Nos. 99-1864 and 99-1865
In the Supreme Court of the United States
JAMES B. HUNT, JR., GOVERNOR OF NORTH CAROLINA,
ET AL., APPELLANTS
MARTIN CROMARTIE, ET AL.
ALFRED SMALLWOOD, ET AL., APPELLANTS
MARTIN CROMARTIE, ET AL.
ON APPEAL FROM THE UNITED STATES DISTRICT COURT
FOR THE EASTERN DISTRICT OF NORTH CAROLINA
BRIEF FOR THE UNITED STATES
AS AMICUS CURIAE SUPPORTING APPELLANTS
SETH P. WAXMAN
Counsel of Record
WILLIAM R. YEOMANS
Acting Assistant Attorney
BARBARA D. UNDERWOOD
Deputy Solicitor General
JAMES A. FELDMAN
Assistant to the Solicitor
DAVID K. FLYNN
LOUIS E. PERAERTZ
Department of Justice
Washington, D.C. 20530-0001
The United States will address the following question:
Whether the district court applied the correct legal standards in findingthat race was the predominant factor in the drawing of District 12 of NorthCarolina's 1997 congressional redistricting plan.
In the Supreme Court of the United States
JAMES B. HUNT, JR., GOVERNOR OF NORTH CAROLINA,
ET AL., APPELLANTS
MARTIN CROMARTIE, ET AL.
ALFRED SMALLWOOD, ET AL., APPELLANTS
MARTIN CROMARTIE, ET AL.
ON APPEAL FROM THE UNITED STATES DISTRICT COURT
FOR THE EASTERN DISTRICT OF NORTH CAROLINA
BRIEF FOR THE UNITED STATES
AS AMICUS CURIAE SUPPORTING APPELLANTS
INTEREST OF THE UNITED STATES
This case concerns a district court's finding that a state election districtingplan was drawn predominantly on the basis of race, in violation of the EqualProtection Clause of the Fourteenth Amendment. The United States enforcesSections 2 and 5 of the Voting Rights Act of 1965 (42 U.S.C. 1973, 1973c),which require, in part, that States and political subdivisions not engagein voting practices that deny citizens an equal opportunity to elect representativesof their choice on account of their race. Those statutes sometimes requireStates to take the racial consequences of their districting decisions intoaccount. The United States has an interest in ensuring that States havereasonable leeway to design districts that comply with both the Voting RightsAct and the Equal Protection Clause. The United States has participatedin all three prior appeals in related litigation. The United States wasa party-defendant in Shaw v. Reno, 509 U.S. 630 (1993), and filed briefsas amicus curiae in Shaw v. Hunt, 517 U.S. 899 (1996), and in Hunt v. Cromartie,526 U.S. 541 (1999).
1. In Shaw v. Hunt, 517 U.S. 899 (1996) (Shaw II), this Court struckdown North Carolina's 1992 congressional districting plan under the EqualProtection Clause of the Fourteenth Amendment. The Court held that District12 in that plan had been drawn predominantly on the basis of race, id. at907, and that it did not satisfy strict scrutiny, id. at 910-918.
After this Court's decision, the North Carolina General Assembly attemptedto enact a new districting plan. The state Senate had a Democratic majorityand the House had a Republican majority. State Senator Roy A. Cooper, III,and State Representative W. Edwin McMahan, the chairmen of the Senate andHouse redistricting committees, provided affidavits and testimony detailingthe goals and purposes of the committees. J.S. App. 81a-87a; J.A. 179-230(Cooper); J.S. App. 137a-154a; J.A. 231-244 (McMahan). Among the avowedgoals of the committees were "curing the constitutional defects ofthe 1992 Plan by assuring that race was not the predominant factor in thenew plan" and "drawing the plan to maintain the existing partisanbalance." J.S. App. 11a. To achieve that partisan goal, "the redistrictingcommittees drew the new plan (1) to avoid placing two incumbents in thesame district and (2) to preserve the partisan core of the existing districtsto the extent consistent with the goal of curing the defects in the oldplan." Ibid.
District 12 in the 1997 Plan is different from the district found unconstitutionalin Shaw II in important respects. As this Court noted in its prior decisionin this case, Hunt v. Cromartie (Hunt I), 526 U.S. 541, 544 (1999), District12 splits six counties, as opposed to ten in the unconstitutional plan.The distance between its farthest points has been reduced from 160 milesto 95 miles. Ibid. African-Americans are no longer a majority in the district,constituting approximately 43% of its voting age population, 46% of registeredvoters, and 47% of its population. Ibid. District 12 is also fully contiguousand, unlike the unconstitutional District 12 in the 1992 plan, it does notemploy artificial devices such as "crossovers" to achieve contiguity.J.S. App. 83a.
The 1997 Plan was enacted by the legislature on March 31, 1997, despitean earlier belief by many that the party division between the two housesof the legislature would make such agreement impossible. J.S. App. 2a, 82a,138a; J.A. 240. Twelve of the 17 African-American members of North Carolina'sHouse of Representatives voted against the plan. J.S. App. 140a.
2. a. Appellees filed an amended complaint alleging that District 12under the 1997 Plan is, like its predecessor, an unconstitutional gerrymander.See Hunt I, 526 U.S. at 544. The parties filed competing motions for summaryjudgment and, in April 1998, the district court, by a 2-1 majority, grantedappellees' motion. Id. at 545; see J.S. App. 243a-282a.
b. On May 17, 1999, this Court in Hunt I unanimously reversed the ordergranting summary judgment to appellees. The Court noted that "[t]hetask of assessing a jurisdiction's motivation * * * is an inherently complexendeavor" and that it "requir[es] the trial court to perform asensitive inquiry into such circumstantial and direct evidence of intentas may be available." 526 U.S. at 546 (citation omitted). Assessingthe summary judgment record, the Court noted that appellees had offered"circumstantial" evidence consisting of geographic and demographicdata, id. at 547, which, "[v]iewed in toto, * * * tends to supportan inference that the State drew its district lines with an impermissibleracial motive-even though they presented no direct evidence of intent."Id. at 548-549. The Court also noted, however, that appellants had producedtestimony by the legislators who drew the plan that their intent was "tomake District 12 a strong Democratic district," and what the Courtdescribed as "[m]ore important" expert testimony examining thedemographics and the entire boundary of the district. Id. at 549. That testimonytended to show "a high correlation between race and party preference,"id. at 552, because "in precincts with high black representation, thereis a correspondingly high tendency for voters to favor the Democratic Party"and vice versa, id. at 550. The expert, Dr. David W. Peterson, concludedthat "the data as a whole supported a political explanation at leastas well as, and somewhat better than, a racial explanation" for theconfiguration of District 12. Ibid.
The Court noted that a political explanation for District 12 would makethe district constitutional, since "a jurisdiction may engage in constitutionalpolitical gerrymandering, even if it so happens that the most loyal Democratshappen to be black Democrats and even if the State were conscious of thatfact." 526 U.S. at 551. To reject that political explanation, the districtcourt had necessarily "either credited appellees' asserted inferencesover those advanced and supported by appellants or did not give appellantsthe inference they were due." Id. at 552. In either event, "itwas error in this case for the District Court to resolve the disputed factof motivation at the summary judgment stage." Ibid.
3. On remand, the three-judge district court held a three- day trial.On March 7, 2000, the court ruled by a 2-1 margin that District 12 "continuesto be unconstitutional." J.S. App. 35a.
a. The majority initially repeated, virtually verbatim, many of the samefacts regarding the racial composition, party registration, and statisticalmeasures of compactness that it had relied on in granting summary judgmentto appellees. That evidence tended to show that cities and counties weredivided such that the portions within District 12 had substantially higherpercentages of African-Americans than the portions outside District 12,see J.S. App. 12a-14a, and that the boundary of District 12 excluded certainprecincts in which 54-69% of the voters had registered as Democrats, seeid. at 13a-14a. It also showed that District 12 scored relatively low onstatistical measures of compactness. Id. at 15a-17a. Compare J.S. App. 247a-253a(district court opinion at summary judgment stage).
The majority also referred to evidence presented by plaintiffs' expert,Dr. Ronald Weber. According to the court, Dr. Weber "showed time andagain how race trumped party affiliation in the construction of the 12thDistrict and how political explanations utterly failed to explain the compositionof the district." J.S. App. 26a. The majority also stated that Dr.Weber had "presented a convincing critique" of the "boundarysegment" analysis presented by the State's expert, Dr. David Peterson,and discussed by this Court in its opinion in Hunt I, and that Dr. Weberhad found that Dr. Peterson's study "'has not been appropriately done,'and was therefore 'unreliable' and not relevant." Id. at 27a. The majoritydid not itself specify the particular respects in which Dr. Peterson's analysiswas deficient.
The majority finally referred to two other items of evidence to supportits conclusion that race, and not politics, was the predominant factor underlyingthe creation of District 12. First, the majority referred to the testimonyof Senator Cooper. The majority stated that "[t]he conclusion thatrace predominated was * * * bolstered by" an allusion by Senator Cooperto a desire to achieve "racial and partisan balance" as factorsunderlying the redistricting plan. J.S. App. 27a. The majority found "simplynot credible" Senator Cooper's contention that "he did not meanthe term 'racial balance' to refer to the maintenance of a ten-two balancebetween whites and African-Americans." Ibid. Second, the court referredto an e-mail to Senator Cooper that had been written by Gerry Cohen, thelegislative employee who had been responsible for technical aspects of drawingthe 1997 and earlier state plans. See id. at 8a. The e-mail discussed theracial composition of a different district-District 1- and then added that"I [Cohen] have moved Greensboro Black community into the 12th, andnow need to take [a]bout 60,000 out of the 12th. I await your directionon this." Ibid.; see J.A. 369 (full text of e-mail). The majority statedthat the e-mail "clearly demonstrates that the chief architects ofthe 1997 Plan had evolved a methodology for segregating voters by race,and that they had applied this method to the 12th District." J.S. App.27a.
The majority concluded that the legislature had "eschewed traditionaldistricting criteria such as contiguity, geographical integrity, communityof interest, and compactness in redrawing the District," but insteadhad "utilized race as the predominant factor in drawing the District."J.S. App. 29a. The court entered an injunction against use of District 12in this year's elections. Id. at 35a.1
b. Judge Thornburg dissented from the panel's holding that District 12is an unconstitutional racial gerrymander. J.S. App. 37a-68a. In his view,appellees-who had the burden of proving that race was the predominant factor-had "failed to carry their burden through either direct or circumstantialevidence." Id. at 45a. He stated that the State had "producedample and convincing evidence which demonstrates that political concernssuch as existing constituents, incumbency, voter performance, commonalityof interests, and contiguity, not racial motivations, dominated the processsurrounding the creation and adoption of the 1997 redistricting plan."Id. at 45a-46a. He noted that the 1997 Plan's drafters "recognizedthe necessity of creating a plan which would garner the support of bothparties and both houses" by "protect[ing] incumbents and therebymaintain[ing] the then existing 6-6 partisan split amongst North Carolina'scongressional delegation." Id. at 46a. Since District 12 had a Democraticincumbent, "common sense as well as political experience dictated ascertainingthe strongest voter performing Democratic precincts in the urban PiedmontCrescent." Id. at 47a. The fact "[t]hat many of those strong Democraticperforming precincts were majority African-American, and that the GeneralAssembly leaders were aware of that fact, is not a constitutional violation."Ibid.
Judge Thornburg addressed Dr. Weber's testimony that District 12 wasdrawn on a predominantly racial, and not political, basis, because the Districtfailed to include some Democratic precincts that had relatively low African-Americanpopulations. Judge Thornburg noted that "there is no dispute that everyone of the majority African-American precincts included in the Twelfth Districtare among the highest, if not the highest, Democratic performing districtsin that geographic region." J.S. App. 50a. He noted that to includeother well-performing Democratic precincts identified by Dr. Weber wouldhave meant excluding "the highest performing Democratic precincts."Ibid. He also explained that "few of the strong Democratic precinctsto which Dr. Weber referred could have easily been included in the TwelfthDistrict" because few of them "actually abutted" the District.Id. at 50a n.21. Judge Thornburg also noted Dr. Weber's testimony that hehad "considered no hypothesis other than race as the legislature'spredominant motive" because he had believed, mistakenly, "thatthe person drawing North Carolina's districts could only see racial data"on his computer screen. Id. at 51a. Finally, Judge Thornburg noted thatDr. Weber had also "specifically failed to inquire about real worldpolitical or partisan factors which might have influenced the process."Ibid.
With respect to the Cooper-Cohen e-mail, Judge Thornburg explained thatit "does little more than reinforce what is already known, and whatis not constitutionally impermissible: North Carolina's legislative leaderswere conscious of race, aware of racial percentages, on notice of the potentialconstitutional implications of their actions, and generally very concernedwith these and every other political and partisan consideration which affectedwhether or not the redistricting plan would pass." J.S. App. 48a n.18.Those facts "contribute little to [appellees'] efforts to show thatracial motives predominated." Ibid.
4. On March 16, 2000, this Court entered an order staying the districtcourt's injunction. 120 S. Ct. 1415.
SUMMARY OF ARGUMENT
This case presents the Court with what is likely to be its final opportunityto clarify the legal standards governing a racial gerrymandering claim beforestate legislatures begin the redistricting process triggered by the decennialcensus. Both in the Shaw context and elsewhere, this Court has frequentlyemphasized the extraordinary sensitivity of redistricting and high costsof unnecessary federal court intrusion into the primary authority of theStates in this area. For those reasons, it is crucial that the "predominantfactor" test that governs a racial gerrymandering claim not be interpretedto give district courts a free-ranging license to substitute their judgmentsfor those of state legislatures in the quintessentially political determinationof how appropriately to draw electoral districts.
This Court's decisions have established that a district is subject tostrict scrutiny when it is drawn with race as the predominant factor; theplaintiff must prove that traditional race-neutral districting principleswere subordinated to race-not to some other factor-before strict scrutinyapplies. As the Court has repeatedly explained, the "predominant factor"test is a demanding one. It does not license a district court to intrudein the core state function of redistricting merely because the State hasdrawn a district that is majority-minority or that has a higher minoritypopulation than neighboring districts. Nor does it permit a district courtto intrude in state redistricting merely because racial considerations werea factor among others in drawing a particular district or in making someof the subsidiary districting decisions that go into a districting plan.Rather, a district court may intrude in districting in this context onlyif the State's dominant and controlling rationale was race.
Under that standard, the district court in this case erred in concludingthat the predominant factor in drawing District 12 was racial. First, thedistrict court relied substantially on evidence that was incompetent todistinguish between race and politics as a factor responsible for the configurationof District 12. The crucial and uncontroverted fact is that in North CarolinaAfrican-Americans reliably vote overwhelmingly-90% or more-for Democraticcandidates. Accordingly, any district that, like District 12, is drawn toconcentrate reliable Democratic voters will tend as well to concentrateAfrican-American voters. The evidence on which the district court reliedthat District 12 is unusually shaped in a way that tends to correspond withrace thus tends only to frame the question-whether the district was drawnwith race or political motives as predominant-but not to answer it. Thedistrict court also relied on evidence showing that District 12 fails toinclude some precincts with high Democratic registration figures. But ina State like North Carolina, in which registered Democrats frequently voteRepublican, that evidence is entirely consistent with the legislature'sprofessed desire to create a district that would be solidly Democratic onelection day, and it provides no basis for doubting the State's professedpolitical motive.
Second, the district court committed clear error in inferring from certainevidence presented by appellees' expert, Dr. Ronald Weber, that race wasthe predominant motive underlying District 12. For example, the districtcourt relied on Dr. Weber's testimony that precincts that had voted forDemocratic candidates in past elections were omitted from District 12. Theevidence on which the court relied would have been sufficient to tend todisprove the State's partisan objective of creating District 12 as a solidlyDemocratic district only if appellees had shown that including the omittedprecincts would have resulted in a District 12 with a higher overall Democraticvoting strength. Omitting precincts with Democratic voting patterns in favorof precincts with even more solidly Democratic voting patterns is entirelyconsistent with the State's professed objective. It cannot support an inferenceof predominant racial motive.
Third, the district court, based in large part on the faulty inferencesdiscussed above, inferred a predominant racial motive from statements bySenator Cooper and legislative employee Gerry Cohen. Insofar as the districtcourt's inferences in this regard were based on its earlier errors, thecourt's conclusions should be disregarded. In any event, however, the inferencesthe district court drew from these statements showed at most that race wasa factor underlying District 12. The district court thus failed to distinguishbetween a State's mere desire to achieve a racial objective in districtingas one factor among others and the desire to achieve a racial objectiveas a predominant motive underlying District 12; only the latter is subjectto strict scrutiny. Where there is as close a coincidence of race and politicsas in this case, a district court may not conclude that race was the predominantfactor based solely on isolated findings that in particular respects theprocess or result of the State's districting shows that the State was awareof the racial consequences of its actions or that race was a factor; todo so would leave States that engaged in entirely constitutional districtingat risk of a district court's inference that, because they had some racialknowledge or motivation, it must have been predominant.
I. THE PREDOMINANT FACTOR TEST REQUIRES A DISTRICT COURT TO ENGAGE INA PARTICULARLY SENSITIVE INQUIRY INTO A STATE'S INTENT IN DRAWING A DISTRICT,AND IT REQUIRES PROOF NOT MERELY THAT RACE WAS A FACTOR IN DRAWING A DISTRICT,BUT THAT IT WAS THE PREDOMINANT FACTOR
A. A Shaw Claim Requires Proof That Race Was The State's "PredominantFactor"
In Shaw v. Reno, 509 U.S. 630 (1993) (Shaw I), this Court first recognizeda claim for racial gerrymandering in violation of the Equal Protection Clause.In Miller v. Johnson, 515 U.S. 900 (1995), the Court articulated the governingstandard: strict scrutiny is triggered only when "race for its ownsake, and not other districting principles, was the legislature's dominantand controlling rationale in drawing its district lines." Id. at 913.Race must thus be shown to be "the predominant factor motivating thelegislature's [re-districting] decision." Bush v. Vera, 517 U.S. 952,959 (1996) (plurality opinion) (emphasis in original); see also Shaw v.Hunt, 517 U.S. 899, 905 (1996) (Shaw II); Lawyer v. Department of Justice,521 U.S. 567, 582 (1997); Hunt v. Cromartie, 526 U.S. 541, 547 (1999) (HuntI).
The "predominant factor" test is not the same inquiry applicable"in cases of 'classifications based explicitly on race,'" Bush,517 U.S. at 958, or in cases in which facially neutral practices are challengedon the ground that race is a "motivating factor in the decision,"Village of Arlington Heights v. Metropolitan Housing Dev. Corp., 429 U.S.252, 266 (1977). A necessary consequence of the Court's holding that a districtis subject to strict scrutiny only when race was the State's "predominantfactor" in drawing it is that a Shaw claim is not made out when raceis merely one of the motives or factors considered-but not the predominantone-in drawing the district. Indeed, the plurality in Bush made that pointexpressly, rejecting the view "that it suffices [in making out a Shawclaim] that racial considerations be a motivation for the drawing of a majority-minoritydistrict." Bush, 517 U.S. at 959 (emphasis in original). In short,"[s]trict scrutiny does not apply merely because redistricting is performedwith consciousness of race," id. at 958, "[n]or * * * is the decisionto create a majority-minority district objectionable in and of itself,"id. at 962. As Justice O'Connor has explained, under the "predominantfactor" test, "States may intentionally create majority-minoritydistricts, and may otherwise take race into consideration, without comingunder strict scrutiny." Id. at 993 (O'Connor, J., concurring).
B. The Predominant Factor Test Is A Demanding One
1. This Court has frequently noted, both in Shaw cases and in other redistrictingcases, that "redistricting and reapportioning legislative bodies isa legislative task which the federal courts should make every effort notto pre-empt." Wise v. Lipscomb, 437 U.S. 535, 539 (1978). Of course,federal courts serve a "customary and appropriate backstop role,"Bush, 517 U.S. at 985, when a state redistricting plan "runs afoulof federal law," Lawyer, 521 U.S. at 577. But because "reapportionmentis primarily the duty and responsi- bility of the State," Chapman v.Meier, 420 U.S. 1, 27 (1975), and is "a most difficult subject forlegislatures," Miller, 515 U.S. at 915, "the States must havediscretion to exercise the political judgment necessary to balance competinginterests," ibid. "The task of redistricting is best left to statelegislatures, elected by the people and as capable as the courts, if notmore so, in balancing the myriad factors and traditions in legitimate districtingpolicies." Abrams v. Johnson, 521 U.S. 74, 101 (1997). See also Growev. Emison, 507 U.S. 25, 34 (1993). Because of the serious consequences offederal judicial intrusion into this most sensitive of state legislativetasks, "[t]he courts, in assessing the sufficiency of a challenge toa districting plan, must be sensitive to the complex interplay of forcesthat enter a legislature's redistricting calculus." Miller, 515 U.S.at 915-916. See also id. at 916 ("[T]he sensitive nature of redistrictingand the presumption of good faith that must be accorded legislative enactments* * * requires courts to exercise extraordinary caution in adjudicatingclaims that a State has drawn district lines on the basis of race.");id. at 915 ("the good faith of a state legislature must be presumed").
2. The extraordinary sensitivity of the redistricting process, coupledwith the high costs of undue federal court intrusion into that process,demands that a district court scrupulously observe the substantive requirementsof the "predominant factor" test before finding a Shaw violation.In some cases, of course, "[t]he evidentiary inquiry is . . . relativelyeasy." Miller, 515 U.S. at 913. For example, "[i]n some exceptionalcases, a reapportionment plan may be so highly irregular that, on its face,it rationally cannot be understood as anything other than an effort to 'segregat[e]. . . voters' on the basis of race." Shaw I, 509 U.S. at 646-647. Similarly,the redistricting record or the subsequent litigation may disclose the relevantState officials making clear that their "overriding purpose was * ** to create * * * congressional districts with effective black voting majorities."Shaw II, 517 U.S. at 906; Miller, 515 U.S. at 918 (State was "drivenby its overriding desire to comply with [racial] maximization demands").These are mere examples; other facts can also demonstrate that race wasthe predominant factor in a particular case.
In other cases, it cannot so readily be inferred that race was the predominantfactor. For example, when (as is true in this case) race correlates highlywith partisan voting behavior, it is predictable that a State that wantsto create a district whose borders tend to concentrate members of a particularpolitical party will, as a byproduct, create a district whose borders tendto concentrate members of a particular race. If that alone were sufficientto support a finding that strict scrutiny applies (and that the districtis unconstitutional absent a compelling interest), a State would have toforego its otherwise lawful option of forming districts on the basis ofpartisan choices. Indeed, it would have to do so only in one category ofcases-where race correlates highly with partisan voting behavior. That contravenesthe settled principles that "incumbency protection, at least in thelimited form of 'avoiding contests between incumbent[s],' [is] a legitimatestate goal," and that "political gerrymandering" should notbe subjected to strict scrutiny. Bush, 517 U.S. at 964.2
Even if the State has taken race into account to some extent in drawingthe district in such a case, that is still not sufficient to show that the"predominant factor" underlying the district is racial. As discussedabove, strict scrutiny is not triggered where race is merely "a motivation,"Bush, 517 U.S. at 959, in drawing a district; a Shaw claim requires proofthat race was the predominant factor. Therefore, where race and partisanvoting behavior correlate highly, and a State draws a district with mixedpolitical, racial, and other motivations, a district court may not merelyseize on isolated evidence tending to show the State's racial motivationin drawing the district to conclude that race was the predominant factor.To permit that inference would paradoxically hamstring state legislaturesin achieving their political objectives in any State where race correlateshighly with partisan voting behavior. For even if a legislature paid noattention whatever to race, its politically motivated districting decisionswould likely be susceptible to a racial interpretation. And if the Stateexercised its lawful authority to take race into account to some extent,it would inevitably risk the finding of predominant racial motive that wasmade here. That result would be inconsistent with bedrock principles recognizingthat state legislatures-and not federal courts-have primary responsibilityfor the politically highly charged task of drawing districts, and that federalcourts must be particularly cautious before intruding into state prerogativesin this area. To trigger strict scrutiny, the party challenging the districtmust satisfy the heavy burden of proving that "[r]ace was the criterionthat, in the State's view, could not be compromised." Shaw II, 517U.S. at 907.
Bush v. Vera illustrates these principles. The plurality in Bush initiallynoted findings "that the State substantially neglected traditionaldistricting criteria such as compactness, that it was committed from theoutset to creating majority-minority districts, and that it manipulateddistrict lines to exploit unprecedentedly detailed racial data." 517U.S. at 962. The plurality stated, however, merely that those factors "togetherweigh in favor of the application of strict scrutiny"-not that theyrequired its application. Ibid. The plurality explained that it must therefore"consider what role other factors played in order to determine whetherrace predominated." Id. at 963. As the plurality explained, "[b]ecauseit is clear that race was not the only factor that motivated the legislatureto draw irregular district lines, we must scrutinize each challenged districtto determine whether the District Court's conclusion that race predominatedover legitimate districting considerations, including incumbency, can besustained." Id. at 965. Only after concluding that there was exceptionallystrong evidence sufficient to show not merely that race was a factor, butthat it was the predominant factor, did the plurality determine that thedistricts in question should be subject to strict scrutiny.3 The same inquirywas required here.
II. THE DISTRICT COURT IMPROPERLY INFERRED A PREDOMINANT RACIAL MOTIVEIN THIS CASE
As this Court noted in its decision in Hunt I, the lines of District12 correlate highly with race. District 12 contains portions of six counties;in each of them, the portion of the county within District 12 has a substantiallyhigher African-American population than does the portion of the county outsideDistrict 12. See Hunt I, 526 U.S. at 548 & n.4. Moreover, the boundarylines of District 12 are irregular in shape. See id. at 547-548. Plaintiffs'claim has always been that this evidence demonstrated that District 12 wasan unconstitutional racial gerrymander.
As the Court explained in Hunt I, however, the State advanced a differentexplanation for the lines of District 12. At the time of the 1997 Plan,the North Carolina legislature was divided between Republicans and Democrats,with the Republicans in control of the House and the Democrats in controlof the Senate. Similarly, the State's congressional delegation was evenlydivided between six Democrats and six Republicans. The State contended thatthe legislators in charge of redistricting concluded that, in this situation,the only way to get a redistricting plan through the legislature would beto adopt a plan that maintained the six-six partisan split in the congressionaldelegation and that protected all of the incumbents. See J.S. App. 82a-83a,138a-139a; J.A. 180-182, 235, 240-241. Because District 12 had a Democraticincumbent, the result was to craft District 12 in such a way as to solidifythe Democratic vote there. Further, it is undisputed that 90% or more ofAfrican-Americans in North Carolina regularly vote Democratic. See, e.g.,J.A. 130 ("over 90 percent" in a series of studies); J.A. 139("95 to 97 percent"). Accordingly, the State contended that thecorrelation between the district lines and race was a mere by-product ofthe State's desire to create a solidly Democratic District 12; the resultof the State's attempt to concentrate Democratic voters in the districtwas that the most reliable Democratic voters-African-Americans-tended tobe included.4
In short, this was a "mixed motive" case, like Bush v. Vera.See 517 U.S. at 959. On remand, what remained for the district court wasto determine whether plaintiffs could carry their burden at trial of provingthat, as between the two motives, race-and not the kinds of partisan considerationsurged by the State-was the predominant factor underlying the District. Thedistrict court's conclusion that plaintiffs had carried that burden wasfatally defective, for three reasons.
A. The District Court Relied Substantially On Evidence That Was IncompetentTo Distinguish Between Race And Politics As A Factor In Drawing District12
Much of the district court's opinion is an almost verbatim repetitionof the court's previous opinion on summary judgment. Compare J.S. App. 10a-17a(final judgment opinion), 23a-26a (same), 28a-30a (same) with J.S. App.246a-253a (summary judgment opinion), 258a-261a (same), 262a-263a (same).The portions of the majority's opinion repeated from its summary judgmentopinion recite findings that District 12's boundaries correspond with race;that District 12 splits each of the cities and counties it enters on linesthat correspond with race; and that District 12 is unusually shaped understatistical and other measures of compactness.
The facts recited by the district court are accurate, and in an appropriatecase they could provide substantial evidence of a predominant racial motive.In the circumstances of this mixed motive case, however, the evidence recitedabove only frames the question; it does nothing to provide an answer. Itmerely shows that there must have been some motive behind this unusuallyshaped district and that that motive might have been race. But the Stateproduced substantial evidence showing that its predominant motives werepolitical, and that political motives would result in a district with thesame unusual shape and the same racial composition. The evidence that District12's boundaries tend to correspond with race does nothing to distinguishbetween the two motives and to determine which was the predominant one-theprimary issue that remained open for trial after this Court's remand.
Nor is that inquiry advanced by the fact, noted by the district courtin its summary judgment opinion and repeated verbatim after trial, that"the uncontroverted evidence demonstrates * * * the legislators excludedmany heavily-Democratic precincts from District 12, even when those precinctsimmediately border the Twelfth and would have established a far more compactdistrict." J.S. App. 25a; see id. at 261a (summary judgment opinion).It is true that District 12 excludes a number of adjacent precincts withhigh Democratic registration; the district court enumerated those precinctsin its opinion. See id. at 13a-14a; compare J.S. App. 249a-250a (summaryjudgment opinion). But, as this Court noted in Hunt I, the State's evidence"showed that, in North Carolina, party registration and party preferencedo not always correspond." 526 U.S. at 551. Indeed, the undisputedevidence showed that a large number of registered Democrats in North Carolinaregularly vote Republican. See J.A. 397, 780; J.S. App. 173a-174a; 213a-225a.5Accordingly, the State asserted that it used actual election returns byprecinct-not registration figures-to assess the partisan makeup of precinctsand to construct its 1997 plan. The fact that District 12's boundaries sometimesomit precincts that are heavily Democratic by registration does nothingto disprove the State's contention that its predominant motive was to createa solidly Democratic District 12, as measured by actual election returns.6
This Court stated in Hunt I that "[e]vidence that blacks constituteeven a supermajority in one congressional district while amounting to lessthan a plurality in a neighboring district will not, by itself, sufficeto prove that a jurisdiction was motivated by race in drawing its districtlines when the evidence also shows a high correlation between race and partypreference." Hunt I, 526 U.S. at 551-552. At bottom, the evidence repeatedfrom the district court's former opinion did no more than show what thisCourt determined would "not suffice" to prove a racial motivation,much less a predominant racial motivation. Accordingly, the district court'sconclusion in this part of the opinion that "where cities and countiesare split between the Twelfth District and neighboring districts, the splitsinvariably occur along racial, rather than political, lines," J.S.App. 25a, must be rejected as unsupported by the evidence.7
B. It Was Clear Error For The District Court To Infer Predominant RacialMotive From Dr. Weber's Testimony
The district court added a brief additional portion to its prior opinion.See J.S. App. 26a-28a. That portion purports to address further the questionwhether race or partisan considerations was the predominant factor in drawingDistrict 12. Some of the evidence to which the district court refers inthis portion of its opinion is essentially repetitious of the evidence discussedabove, and it is thus no more helpful in distinguishing between racial andpartisan motivations underlying District 12. But the district court alsorelied on a number of portions of the testimony of Dr. Ronald Weber, appellees'expert, which the district court stated showed "time and again howrace trumped party affiliation in the construction of the 12th Districtand how political explanations utterly failed to explain the compositionof the district." J.S. App. 26a. That conclusion, however, was plainlywrong.
Initially, as discussed above, "party affiliation"-as opposedto actual partisan voting conduct-is of little relevance in this case andof no use in the analysis. See pp. 19-20, infra. It was therefore errorto rely on portions of Dr. Weber's testimony that were based on registrationdata. Beyond that, however, the evidence presented by Dr. Weber on whichthe district court relied was not significantly probative of race as thepredominant factor in drawing District 12. Accordingly, the court committedclear error in relying on that evidence.
1. The district court cited a portion of Dr. Weber's testimony in whichhe referred to the fact that District 12 has more Democratic voters thanadjoining Democratic District 8. He stated that the State, had it been followingits partisan objectives, would have "want[ed] to take some of the votersin the district that you are drawing that's overly safe and put them into[an] adjacent district so as to make that district more competitive."Tr. 162 (J.A. 91).8
The State, however, explained the reason for this configuration. District12, in general, is no more solidly partisan than are at least two RepublicanDistricts-Districts 6 and 10. See J.S. App. 80a (election results). Theproportion of Democrats in District 12 is therefore not suspect. And withrespect to the specific line dividing Districts 12 and 8, the State explainedthat that line runs along the border between Cabarrus County (in District8) and Mecklenburg County (in Districts 9 and 12). See J.A. 501 (map). Toput some District 12 Democrats into District 8, the State would have hadto violate two political constraints that were important to the legislature:it would have had to move some of Mecklenburg County into District 8, whichwould have divided the county into three districts and thus violated theState's consistent policy in the 1997 Plan of placing no county in morethan two districts, see J.A. 179, 474-475, 780-782; see also J.A. 658; andit would likely have required moving some of Cabarrus County out of District8 to District 12 in return, thus violating the desire of Democratic incumbentHefner in District 8, who lived in Cabarrus County, to represent his entirehome county, see J.S. App. 85a; J.A. 205-206.
The district court did not discuss the State's proffered explanationor otherwise explain why it might be deficient.9 Dr. Weber admitted thathe did not take into account any of the political considerations advancedby the State. See J.A. 135 ("I don't know anything about what CongressmanHefner asked."), 136 (answering "No" to question whetherhe "inquired about any real world political issues that might havebeen going on that might have determined why the Legislature drew the linewhere it did"). Without some reason to discredit the State's explanation,Dr. Weber's analysis does not provide significant evidence of discrimination.Accordingly, the district court's inference of predominant racial motivefrom Dr. Weber's evidence was "illogical" and, hence, clearlyerroneous. See Anderson v. City of Bessemer City, 470 U.S. 564, 577 (1985).
2. The district court also relied on Dr. Weber's testimony that District12 contains virtually all (76 out of 79) precincts that are 40% or moreAfrican-American in the six counties that comprise the district, but itdoes not contain as high a percentage of precincts with Democratic tendencies,even as measured by election results. Tr. 204-205 (J.A. 105-106). The districtcourt clearly erred in inferring a racial motive- much less a predominantracial motive-from that testimony. The question is not whether there wereother precincts in the six counties with Democratic voting patterns thatwere left out of District 12; the question is whether, if there are suchprecincts, including them in District 12 would have raised or lowered theoverall likely Democratic vote in District 12. If the omitted Democraticprecincts are far from the borders of District 12, including them wouldfrequently not have been practical, and, even if it would, expanding thedistrict to include them could easily have required including or excludingother precincts that would have resulted in an overall boost in Republicanstrength in District 12.10 Dr. Weber, however, did not attempt to show thatthe omitted precincts could have reasonably been included in District 12or that their inclusion would have in fact raised Democratic strength inthe district. Cf. J.S. App. 50a n.21 (Thornburg, J., dissenting) (State'sevidence showed that "few of the strong Democratic precincts to whichDr. Weber referred could have easily been included in the Twelfth District").Without such evidence, Dr. Weber's testimony on this point proves nothing.11
3. The district court also relied on page 221 of Dr. Weber's testimony(J.A. 111) in which he argued that splitting a single precinct in MecklenburgCounty (Precinct 77, the only split precinct in District 12, see J.S. App.84a) showed that race was the predominant motive. The State explained thatthe purpose of splitting that precinct, located at the southernmost tipof Mecklenburg County, was to connect the two portions of Republican RepresentativeMyrick's district without including additional Democratic voters in herdistrict. See J.S. App. 208a; J.A. 20, 617-618. That in turn was in serviceof the overall goal of protecting incumbents and therefore splitting MecklenburgCounty between the two incumbents who lived there-the Democratic incumbentin District 12 and the Republican incumbent in District 9. See J.A. 597-598.Neither the court nor Dr. Weber addressed that explanation. Although evidenceof a single split precinct is unlikely to be significantly probative inany event, the failure by Dr. Weber or the court to explain why the State'sexplanation was deficient undermines the court's reliance on this testimonyto infer predominant motive. See n. 9, supra.
4. Taken individually or together, none of the portions of Dr. Weber'stestimony on which the district court relied were significantly probativeeven of race as a factor in drawing District 12. Moreover, even if it wereotherwise and Dr. Weber's testimony on these points were significantly probativethat race was a factor in drawing District 12, neither a slight increasein the percentage of Democrats in District 12, a failure to include someisolated Democratic precincts, nor the splitting of a single precinct wouldsuffice to show that race was the predominant factor. The district courtcommitted clear error in finding Dr. Weber's testimony sufficient to supportan inference that the State's predominant motive in drawing District 12was race.
C. The District Court's Conclusions From Other Testimony Were InfectedBy Its Earlier Errors And In Any Event Confuse Evidence That Race Was AFactor In Drawing District 12 With Evidence That It Was The PredominantFactor
1. The district court stated that "[t]he conclusion that race predominatedwas further bolstered by Senator Cooper's allusion to a need for 'racialand partisan balance'" in a statement made to the state House Committeeon Congressional Redistricting. J.S. App. 27a. At trial, Senator Coopertestified that by "partisan balance," he meant "[k]eepingthe 6-6 split," and by "racial balance," he meant "thatAfrican Americans would have a fair shot to win both the First and 12thDistricts, and I think that's racially fair." J.A. 222. The districtcourt stated, however, that "[t]he Senator's contention that althoughhe used the term 'partisan balance' to refer to the maintenance of a six-sixDemocrat-Republican split in the congressional delegation, he did not meanthe term 'racial balance' to refer to the maintenance of a ten-two balancebetween whites and African-Americans is simply not credible." J.S.App. 27a.
When the district court made that credibility finding regarding SenatorCooper's testimony, it had already made the errors recounted above in determiningthat the statistical and demographic evidence in the case supported an inferenceof race as the predominant motive. The district court was no doubt influencedby those erroneous conclusions in determining that Senator Cooper's contrarytestimony was not credible. Moreover, the district court's inference thatbecause "partisan balance" meant a six-six split, "racialbalance" must have also meant a fixed numerical split, is belied bythe fact that Senator Cooper's original testimony did not merely refer to"partisan and racial balance," see J.S. App. 27a, but to "geographic,racial and partisan balance," J.A. 460 (emphasis added). Because theterm "geographic balance" does not suggest the kind of divisioninto neat numerical categories that the term "partisan balance"does, it is apparent that Senator Cooper did not consistently mean by "balance"a fixed numerical division of the districts, as the district court apparentlybelieved.
For the above reasons, the district court's credibility finding regardingSenator Cooper is unsupported. Even if the district court's finding wereaccepted, however, it would show at most that race was a motivation in SenatorCooper's attempt to configure District 12. He had already testified, however,that "we did pay attention to race," and that "[t]hat wasone of the factors that was considered," but that "it was certainlynot the predomina[nt] factor." J.A. 222. The question in the case thuswas never whether race was considered, but whether race was the predominantfactor. Neither Senator Cooper's statement that he was seeking "geographic,racial and partisan balance," nor his asserted failure to explain whathe meant by "racial balance" suggests that racial balance wasthe predominant motive underlying the creation of District 12-that "[r]acewas the criterion that, in the State's view, could not be compromised."Shaw II, 517 U.S. at 907.
2. Finally, the district court relied on the Cooper-Cohen e-mail, inwhich Gerry Cohen, the legislative employee responsible for actually drawingthe 1997 Plan on the computer, had said "I [Cohen] have moved GreensboroBlack community into the 12th, and now need to take [a]bout 60,000 out ofthe 12th. I await your direction on this." J.S. App. 8a; see J.A. 369(full text of e-mail). Cohen's e-mail on its face merely identified thegeneral characteristics of the community that had been moved into District12 by referring to its racial composition-which, as this Court has noted,"the legislature always is aware of * * * when it draws district lines,just as it is aware of age, economic status, religious and political persuasion,and a variety of other demographic factors." Shaw I, 509 U.S. at 646;see also Bush, 517 U.S. at 958 ("Strict scrutiny does not apply merelybecause redistricting is performed with consciousness of race."). Accordingly,the question presented by the e-mail is whether the district court properlyinferred from that awareness that "the chief architects of the 1997Plan had evolved a methodology for segregating voters by race, and thatthey had applied this method to the 12th District." J.S. App. 27a.
As with the Cooper statement, the district court made its inference withrespect to the e-mail only after having made its erroneous findings thatthe statistical and demographic evidence demonstrated a predominant racialmotive. Had the district court not made the earlier errors, it might haveseen the e-mail in a different light, and it might not have drawn the dramaticconclusion from the e-mail that it did. Indeed, the State had explainedthat the reason for moving the community into District 12 was in part toavoid splitting Guilford County into three districts-a goal that, as notedabove, see p. 23, supra, the State followed consistently with respect toevery county in the State in the 1997 Plan-and in part to bolster the Democraticvote in District 12 (a goal desired by the Democratic state Senate and CongressmanWatt, the incumbent there) and to subtract Democrats from the vote in neighboringDistrict 6 (a goal desired by Republican Congressman Coble, the incumbentthere). See J.A. 192, 193, 195-196, 216, 264-265, 268. The district courtdid not specifically address or assess the State's evidence that these werethe primary motivations for moving the portion of Greensboro into the TwelfthDistrict. See n. 9, supra. Without an explanation of the district court'sreasons for rejecting the State's proffered explanation, the district court'sconclusion from the e-mail is insupportable.
Finally, even if the e-mail were viewed as persuasive evidence that racewas a factor in moving that portion of Greensboro into District 12, it wouldnot provide sufficient evidence to infer that race was the predominant factorin constructing District 12 as a whole. In this respect, again, Bush isinstructive. In that case, the plurality noted evidence that "the decisionto create the districts now challenged as majority-minority districts wasmade at the outset of the process and never seriously questioned,"517 U.S. at 961, and that those drawing the challenged districts made useof "uniquely detailed racial data," id. at 961-962. Nonetheless,the plurality viewed that evidence merely as setting forth the questionwhether race or politics predominated in drawing the challenged districts,not as providing an answer for that question. Similarly here, even scatteredevidence that race was a factor taken into account in determining one oranother particular feature of District 12 is insufficient to show that racewas the predominant motive underlying District 12 as a whole.
3. As is apparent from a review of the district court's opinion, thecourt erred in concluding that race was the predominant motive in the creationof District 12. To a significant extent, the court relied on evidence thatcould not resolve the central question before the court: whether race orpolitics predominated in the construction of District 12. Even insofar asthe district court, however, relied on evidence that had to do with racialconsiderations, the evidence showed at most that race was taken into accountin creating District 12-a fact that the State conceded from the beginning.Because the district court failed correctly to appreciate and apply thedifference between race as a factor and race as the predominant factor,the district court's conclusion that District 12 is an unconstitutionalracial gerrymander cannot stand. To permit a district court to find a predominantracial motive in a case like this would put state legislatures that haveacted entirely constitutionally at risk that a district court, finding thatrace was a factor in one or another feature of a districting plan, coulddeclare the entire plan unconstitutional. That would threaten to immersethe district courts deeply in the highly political thicket of redistricting,and it cannot be squared with the kind of sensitivity toward state legislativeefforts in this field that this Court has always required.
The judgment of the district court should be reversed.
SETH P. WAXMAN
WILLIAM R. YEOMANS
Acting Assistant Attorney
BARBARA D. UNDERWOOD
Deputy Solicitor General
JAMES A. FELDMAN
Assistant to the Solicitor
DAVID K. FLYNN
LOUIS E. PERAERTZ
1 The district court also held that District 1 was subject to strictscrutiny, but it found that the State had satisfied that standard. J.S.App. 30a-35a. Appellees did not perfect their appeal from that ruling, andthe district court granted appellants' motion to dismiss appellees' appealon August 3, 2000 (Docket No. 178). Accordingly, although the district court'sruling that District 12 is unconstitutional is now before this Court, thedistrict court's ruling that District 1 is constitutional is no longer atissue, and will not be further addressed herein.
2 See Bush, 517 U.S. at 968 ("If the State's goal is otherwise constitutionalpolitical gerrymandering, it is free to use * * * political data * * * -precinctgeneral election voting patterns, precinct primary voting patterns, andlegislators' experience-to achieve that goal regardless of its awarenessof its racial implications and regardless of the fact that it does so inthe context of a majority-minority district.") (citations omitted).
3 See Bush, 517 U.S. at 969 (evidence that the State itself explainedthe district "in exclusively racial terms"), 970 (evidence that"districting software * * * provided only racial data at the block-by-blocklevel," and that district lines were in fact determined at that level),970 (evidence of use of "race as a proxy"), 971 (evidence thatshape of district was "far from the shape that would be necessary tomaximize the Democratic vote in that area"), 972-973 ("intensiveand pervasive use of race both as a proxy to protect the political fortunesof adjacent incumbents, and for its own sake in maximizing the minoritypopulation of District 30 regardless of traditional districting principles"),975 ("racial demographics and voting patterns * * * belie any suggestionthat party politics could explain" two adjoining districts, because"[t]he district lines correlate almost perfectly with race, while bothdistricts are similarly solidly Democratic") (citation omitted).
4 It is also significant that District 12-with a 43% African-Americanvoting age population and a 47% total African-American population-is nota majority-minority district. As the Court explained in Lawyer, "[t]hefact that [the challenged district] is not a majority black district * ** supports the * * * finding that the district is not a 'safe' one for black-preferredcandidates, but one that offers to any candidate, without regard to race,the opportunity to seek and be elected to office." 521 U.S. at 581(internal quotation marks omitted).
5 For example, in 1996, 54% of the State's voters were registered asDemocratic, while only 34% were Republicans. The Almanac of American Politics1998, at 1056 (1997). Yet the Republican candidates won in the 1992 and1996 presidential elections, the State's two Senators at the time of theredistricting were both Republicans (although a Democrat defeated one ofthem in the 1998 election), and the State's delegation to the 105th Congressconsisted of six Republicans and six Democrats (although one of the Democraticseats was won by a Republican in the 1998 election). Id. at 1057.
6 There is an additional defect in the district court's inference, becausethe district court disregarded "the necessity of determining whetherrace predominated in the redistricters' actions in light of what they hadto work with." Bush, 517 U.S. at 972 n.*. The fact that District 12excludes even some adjacent precincts with Democratic voting patterns wouldbe of little significance, unless it could be shown as well that includingthose precincts would make the District as a whole more Democratic. Where,for instance, the district lines tend to exclude precincts with Democratictendencies while including precincts with more pronounced Democratic tendencies,the exclusion of the former casts no doubt on the State's claim that itwas attempting to draw as highly a Democratic district as possible. Theevidence in fact showed that this was precisely what happened here, whethermeasured by party registration or actual election returns. See Jt. Exhs.107-109; see also J.A. 140 (testimony by Dr. Weber agreeing that excludedwhite precincts are not "as heavily Democratic" as the precinctswithin District 12.).
7 The State supported its conclusion that the district was drawn alongpolitical lines by showing that Republican victories were common in precinctsabutting District 12, see J.S. App. 213a- 225a, and that the splits in countiesand municipalities divided Democratic portions in District 12 from Republicanportions outside District 12, see id. at 189a, 191a-192a. The district courtdid not address that evidence.
8 The district court referred to another portion of Dr. Weber's testimony,in which he made essentially the same point, when it stated that "Dr.Weber showed that, without fail, Democratic districts adjacent to District12 yielded their minority areas to that district, retaining white Democraticprecincts." J.S. App. 26a (citing Tr. 255-256 (J.A. 134-135)). Thedistrict court's misapprehension of the record is apparent from its referencesto "Democratic districts adjacent to District 12," since it isundisputed that, of the five districts adjacent to District 12, only one(District 8) had a Democratic incumbent in 1997. Moreover, the districtcourt did not specify any majority-minority precincts that had been in District8 in a prior plan and subsequently were "yielded" to District12, and we are unable to identify any. As the map of District 12 and itssurroundings reveals, see J.A. 483, there are no majority-minority precinctsnear the border between Districts 8 and 12.
9 Under Rule 52(a) of the Federal Rules of Civil Procedure, a districtcourt "shall find the facts specially and state separately its conclusionsof law thereon." As this Court has stated, "there comes a pointwhere findings become so sparse and conclusory as to give no revelationof what the District Court's concept of the determining facts and legalstandard may be." Commissioner v. Duberstein, 363 U.S. 278, 292 (1960).The courts of appeals, led by the Fifth Circuit, have required that districtcourts exercise special care under Rule 52(a) in the redistricting context,and the district court's failure to exercise such care is itself groundsfor reversal. As the Fifth Circuit has explained, "[b]ecause the resolutionof a voting dilution claim requires close analysis of unusually complexfactual patterns, and because the decision of such a case has the potentialfor serious interference with state functions," district courts must"strictly adhere to the [Rule] 52(a) requirements" that they"find the facts specially" and must "explain with particularitytheir reasoning and the subsidiary factual conclusions underlying theirreasoning." Westwego Citizens for Better Gov't v. City of Westwego,872 F.2d 1201, 1203 (1989) (quoting Velasquez v. City of Abilene, 725 F.2d1017, 1020 (5th Cir. 1984)). Other courts of appeals similarly "requirea particularly definite record for voting rights cases." Cousin v.McWherter, 46 F.3d 568, 574 (6th Cir. 1995); accord Johnson v. Hamrick,196 F.3d 1216, 1223 (11th Cir. 1999); Lee County Branch of the NAACP v.City of Opelika, 748 F.2d 1473, 1480 (11th Cir. 1984); Harvell v. Ladd,958 F.2d 226, 229 (8th Cir. 1992); Buckanaga v. Sisseton Indep. Sch. Dist.,804 F.2d 469, 472 (8th Cir. 1986). The "bedrock rule" that a districtcourt's findings must be "sufficiently detailed to permit a reviewingcourt to ascertain the factual core of, and the legal foundation for, therulings below * * * has particular force in cases of this genre." Unov. City of Holyoke, 72 F.3d 973, 988 (1st Cir. 1995). That includes therequirement that "the district court must discuss 'not only the evidencethat supports its decision but also all the substantial evidence contraryto its opinion.'" Ibid.; see also Velasquez, 725 F.2d at 1021 (remandingfor district court, which wrote a "long and detailed" opinion,to "take note of substantial contrary evidence presented by the appellants").It also includes the requirement that "when the statistics are theprincipal evidence offered * * *, the district court must ensure that itthoroughly discusses its reasons for rejecting that evidence." Clarkv. Calhoun County, 21 F.3d 92, 96 (5th Cir. 1994).
10 Insofar as Dr. Weber referred to precincts with Democratic votingpatterns adjoining District 12, the evidence showed that those precinctswere uniformly less Democratic than the precincts included in the district.See p. 20 n.6, supra.
11 The district court also referred to pages 262 (J.A. 139-140) and 288(J.A. 156-157) of the transcript. In those portions of his testimony, Dr.Weber was being cross-examined regarding his claim that Democratic precinctswere left out of District 12. His testimony on cross-examination adds nothingto the analysis. At page 251 of the transcript (J.A. 131), Dr. Weber simplystates the conclusion that "[r]ace is the predomina[n]t factor."That too adds nothing to the analysis.