REAVLEY, Circuit Judge:
This is a voting rights case. We uphold the district court’s finding, as not clearly erroneous, that the at-large election of City of Baytown council members constitutes a violation of § 2 of the Voting Rights Act of 1965, 42 U.S.C. § 1973 (Supp.1987) (as amended June 29, 1982), because of vote dilution of the politically cohesive combination of Blacks and Mexican-Americans. However, the 5-3-1 plan, proposed by the city and adopted by the district court, has not been precleared.
See
42 U.S.C. § 1973c. We, therefore, vacate the district court’s approval of the plan and remand for compliance with § 1973c.
I.
The City of Baytown, located in Harris County, Texas, had a 1980 census population of 56,917. Of that, 9348 (16.42%) were Hispanic and 5096 (8.95%) were Black. The combined minority population was 14,444 (25.4%).
Baytown has an at-large election system for its six city council members and its mayor. Although elected at-large, each council member has to reside in a particu
lar district resulting in a numbered-post system.
Additionally, there is a majority vote requirement.
The system has remained unchanged since Baytown was formed in its present shape in 1947. No minority member, either Black or Hispanic, has ever been elected to the Baytown City Council.
A number of Hispanic and Black citizens of Baytown brought this suit, individually and as a class action, alleging that the at-large system was a violation of § 2 of the Voting Rights Act. 42 U.S.C. § 1973. Combining Blacks and Hispanics as one minority group, the district court found that the minority group was sufficiently large and geographically insular to form a majority in a single member district. Focusing exclusively on elections with minority candidates, the court found that Blacks were cohesive, Hispanics were cohesive, together the minority group was cohesive, and that Anglos voted sufficiently as a bloc to usually defeat the minority’s preferred candidate. To review the totality of the circumstances, the court considered the factors enumerated in
Zimmer v. McKeithen,
485 F.2d 1297 (5th Cir.1973) (en banc),
aff'd sub. nom. on other grounds, East Carrol Parish School Board v. Marshall,
424 U.S. 636, 96 S.Ct. 1083, 47 L.Ed.2d 296 (1976) (per curiam), and concluded that a § 2 voter dilution violation was established.
The court then ordered the city to submit an alternate plan. Baytown responded with a 5-3-1 plan that had five council members elected from districts including one minority district, three council members elected at-large, and one mayor elected at-large. Paying heed to legislative deference, the district court adopted the city’s plan.
II.
A.
In
Thornburg v. Gingles,
478 U.S. 30, 106 S.Ct. 2752, 92 L.Ed.2d 25 (1986), the Supreme Court, for the first time, considered the 1982 amendments to § 2 of the Voting Rights Act, 42 U.S.C. § 1973,
in the context of a challenge to an at-large election system. The Court noted that the purpose for the congressional amendment to § 2 was to eliminate the requirement of showing discriminatory intent in a challenge to a contested electoral mechanism.
Gingles,
106 S.Ct. at 2759;
see City of Mobile, Ala. v. Bolden,
446 U.S. 55, 100 S.Ct. 1490, 64 L.Ed.2d 47 (1980). Instead, the “results test” of
White v. Regester,
412 U.S. 755, 93 S.Ct. 2332, 37 L.Ed.2d 314 (1973), was re-established. Specifically, the Court found that the Senate Judiciary Report focused on the
Zimmer
factors in showing a § 2 voter dilution claim.
Gingles,
106 S.Ct. at 2759-60.
Although
many or all of the factors are relevant in a challenge to an at-large system, the Court concluded that “the use of multimember districts generally will not impede the ability of minority voters to elect representatives of their choice,” unless “a bloc voting majority must
usually
be able to defeat candidates supported by a politically cohesive, geographically insular minority group.”
Gingles,
106 S.Ct. at 2766 (emphasis in original).
That test breaks down into three parts: 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. ... Second, the minority group must be able to show that it is politically cohesive. ... Third, the minority must be able to demonstrate that the white majority votes sufficiently as a bloc to enable it — in the absence of special circumstances, such as the minority candidate running unopposed, — usually to defeat the minority’s preferred candidate.
Gingles,
106 S.Ct. at 2766-67 (cross citation omitted). The second and third parts, cohesion and majority bloc voting, are usually proven by statistical evidence of racially polarized voting.
Gingles,
106 S.Ct. at 2768-69. As the Court explained:
the question whether a given district experiences legally significant racially polarized voting requires discrete inquiries into minority and white voting practices. A showing that a significant number of minority group members usually vote for the same candidates is one way of proving the political cohesiveness necessary to a vote dilution claim, and, consequently, establishes minority bloc voting within the context of § 2. 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.
Gingles,
106 S.Ct. at 2769-70 (citation omitted).
The ultimate finding of minority voter dilution is then to be based on the totality of the circumstances.
Gingles,
106 S.Ct. at 2782. That finding, and its subsidiary findings, are subject to the clearly erroneous standard of appellate review for fact finding.
Gingles,
106 S.Ct. at 2781; Fed.R. Civ.P. 52(a). Under that standard, “a finding is ‘clearly erroneous’ when although there is evidence to support it, the reviewing court on the entire evidence is left with the definite and firm conviction that a mistake has been committed.”
Anderson v. City of Bessemer City, N.C.,
470 U.S. 564, 573, 105 S.Ct. 1504, 1511, 84 L.Ed.2d 518 (1985) (quoting
United States v. United States Gypsum Co.,
333 U.S. 364, 395, 68 S.Ct. 525, 542, 92 L.Ed. 746 (1948)). If the factual determinations are based on determinations of the witnesses’ credibility, or on reasonable interpretation or inferences from the testimony or other evidence, the district court’s findings cannot be clearly
erroneous.
Anderson,
470 U.S. at 574-75, 105 S.Ct. at 1511-12.
B.
There is nothing in the law that prevents the plaintiffs from identifying the protected aggrieved minority to include both Blacks and Hispanics. Section 1973(a) protects the right to vote of both racial and language minorities.
See
42 U.S.C. §§ 1973(a), 1973b(f)(2). Congress itself recognized “that voting discrimination against citizens of language minorities is pervasive and national in scope,” 42 U.S.C. § 1973b(f)(l), and similar discrimination against Blacks is well documented. If, together, they are of such numbers residing geographically so as to constitute a majority in a single member district, they cross the
Gingles
threshold as potentially disadvantaged voters. To prove the fact of their electoral dilution, plaintiffs must prove that the minorities so identified actually vote together and are impeded in their ability to elect their own candidates by all of the circumstances, including especially the bloc voting of a white majority that usually defeats the candidate of the minority.
III.
Gingles
requires that the minority group be “sufficiently large and geographically compact to constitute a majority in a single member district.”
Gingles,
106 S.Ct. at 2766. The district court found that the minority group was concentrated in the southern sections of Baytown. Evidence was introduced that contiguous single member districts with 72.3% minority population (50% Hispanic, 22.3% Black), 75.4% minority population (50.7% Hispanic, 24.7% Black), and 76.9% minority population (49.5% Hispanic, 27.4% Black)
were possible. Additionally, the City’s plan, adopted by the district court, has a single member district (out of five) that has a 65.9% minority population (44.6% Hispanic, 21.3% Black). The district court was not clearly erroneous in deciding that this
Gingles
requirement was satisfied.
Baytown argues that since there are Blacks and Hispanics spread throughout the city and since 51% of Baytown’s Blacks and 44% of its Hispanics do not live in the minority district, the plaintiffs have failed to show that it is compact and insular enough to meet the
Gingles
requirement. We disagree. The fact that there are members of the minority group outside the minority district is immaterial. All that is required is that the minority group be
“sufficiently
large and geographically compact to constitute a majority in a single member district.”
Gingles,
106 S.Ct. at 2766 (emphasis added). As the Court explained, the purpose behind the requirement is that “[ujnless 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,
106 S.Ct. at 2766-67 n. 17 (emphasis in original). The plaintiffs demonstrated that potential here. Just because not all of the minorities in Baytown are in the district does not mean that
Gingles’
first part is not satisfied.
The next important step under
Gingles
is to determine whether the minority group is politically cohesive. That is not an inquiry to be made prior to and apart from a study of polarized voting, as Baytown argues, because the central focus is upon voting patterns. The Supreme Court has made clear that “[t]he 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 bloc usually to defeat the minority's preferred candidates.”
Gingles,
106 S.Ct. at 2769. It follows that a minority group is politically cohesive if it votes together.
One question that arises, in the admission of evidence or the court's evaluation of the evidence, is identifying the elections to
be examined. The district court here examined only those elections that had a minority member as a candidate. Baytown argues that the court should have examined all elections, including Anglo-Anglo races.
The Supreme Court stated that “[t]he number of elections that must be studied in order to determine whether voting is polarized will vary according to pertinent circumstances. One important circumstance is the number of elections in which the minority group has sponsored candidates.”
Gingles,
106 S.Ct. at 2770 n. 25. Baytown responds that the Court also stated that it is “the
status
of the candidate as the
chosen representative of a particular racial group,
not the race of the candidate, that is important.”
Gingles,
106 S.Ct. at 2776 (emphasis in original). Therefore, Baytown concludes, any time a candidate gets a majority of the minority votes he is the “chosen representative" of the minority group.
We disagree. The district court was warranted in its focus on those races that had a minority member as a candidate.
Cf. Citizens For a Better Gretna v. City of Gretna, La.,
834 F.2d 496, 503 (5th Cir.1987) (Black candidate was sponsored by minority group “because he received a significant portion of black vote and because he is black."). As
Citizens For a Better Gretna,
834 F.2d at 503, points out, Justice Brennan’s statement that the race of the candidate is unimportant carried only four members of the Court in
Gingles.
Justice White specifically disavowed it and the other four Justices agreed with Justice White.
See Gingles,
106 S.Ct. at 2783-84 (White, J., concurring) & 106 S.Ct. at 2793 (O’Connor, J., concurring in the judgment). The focus in
Gingles
was on a candidate
sponsored
by the minority group. 106 S.Ct. at 2770 n. 25. Indeed,
Gingles
itself looked only to elections where Black candidates were running. The
Gretna
panel stated:
“Gingles
is properly interpreted to hold that the race of the candidate is in general of less significance than the race of the voter — but only within the context of an election that offers voters the choice of supporting a viable minority candidate.”
Citizens For a Better Gretna,
834 F.2d at 503. There was no evidence that any Anglo-Anglo race for Baytown City Council offered the voters the choice of a “viable minority candidate.” Indeed, the evidence showed, as the district court found, that minority voter turnout increased dramatically when there was a candidate who was a member of the minority group. Contrary to Baytown’s assertions, the district court properly focused only on those races that had a minority member as a candidate.
Baytown argues that, in order to show cohesion when there are two minorities that make up the minority group, plaintiffs must show that Blacks are cohesive, that Hispanics are cohesive and that Blacks and Hispanics are together cohesive. We think that that burden is too great, if not impossible, in certain situations.
Instead, the proper standard is the same as
Gingles:
whether the minority group together votes in a cohesive manner for the minority candidate. That can be done by examining elections when there was a minority candidate. The key is the minority group as a whole. Of course, if one part of the group cannot be expected to vote with the other part, the combination is not cohesive. If the evidence were to show that the Blacks vote against a Hispanic candidate, or vice versa, then the minority group could not be said to be cohesive. But if the statistical evidence is that Blacks and Hispanics together vote for the Black or Hispanic candidate,
then cohesion is shown.
The evidence presented here by the plaintiffs’ expert, Dr. Brishetto, embraced five elections. Two of those elections pitted a Hispanic candidate versus an Anglo candi
date in the contest for the Baytown City Council District 1 seat. Another election pitted a Black against an Anglo in the 1979 race for City Council District 3. The other two elections were the 1986 general election and runoff for the Goose Creek Independent School District
between a Black and an Anglo. The results of the elections were examined by Dr. Brishetto’s use of ecological correlation and regression analysis.
The “r” and “r2” values for the Black candidates all show an excellent correlation. Baytown does not contest that both Blacks and Hispanics vote as a bloc for the Black candidate. Baytown, however, does challenge the district court's conclusion that the minority group is cohesive. Specifically, Baytown claims that Blacks do not vote for the Hispanic candidate. They rely on their expert’s extreme case analysis
to show that Blacks do not vote for the His
panic candidate. Additionally, Baytown points out that the “r” values are higher for the Campos and Delgado races when only Hispanics are plotted as the minority group as opposed to when Blacks and Hispanics are the minority group.
The district court made explicit findings on the statistical evidence of the voting pattern of the combined minority in the Delgado and Campos races. For the Delgado race, the court found that the “r” value of .69, slope of .46 and significance of .0031 was “quite a high correlation with good statistical significance.” For the Campos race, the court found that the “r” value of .52, slope of .33 and significance of .0456 were the lowest values introduced but, as testified by Brishetto, they were still statistically significant.
The court explicitly rejected the use of Precinct 248. That is the heart of Bay-town’s complaint. Precinct 248 is the precinct that Baytown used as its extreme case analysis for Blacks in Baytown. Additionally, inclusion of Precinct 248 on the statistical graphs for the Campos and Delgado contributes significantly to the smaller “r” values and slopes for the two races.
In essence, Baytown’s whole argument that Blacks are not cohesive with Hispanics comes down to a challenge that the district court was clearly erroneous in rejecting the use of Precinct 248. If Precinct 248 is an aberration, and could properly be ignored, then the statistical evidence showed that Blacks and Hispanics as one minority were politically cohesive.
There are two things about Precinct 248 on which everyone agrees: it is extremely small, both in area and population, and it is overwhelmingly Black. Its total population is 697 people of whom 658 or 94.4% are Black. In the 1986 race between the Hispanic Delgado and the Anglo Simmons, Simmons received 51 votes and Delgado received 7 votes. In other words, the Anglo received 88% of Precinct 248’s votes. In the 1984 race between the Hispanic Campos and the Anglo Simmons, Simmons received 75 votes and Campos received 13. That means that the Anglo received approximately 85% of Precinct 248’s votes over the Hispanic candidate. That, Bay-town concludes, shows that the Blacks in Baytown vote for the Anglo candidate over the Hispanic candidate.
Although Precinct 248 is overwhelmingly Black, it contains less than 13% of the Black population. There have never been more than 90 voters in local elections in Precinct 248. If Precinct 248 is not representative of the Black population in Baytown, it cannot serve as an extreme case analysis. For a number of reasons, we conclude that the district court was not clearly erroneous in rejecting the use of Precinct 248 as a measure of the Black vote. First, there was testimony from Tony Campos that Precinct 248 was controlled by one man. That man was the first Black principal at a local high school, who worked for the county at the time of these elections in 1984 and 1986, and felt that support of Simmons was important to retaining his job. The one-sidedness of both the 1984 and 1986 election results in Precinct 248 indicates that neighborhood politics were at work. Baytown argues that a local Precinct 248 businessman and the pastor of the local church both testified that Precinct 248 was representative of the Black community in Baytown. The district court rejected both witnesses as not being credible. Both, the court found, were familiar only with Precinct 248; the businessman could not understand a number of
questions and the pastor did not live in Baytown. When the findings of a district court are based on the credibility of the witnesses, that finding can almost never be clearly erroneous.
Anderson,
470 U.S. at 575, 105 S.Ct. at 1512. The district court could properly conclude that Precinct 248 was an aberration based on the witnesses’ testimony.
A close examination of the evidence also reveals a second reason for rejecting the use of Precinct 248 as indicative of Black voting preference in Baytown. If Precinct 248 is to be used as an extreme case analysis, it should be predictive of other Black votes. Precinct 102 contains approximately 64% more Blacks than Precinct 248. Precinct 102 contains the largest absolute number of Blacks of any precinct in Bay-town.
Assuming that Blacks vote in approximately equal percentage in both precincts, the results in Precinct 248 should, if that is a valid extreme case analysis, predict the approximate number of Black votes for a candidate in Precinct 102. In the 1986 Delgado-Simmons race, Simmons received 51 votes to Delgado's 7 votes in Precinct 248. Since Precinct 102 has 64% more Blacks, Simmons should have received approximately 84 votes from Blacks in Precinct 102. Actually, in Precinct 102 Simmons received only 35 votes and Delgado received 243. Those 35 votes are far short of the predicted 84 votes that Simmons should have received from Blacks if Precinct 248 were representative. Instead, given that Precinct 102 is also 11% Anglo, the opposite conclusion is appropriate: Blacks in Precinct 102, the largest number of Blacks in any precinct, voted overwhelmingly for Delgado. The 1984 Simmons-Campos race also breaks down the same way. Precinct 248 voted 75-13 for Simmons. That should indicate that approximately 123 Blacks in Precinct 102 should have voted for Simmons. But, in fact, only 58 votes were cast in that race for Simmons with Campos receiving 167 votes in Precinct 102.
We are not “left with the definite and firm conviction,”
Anderson,
470 U.S. at 573, 105 S.Ct. at 1511, that the district court made a mistake in rejecting the use of Precinct 248. The court could properly conclude that Precinct 248 was an aberration based both on the testimonial and statistical evidence. The plaintiffs showed that Blacks and Hispanics, as one minority, were cohesive. Using extreme case analysis with Precinct 248, Baytown attempted to show that Blacks were not cohesive with Hispanics since Blacks did not vote for the Hispanic candidate. If that evidence had been unrebutted, then cohesion would be doubtful. However, the plaintiffs presented probative evidence to rebut Baytown’s showing.
Baytown’s other arguments on cohesion also fail. The district court could reject Baytown’s expert’s testimony which relied primarily on Precinct 248 as an extreme case analysis as well as a large number of Anglo-Anglo races to indicate minority preference. The standing evidence showed that Blacks and Hispanics, as one minority, were politically cohesive. The district court’s finding on the second part of the
Gingles
test was not clearly erroneous.
C.
Statistical evidence of racially polarized voting also bears on the third part of the
Gingles
test: “whether whites vote sufficiently as a bloc usually to defeat the minority’s preferred candidates.” 106 S.Ct. at 2769. As explained by the court, “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.”
Id.
at 2770.
Here, the evidence of racially polarized voting supports the district court’s
conclusion that white bloc voting usually defeats the minority candidate. There have been eight different elections for the Baytown City Council where minority members were candidates, and the minority candidate has never won. Three election results are illustrative of the white bloc voting. In the 1986 race for City Council Position 1, Mario Delgado received approximately 83% of the minority votes but only 37% of the Anglo vote and lost the election. In the 1984 race for Position 1, Tony Campos received 63% of the minority vote but only 29% of the white vote and lost the election. In the 1979 race for Position 3, Ruby Hardy received approximately 78% of the minority vote but only 3% of the white vote and lost the election. In all, the evidence was sufficient to show that the whites vote sufficiently as a bloc to overcome the minority votes plus the white “crossover” vote.
Baytown argues that the preferred candidate of the minority does win a significant percentage of the elections. Here the argument again points to Anglo-Anglo contests. As already stated the inquiry of racially polarized voting properly focused only on those contests that had a candidate sponsored by the minority group. Those were the contests that had a minority member as a candidate.
Baytown also argues that there is no white bloc voting. Specifically, they argue that Mario Delgado received 44% of the total vote including many whites. Therefore, they conclude, the whites do not vote as a bloc. The argument misses the point.
Gingles
does not require total white bloc voting. Instead, it requires only that “white majority votes sufficiently as a bloc to enable it... usually to defeat the minority’s preferred candidate.”
Gingles,
106 S.Ct. at 2767. The Delgado election is a good example: despite overwhelming minority support (83%), the whites voted as a bloc (63%) to defeat him. The fact that Delgado had support of 37% of the whites, and came relatively close to winning, does not mean that there was not determinative white bloc voting. Instead, the white bloc voting defeated the combined strength of minority votes plus white crossover votes.
Gingles,
106 S.Ct. at 2769.
Finally, Baytown argues that the minority candidate does not always lose. They point to the fact that David Smith, a Black man, won an at-large seat for the Goose Creek I.S.D. Although the Goose Creek I.S.D. is a relevant political entity, it also includes precincts 'outside of Baytown. Those precincts voted overwhelmingly for Smith and the evidence was conflicting as to whether Smith would have won if the election was exclusively in Baytown. Indeed, the statistical evidence,
see supra
n. 9, indicates that despite near unanimous minority support, Smith received only 37% of the white vote in the Baytown precincts. That was the same percentage that Delgado, who lost, received. In other words, that election does not necessarily indicate that a minority member could win a seat on the Baytown City Council. In fact, none ever have. Additionally, the occasional win by a minority’s preferred candidate would not preclude the finding of racially polarized voting.
See Gingles,
106 S.Ct. at 2779-80.
The district court’s finding of white bloc voting, the third part of the
Gingles
test, was not clearly erroneous.
D.
After making the initial
Gingles’
factors determination, the district court turned to the
Zimmer
factors
to determine whether, under the totality of the circumstances, there was a § 2 voter dilution claim. The court found that there was racially polarized voting, that the Blacks and Hispanics suffer the lingering socio-economic effects of past official discrimination, and that no minority has ever been elected to the Baytown City Council. The district court did find that the city had been responsive to the needs of the minority but concluded that that one factor does not mitigate the seriousness of the minorities “inability to participate fully in the electoral system in Baytown.” Hence, the court
concluded that a § 2 violation of minority voter dilution existed.
Baytown’s principal argument at this point is that a finding of responsiveness precludes a finding of a § 2 violation. We have rejected that argument before,
Jones v. City of Lubbock,
727 F.2d 364, 381 (5th Cir.1984), and we do so again. A finding of responsiveness does not necessarily preclude a § 2 violation. As we stated before, “Congress has expressly disapproved excessive reliance on responsiveness.”
Id.
Baytown’s argument that the district court’s finding of a § 2 violation was clearly erroneous is without merit.
IV.
Having upheld the district court’s finding of a § 2 violation, we cannot affirm the court’s remedy. The 5-3-1 plan was proposed by the City and accepted by the district court. Since that was a legislative plan (i.e., proposed by the political body) it must be precleared.
McDaniel v. Sanchez,
452 U.S. 130, 153, 101 S.Ct. 2224, 2238, 68 L.Ed.2d 724 (1981) (error for court to act on county’s proposed plan before submitted for preclearance);
Wise v. Lipscomb,
437 U.S. 535, 542, 98 S.Ct. 2493, 2498, 57 L.Ed.2d 411 (1978) (district court should not address a plan proposed by the legislative body in response to a voter dilution claim until the proposed plan has been pre-cleared);
see also
42 U.S.C. § 1973c (pre-clearance). The 5-3-1 plan has not been precleared. Both sides agree that we must vacate the adoption of the plan. We vacate the judgment and remand to the district court in order that Baytown may submit the plan for preclearance.
VACATED and REMANDED.
[[Image here]]
[[Image here]]