills children are guaranteed the chance to achieve.” CFE, 769 N.Y.S.2d 106, 801 N.E.2d at 332; id. 769 N.Y.S.2d 106, 801 N.E.2d at 330 (“[A] sound basic education conveys not merely skills, but skills fashioned to meet a practical goal: meaningful civic participation in contemporary society.”).

146 . The parties have stipulated that:

• A 2002 report of the Joint Center for Political and Economic Studies reported that there was a jump of 332 percent in the number of black elected officials in New York State from 1970 to January 2000.

• With respect to Hispanics, the number with seats in the legislature has doubled over the past dozen years — a jump from 5 to 10 seats in the Assembly and from 2 to 4 in the Senate. There are currently nine Hispanic members on the New York City Council, and two Hispanics are members of the United States House of Representatives — Nydia Velazquez and Jose Serrano.

• Under the Senate Plan, the number of minority New York State senators has risen from 12 (8 black and 4 Hispanic) to 13 (9 black and 4 Hispanic).

(See generally Stip. ¶¶ 141-45.)

147 . The claim, as alleged in the Complaint, was based on the Senate Plan’s failure to create two additional majority-minority districts (to make sixteen total). (See Compl. ¶ 295.) The two additional districts the plaintiffs had contemplated were Proposed District 36 in the Bronx and Proposed District 8 in Nassau County. But as the plaintiffs acknowledged, the upstate-downstate population distribution has no effect on Long Island, and Proposed District' 8 is not part of this count. (See Summ. J. Tr. at 54.)

148 . Dr. de la Garza used the terms "Latino” and "Hispanic” interchangeably in his report. (See de la Garza Report at 3 n. 3.) While we have used the label "Hispanic” throughout this opinion, we will generally use the label "Latino” for this count to reflect the terminology used most often by the LVRC intervenors.

149 . There is no suggestion in this claim that blacks and Latinos vote cohesively in either SD 31 or Proposed District 32.

150 . SD 31 was essentially created anew in the 2002 Senate Plan, drawing 39.40% of its population from the 1990s SD 28; 39.15% of its population from the old SD 30; 21.35% of its population from the old SD 29; and 0.10% of its population from the old SD 34. (Braatz Decl. Ex. 13.) Thus the only endogenous Senate election for SD 31, as it currently exists, was in 2002.

Dr. de la Garza analyzed the local exogenous primaries based on the precincts that are now encompassed in the current SD 31. He did not analyze endogenous State Senate elections for the portions of SDs 28, 29, and 30 that now constitute SD 31.

151 . It does not say in Dr. de la Garza's report that Ramirez failed to receive the most votes in the areas constituting the current SD 31, but the numbers strongly suggest that he did.

152 . As explained supra note 150, SD 31 was essentially created anew by the 2002 Senate Plan, but Schneiderman was the incumbent for the predecessor district, which encompassed the areas of the Upper West Side and Riverdale that constituted substantial portions of the 2002 SD 31. (See Chin Aff. ¶ 29.)

153 . The defendants count the population of COBED-CD 17 somewhat differently. They find that the non-Hispanic black VAP would be 35.77%, the Hispanic VAP would be 30.20%; the Asian VAP would be 3.75%; and - the white. VAP would be 26.85%. (See Chill Aff. ¶ 25.) The different counts between the parties results from variations in the classifications of individuals who identify with more than one race in response to census questions. For the purposes of the summary judgment motion, we assume that the Rivers/Barbour intervenors’ numbers are correct.

154 . To the extent that the Rivers/Barbour in-tervenors argue in the alternative that COBED-CD 17 should be construed as a black or Hispanic "influence district,” we do not find such claims cognizable under section 2 of the Voting Rights Act, as we have explained with respect to the Suffolk County claim. Moreover, here, it is not even clear which minority group should be construed to have the relevant influence, blacks or Hispanics, or what percentage of the VAP would satisfy the "influence” claim. Finally, the Rivers/Barbour intervenors did not offer any evidence of either group's “ability to elect” on its own.

155 . Cohesion could be suggested Without blacks and Hispanics preferring the same candidate if, for example, black and Hispanic voters split their votes among two Hispanic candidates while voting against candidates of other races. That situation is not raised by the election results relied on by the Rivers/Barbour intervenors.

156 . While Hispanics seem more cohesive with whites in these elections, there are also examples of elections where blacks cohere with whites — or at least prefer white-preferred candidates to Hispanic-preferred candidates. For example, in the 1994 congressional Democratic primary between William Colon and Eliot Engel, Hispanics preferred Colon while blacks and whites preferred En-gel. Similarly, in the 2001 Democratic primary for Public Advocate, Hispanics again preferred Colon, but blacks split their vote between two white candidates, Betsy Gotbaum and Norman Siegel. (See Chill Aff. Ex. A at 11-17.) These elections, offered by the defendants and ignored by the Rivers/Barbour intervenors in their expert reports, further support the claim that blacks and Hispanics are not politically cohesive in CD 17.

157 . Almost all, 96.8%, of the Bronx portion of SD 34 as enacted in 2002 comes from the preexisting portion of SD 34 in Bronx County. (Burgeson Aff. ¶ 32.)

158 . In Shaw I, the plaintiffs challenged two majority-black districts. See Shaw, 509 U.S. at 636-37, 113 S.Ct. 2816; see also Vera, 517 U.S. at 957, 116 S.Ct. 1941 (challenging three majority-black and majority-Hispanic districts); Miller, 515 U.S. at 907-08, 115 S.Ct. 2475 (challenging three majority-black districts); Clark v. Putnam County, 293 F.3d 1261, 1265 (11th Cir.2002) (challenging two majority-black districts); Diaz v. Silver, 978 F.Supp. 96, 103-05 (E.D.N.Y.) (challenged district was majority-Hispanic), aff'd mem., 522 U.S. 801, 118 S.Ct. 36, 139 L.Ed.2d 5 (1997).

159 . In fact, incumbent State Senator Guy J. Vellela was preceded in office by another Republican, John D. Calandra, who served as State Senator of SD 34 from 1973 to 1986. Sen. Calandra represented SD 33 from 1967 to 1972 and SD 38 in 1966. The New York Red Book, supra note 78, at 197.

160 . For voter enrollment information, see N.Y. State Bd. of Elections, at http://www.elections.state.ny.us/enrollment/enroll.htm (last visited Mar. 14, 2004). The enrollment figures are as of June 27, 2002. (See also DX 116-21.)

161 . The towns and cities of Westchester County covered by SD 34 are individually named in N.Y. State Law § 124 (McKinney 2002).

162 . See The New York Red Book, supra note 78, at 197.

163 . The record offers only partial evidence of district shape from the 1980s and none for the period before 1980. Senate district maps and statutory contours of each Senate district are published in the New York State Legislative Manual, which is prepared by the New York Secretary of State pursuant to N.Y. Executive Law § 95. The 1980s district lines for SD 34 can be found at New York State Legislative Manual 831-33 (1986-1987).

164 . (See DX 106; PX 9 tbl.2; Beveridge Deck ¶ 53.)

165 . The black VAP increase is based upon the difference between the percentage of blacks in the old and new districts excluding mixed-race responses. The record does not reflect the comparative numbers including mixed-race responses, which are the numbers used elsewhere in this decision.

166 . The seventh compactness measure, called the sum of perimeters, is not meaningful in this context.

167 . "The' racial composition of the Westches-ter portion changed to a slightly greater degree, but this was simply through the transfer of a white town (Eastchester) from one majority-white district to another.” (Burgeson Aff. ¶ 46.)

168 ."The shape of District 34 in the 1990s was affected by the shape of the majority black Bronx-Westchester district drawn adjacent to District 34, which empowered black voters.” (Burgeson Aff. ¶ 42.)

169 . In fact, at least two districts have crossed the Bronx County border since 1972. See New York State Legislative Manual 962-63, 966-67 (1973) (SDs 30, 35); id. at 829-33 (1986-1987) (SDs 30, 34); (DX 6 (SDs 28, 30, 33, and 34 (twice))).

170 . The Court also observed that "[t]he Majority Leader [Sen. Marino] has marshaled a considerable amount of statistical and demographic data to support his contention that these districts were drawn in a 'good faith effort’ (Matter of Schneider v. Rockefeller, supra, at 428, 340 N.Y.S.2d 889, 293 N.E.2d 67) to comply with Reynolds v. Sims and the Voting Rights Act and not for partisan political reasons, as petitioners argue." Wolpoff, 80 N.Y.2d at 78-79, 587 N.Y.S.2d 560, 600 N.E.2d 191. It would be an unsustainable leap of logic to conclude from these remarks in Wolpoff that the defendants here are es-topped from arguing that politics played a role in the 2002 Senate districting processes. (See PFF ¶¶ 685-86.)

171 . "That an alternative plan might have been devised that conflicted less with article III, § 4 but did greater violence to the equal representation principle is no basis for rejecting the Senate plan. Further, we cannot focus solely on the challenged districts and ignore the fact that a redistricting plan must form an integrated whole.” Wolpoff, 80 N.Y.2d at 79, 587 N.Y.S.2d 560, 600 N.E.2d 191.

172 . The defendants have also presented written materials, including transcripts from several LATFOR public hearings and meetings (DX 126-28, 131-33) and the Assembly and Senate floor debates (DX 134), which took place prior to passage of the Senate Plan. Sen. Skelos and Democratic members of LAT-FOR spoke at these meetings.

173 . The defendants argued at the summary judgment stage that "[s]ince there is a blaz-ingly obvious political explanation for the legislature's District 34, the plaintiffs cannot show that it is unexplainable on this political ground and, since the political explanation is so blazingly obvious, no member of the legislature needed to be handed political data to know the legislature’s District 34 was better for Senator Velella than the overwhelmingly Democratic Hispanic majority district (and, in any event, all such political data was readily available on LATFOR’s website).” (Defs. Reply at 15.)

174 . In 2002, 54 of the 56 incumbent senators who sought reelection won. In addition, because the Senate Plan paired two incumbent Democrats in SD 23, one senator ran for the "open” seat in SD 22 and lost. See N.Y. State Bd. of Elections, 2002 State Senate General Elections Results, at http:// www.elec-tions.state.ny.us/elections/2002/general/2002 _sen.pdf (last visited Mar. 3, 2004); (DX 93 (listing incumbents assigned to each district under 2002 Senate Plan)); see also Martinez v. Bush, 234 F.Supp.2d 1275, 1307 n. 36 (S.D.Fla.2002) (three-judge court) (taking judicial notice, pursuant to Fed.R.Evid. 201(b)(2), of election results posted on Florida Department of State website).

175 .Burgeson, who testified in this case as one of the defendants’ experts, has been employed by the LATFOR Task Force (or the Senate majority offsite redistricting office) since 1980.

176 . The plaintiffs also contend thát in 1992 the Senate majority was following a “maximization policy" that the Supreme Court would later criticize in Miller, 515 U.S. at 920-27, 115 S.Ct. 2475, and other decisions, by pointing out that Mr. Burgeson’s 1992 affidavit notes that a 1992 goal was to “[e]stablish, to the extent possible, districts in which minorities comprise a májority of the voting age population, as required by the Voting Rights Act.” (PX 30 ¶ 3(b); see also PX-31 at 15, 52-58.) We do not believe the quoted remarks establish a 1992 legislative "maximization policy" or racial predominance.

177 . The Senate redistricting plans of 1992 and 2002 were submitted to the Department of Justice for preclearance, and both times the three majority-Hispanic, one majority-black, and one majority-white district configuration of the Bronx was approved. (See DX 138 (Letter from Ralph F. Boyd, Jr., Ass't Attorney General, U.S. Department of Justice, to Michael A. Carvin, Esq., Jones, Day, Reavis & Pogue, et al., (June 17, 2002)); DX 137 (Letter from Sen. Dean G. Skelos and Assemblyman William L. Parment, Co-Chairmen, LAT-FOR, to Chief of Voting Section, Civil Rights Division, U.S. Department of Justice (Apr. 30, 2002) ("The redistricting plans for the Senate and Assembly promulgated after the 1990 census were precleared in July 1992; 29 CFR 51.27(p). All court challenges to the 1990 plan were dismissed.”)).)

178 . The May 4, 2001 memorandum was sent to Sen. Skelos and Steve Boggess (whose affiliation has not been established), and copied to Vinny Bruy and is entitled "Reapportionment Areas” ("R/As”). It sets forth the eleven R/As that were under consideration, including a Queens R/A, Brooklyn R/A, and Hudson R/A, which included Richmond (Staten Island), New York (Manhattan), Bronx, Westchester, Putnam, Dutchess, and Columbia Counties. (PX 18 at 1-3.) The June 21, 2001 memorandum, sent to Sen. Skelos, Steve Boggess, and Vinny Bruy and entitled "Westchester,” "discuss[es] the particulars of how [Mr. Burge-son] determined the population total in Westchester to be connected to Bronx.” (PX 19 at 1.) Three alternative versions of the Hudson R/A are set forth, all of which state "[tjotal left in Westchester for SD's [old] 33 & 34.” Under two scenarios, approximately 131,000 total population is "left for SD's 33 & 34”; under the third scenario, 135,033 people are designated for SDs 34 and 35. (Id. at 2.)

179 .Dr. Grofman’s affidavit, dated May 11, 2002, was submitted to the Court in connection with the development of Special Master Lacey’s congressional plan. (DX 141 ¶ 2-5; see also Burgeson Aff. ¶ 6 ("The 2002 Senate plan furthers the important neutral principle of preserving the cores of the then-existing Senate districts to the extent practical, while keeping cognizant of population shifts.”).)

180 . As the parties have stipulated, no Republican senator has ever been paired with another incumbent senator, Republican or Democrat, in the history of redistrieting in New York. (Stip^ 234.)

181 . We also agree with the defendants that the contours of SD 33 in 1992 were likely influenced, in part, by section 5 of the Voting Rights Act, which required the Legislature in 1992 to preserve (not diminish) the sole ma jority-minority black district in the Bronx. (Tr. at 268-69 (Beveridge).) In 2002, the four majority-minority districts in the Bronx were maintained in accordance with section 5. (See DX 138.)

182 . “There are ... a number of important nonracial principles directly served by preserving District 34, and its shape was not related to diluting minority voting strength or segregating any racial group.” (Burgeson Aff. ¶ 31.)

183 . Judge Maas generally ruled in favor of document disclosure in this case but, describing these documents as "political,” he concluded that "disclosure ... would add nothing to the discussion of the legality of the Senate majority’s final plan if, as is generally the case, the request was couched in political rather than racial or ethnic terms." Id.

184 . To take one example from our discussion of Count III, the plaintiffs’ expert, Dr. McDonald, estimated that in the 2000 SD 34 Senate Democratic primary, 72.2% of whites supported the winning candidate (Lorraine Coyle Koppell), as compared to 61.7% of black voters and 46.2% of Hispanic voters. (McDonald Decl. ¶ 24.) But in the general election against Republican Sen. Velella, Kop-pell received nearly unanimous support of black and Hispanic voters (99.3% and 99.2% respectively), but only 31.9% from white voters (and was defeated).