Seven black citizens of Norfolk, Virginia, and the Norfolk Branch of the National Association for the Advancement of Colored People appeal the district court’s judgment denying their claim that the at-large system of voting for city council violates rights secured by the Voting Rights Act of 1965 as amended in 1982, 42 U.S.C. § 1973. The principal issue is whether the complainants have less opportunity than other members of the electorate to elect more than one councilman as the “representatives of their choice.” § 1973(b).1 The complainants assign error to the district court’s construction of the Act with respect to the meaning of the statutory term “representatives of their choice.” They also contend that the district court erred by giving undue weight to the election of a second black councilman after this action was filed. In accordance with the Supreme Court’s mandate, we have reviewed the district court’s judgment in the light of Thornburg v. Gingles, 478 U.S. 30, 106 S.Ct. 2752, 92 L.Ed.2d 25 (1986).2 Finding merit in the assign-*1235merits of error, we reverse the district court’s judgment and remand the case for further proceedings.
I
In Collins v. City of Norfolk, 605 F.Supp. 377, 382-85 (E.D.Va.1984) (Collins I), the district court recounted the history of the government of the city from the grant of its royal charter in 1736 to the present. During the intervening years, there were various methods of electing the city council. Much of the time the common council was elected by wards, and the select council was chosen by members of the common council. In 1918, the city adopted a new charter, which provided for a city council of five members elected at-large. Since 1952 the council has consisted of seven members elected at-large. Council members serve four-year, staggered terms, so every two years three or four of the seven seats are contested, except on occasion when vacancies open more seats. All candidates compete for the open seats. The top three or four, as the case may be, are elected without any run-off. The elections are nonpartisan.
The district court also found that black voters in Norfolk were effectively disenfranchised by the Virginia Constitution of 1902. Among the devices used to limit black participation in elections were the literacy test, which remained in effect until the enactment of the Voting Rights Act of 1965, and the poll tax, which was declared unconstitutional in Harper v. Virginia Board of Elections, 383 U.S. 663, 86 S.Ct. 1079, 16 L.Ed.2d 169 (1966). Prior to the passage of the Voting Rights Act, black voter registration was proportionately lower than that of whites. Now the facilities of the registrar’s office are equally accessible to all citizens. The black voter registration rate and the turnout rate for council elections, based on the percentage of the black voting age population, are now greater than the registration and turnout rates of white voters. See Collins I, 605 F.Supp. at 384-85.
The 1980 census disclosed that Norfolk has a population of 266,979 persons of whom 60.8% (162,300) are white and 35.2% (93,987) are black. According to the census, the city’s voting age population is 201,-366 of whom 64.85% (130,595) are white and 31.48% (63,396) are black. Collins I, 605 F.Supp. at 382.
From 1918 until 1968, every member of Norfolk’s city council was white. In 1968, Joseph A. Jordan, Jr., a black citizen, was elected to the council, and from that time until this action was filed the council had one black member. Jordan served until 1977 when he resigned, and Rev. Joseph Green, a black citizen, was appointed to fill the vacancy. Rev. Green was elected in 1978 and re-elected in 1982 and 1986. Although the city’s population is 35% black and the rate of black participation in the electoral process is high, black citizens were unable to elect more than one black member (14%) to the seven-member council until after this case commenced. Then in 1984, Rev. John Foster, a black citizen, was elected, under circumstances that will be discussed in part VI of this opinion. Since 1984, two black members (28%) have sat on the council simultaneously.
The district court construed the statutory term “representatives of their choice” to include successful candidates who received more than 50% of the black vote, even though unsuccessful candidates received a much higher percentage of the black vote. It also held that the election of Rev. Foster in 1984, after this action was commenced, and the subsequent re-election of incumbent black councilmen demonstrated that Norfolk’s black citizens can elect representatives of their choice. These conclusions of the district court are critical in determining the validity of the district court’s judgment. They are the subject of the complainants’ primary assignments of error.
*1236II
Section 2 of the Voting Rights Act of 1965 as amended in 1982 was enacted to prohibit the “denial or abridgement of the right of any citizen of the United States to vote on account of race or color.” The Act seeks to ensure that black citizens shall not “have less opportunity than other members of the electorate to participate in the political process and to elect representatives of their choice.” 42 U.S.C. § 1973(b). The Senate Report accompanying the 1982 amendments to the Act sets forth typical factors to be considered in determining whether the Act has been violated.3 These factors are neither comprehensive nor exclusive, and complainants need not prove any particular number of them. Gingles, 478 U.S. at 45, 106 S.Ct. at 2763. Minority voters do not have to prove that the electoral system was created or is maintained for a discriminatory purpose. Congress intended that a violation could be proved by showing discriminatory effect alone. Gingles, 478 U.S. at 35, 43-44, 106 S.Ct. at 2758, 2762-63.
“The essence of a § 2 claim is that a certain electoral law, practice, or structure interacts with social and historical conditions to cause an inequality in the opportunities enjoyed by black and white voters to elect their preferred representatives.” Gingles, 478 U.S. at 47, 106 S.Ct. at 2764. The Supreme Court has long recognized that at-large voting in a multimember political unit, such as Norfolk’s city council, may prevent minorities from electing representatives of their choice by diluting their voting strength. See Gingles, 478 U.S. at 47-48, 106 S.Ct. at 2764-65. The potential for this type of discrimination may be enhanced by staggered terms. Cf. City of Rome v. United States, 446 U.S. 156, 185, 100 S.Ct. 1548, 1565, 64 L.Ed.2d 119 (1980). But at-large voting for multimember offices does not per se violate minority voters’ rights. Minority voters who assert that at-large voting in multimember political units violates the Act must prove that this electoral system “operates to minimize or cancel out their ability to elect their preferred candidates.” Gingles, 478 U.S. at 48, 106 S.Ct. at 2765. More specifically, when at-large voting is alleged to be the diluting device that violates the Act, complainants must prove three facts: 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 votes sufficiently as a bloc to enable it, in the absence of special circumstances, “usually to defeat the minority’s preferred candidate.” Gingles, 478 U.S. at 50-51, 106 S.Ct. at 2766.
*1237Racial bloc voting — or as it is sometimes called, racially polarized voting4 — is one of the factors the Senate Report identified as probative in determining whether the Act has been violated.5 Gingles recognized it as a key element of a vote dilution claim. 478 U.S. at 55, 106 S.Ct. at 2768. Racially polarized voting is important for two reasons. First, it is a means of ascertaining whether minority groups constitute a politically cohesive unit. Gingles, 478 U.S. at 56, 106 S.Ct. at 2769.6 The second purpose for inquiring about racially polarized voting is “to determine whether whites vote sufficiently as a bloc usually to defeat the minority’s preferred candidates.” Gingles, 478 U.S. at 56, 106 S.Ct. at 2769. It is in this aspect of the district court’s analysis of racially polarized voting that we detect error.
Ascertaining whether legally significant white bloc voting exists begins with the identification of the minority members’ “preferred candidates” or “representatives of their choice.”7 Then the court must inquire whether in general a white bloc vote normally will defeat the combined strength of minority support plus white “crossover” vote for the minority’s preferred candidates. Gingles, 478 U.S. at 56, 106 S.Ct. at 2769.
Whether an electoral system impairs the ability of black citizens to elect “representatives of their choice” is a question of fact subject to the clearly erroneous standard of review. Gingles, 478 U.S. at 78-79, 106 S.Ct. at 2780-81. But this standard “ ‘does not inhibit an appellate court’s power to correct errors of law, including those that may infect a so-called mixed finding of law and fact, or a finding of fact that is predicated on a misunderstanding of the governing rule of law.’ ” Gingles, 478 U.S. at 79, 106 S.Ct. at 2781 (citations omitted). We cannot accept the district court’s construction of the Voting Rights Act or its analysis of the facts based on its construction of the Act. These are errors of law or of mixed law and fact, which the clearly erroneous standard does not preclude us from correcting. Gingles, 478 U.S. at 79, 106 S.Ct. at 2781. The district court rejected the complainant’s statistical analysis of election returns and accepted the city’s. For the purpose of this opinion, we will refer only to the city’s statistics.
III
We agree with the district court that the complainants have proved the first two elements of their cause of action. The court found that Norfolk’s black population is sufficiently large and geographically compact to constitute a majority in two districts. The complainants contend that they would constitute a majority in three. In any event, the first element is satisfied. The district court also found that Norfolk’s black citizens are politically cohesive. This satisfies the second element. It is the third element that the district court found lacking. The district court held that “Norfolk’s whites do not vote sufficiently as a bloc that they usually defeat the minority’s preferred candidate.” Collins V, 679 F.Supp. at 566.
IV
The proper identification of minority voters’ “representatives of their choice” is critical. The Act was intended to assure that minorities do not have less opportunity *1238than other members of the electorate “to elect representatives of their choice.” 42 U.S.C. § 1973(b). In Collins IV, we cautioned that great care must be exercised in identifying the minority’s preferred candidates or representatives, saying:
The mere election of a candidate who appears to have received votes from more than fifty percent of minority ballots does not count as a minority electoral success, when each ballot may contain votes for more than one candidate. In such a situation, if there were other candidates, preferred by a significantly higher percentage of the minority community, Who were defeated in the same election, then it cannot fairly be said that the minority community has successfully elected representatives of its choice. Each such situation must be reviewed individually to determine whether the elected candidates can be fairly considered as representatives of the minority community. The presumption must be that they cannot, if some other candidate has received significantly more minority votes.
Collins IV, 816 F.2d at 937 (footnote omitted).
The district court treated successful candidates who received more than 50% of the minority vote as the minority’s representatives of choice even though candidates who received a much higher percentage of the minority vote were defeated. Collins V, 679 F.Supp. at 568. In 1974, Lenious G. Bond (B) received 73.4% of the black vote; he was defeated. Elizabeth M. Howell (W) received 58.2% of the black vote; she was elected. Claude J. Staylor, Jr., (W) received 56.5% of the black vote; he was elected. The district court held that Howell and Staylor, rather than Bond, were the minority’s representatives of choice. A similar situation arose in the 1980 election. Evelyn T. Butts (B) received 92.9% of the black vote; she was defeated. Howell (W) received 72.9% of the black vote; she was elected. The district court held that Howell, rather than Butts, was the minority’s representative of choice. Collins V, 679 F.Supp. at 574-75.
According to the principles we explained in Collins IV, 816 F.2d at 937, Howell and Staylor are presumed not to be the minority’s preferred candidates or representatives of choice. This presumption imposed on the city the burden of going forward with evidence to rebut or meet the presumption. Fed.R.Evid. 301. The city relied chiefly on testimony that Howell and Staylor were endorsed by a black political organization, Concerned Citizens of Norfolk, and had black support greater than some of the lesser black candidates. The district court held that the presumption that Howell and Staylor were not the minority’s representatives of choice was “overwhelmingly rebutted.” See Collins V, 679 F.Supp. at 574-75.
The proof offered by the city is insufficient to overcome the presumption that Howell and Staylor were not the black community’s representatives of choice. The Concerned Citizens’ endorsement is not dis-positive because it endorsed several candidates, black and white. Moreover, the difference between the level of black support for Howell and Staylor and the level of black support for black candidates who received less than 50% of the black vote is immaterial. Those black candidates who received few votes were not the black community’s preferred candidates. As we clearly explained in Collins IV, 816 F.2d at 937, the critical fact is the difference between the black support for the candidate who received the most black votes yet lost and for the candidates who won with fewer black votes.
Gingles, quoting the Senate Report, emphasizes that “ ‘whether the political processes are “equally open” depends upon a searching practical evaluation of the “past and present reality,” ’ and on a ‘functional’ view of the political process.” 478 U.S. at 45, 106 S.Ct. at 2763. Therefore, in addition to the bare statistics, it is appropriate to consider testimony revealing how political observers and the candidates themselves viewed the city’s claim that Howell and Staylor were the minority’s preferred candidates and representatives of choice. When questioned by the city’s attorney, Rev. Green, a black member of council who *1239served with Howell and is a defendant in this action, testified:
Q. Is Mrs. Howell looked upon in the black community as a representative of that community?
A. Not as a representative of that community, but as a supporter of that community.
Exhibit Vol. I at 1616. William E. Swin-dell, Jr., a black deputy sheriff who is one of the complainants, also testified at trial that Howell is “not a representative of our community.” App. Vol. I at 161. He explained: “Even though she receives black votes, she doesn’t live in our community. She doesn’t share our experiences.” App. Vol. I at 160.
Howell’s testimony also reveals that she cannot be considered to be the black community’s representative of choice. While being thoroughly questioned on her sources of political support, Howell admitted that she had received high levels of black support in the past, but she did not characterize herself as a representative of the black community. Exhibit Vol. II at 1831. She testified: “There are problems in the black community right now which I don’t care to be involved in, so I’m not.” Exhibit Vol. II at 1855. Howell stated that she had no opinion on the issue of school busing. Exhibit Vol. II at 1854. She did not think that the black community of Norfolk had any special needs or interests. Exhibit Vol. II at 1904. She was unaware that black people in Norfolk had higher rates of poverty and unemployment and lower levels of education than whites. Exhibit Vol. II at 1899-1900.8
Staylor also did not consider himself to be the black community’s candidate of choice. In 1974, Staylor was first elected to city council with black support after he resigned from a police officer’s organization because it refused to admit black people. Exhibit Vol. I at 1604-05. He testified that he voted against a proposal to build a- memorial to Martin Luther King because he believed King was responsible for inciting riots, associated with communists, and was not a “patriotic, good American.” Exhibit Vol. I at 1646-47. Staylor has also been an outspoken opponent of school busing. Exhibit Vol. I at 1659-63. In 1978, Staylor lost virtually all black support because of his position on issues of importance to black voters. Nevertheless, he was re-elected by white voters.
The district court’s error in finding successful candidates who received over 50% of the minority vote to be the chosen representatives of the minority community, despite the fact that other candidates received a much greater percentage of the minority vote, is not simply technical or semantic. It is an error of mixed law and fact that defeats the Voting Rights Act’s purpose of securing equal opportunity for minorities to “elect the representatives of their choice.” By declaring that Howell in two elections and Staylor in one election were the chosen representatives of the black community, the district court was able to find that there was no significant white bloc voting. On this basis the district court held that the complainants had not satisfied the third requirement of Gingles, 478 U.S. at 51, 106 S.Ct. at 2766, for showing a violation of the Act.
A moment’s reflection shows how the district court’s method of identifying the black community’s representatives of choice defeats a primary purpose of the Act. In 1974, 21 candidates sought the 4 open seats on the council. In 1980, 12 candidates vied for 3 open seats. In 1974, each voter — black and white — could cast four votes and in 1980 three votes. If black voters exercised their right to cast all of their allotted votes, they ran the risk that their second and third choices would be declared their preferred candidates. Only by single-shot voting — withholding all votes save for their first choice, and forfeiting the opportunity to cast all votes allotted to each voter — could the minority be assured that its second and third choices would not be declared its preferred candi*1240dates. In contrast, under the at-large system, the white voters can freely cast all votes allotted to them without suffering the penalty imposed on the minority voters. The district court’s construction of the Act defeats the congressional purpose of assuring that the opportunity to participate in the electoral process is equally open to all citizens.
We are aware of no case that supports the district court’s construction of the Act. Two cases have rejected it. In Citizens for a Better Gretna v. City of Gretna, 834 F.2d 496, 502 (5th Cir.1987), the court said:
In a multiple seat election such as Gret-na’s aldermanic race, the minority necessarily will have more than one preferred candidate. In the Gretna elections studied, blacks exercised their right to vote to fill all at-large positions, but only one available candidate was black. Thus, it was virtually unavoidable that certain white candidates would be supported by a large percentage of Gretna’s black voters. Significance lies in the fact that the black candidate preferred by the minority was defeated by white bloc voting. That blacks also support white candidates acceptable to the majority does not negate instances in which white votes defeat a black preference.
In Campos v. City of Baytown, 840 F.2d 1240, 1245 (5th Cir.1988), the court rejected the argument that “any time a candidate gets a majority of the minority votes he is the ‘chosen representative’ of the minority group.” We are aware that the factual situations in these cases differs from that in Norfolk, but their underlying premise is applicable: support for some successful candidates by a majority of minority voters in multimember district races does not prove that the successful candidates were the chosen representatives of the minority when a candidate who received much greater minority support was defeated.
“Definitions and legal criteria ... must accord with law, and will be reviewed as legal questions.” Collins IV, 816 F.2d at 938. The district court erred in construing the statutory term “representative of their choice.” This skewed its analysis on which it relied to decide the case.
V
If the degree of minority electoral success necessary to defeat this voting rights action were the ability to elect a single candidate to the council, this action would border on the frivolous. Since the passage of the Voting Rights Act of 1965 and the elimination of the poll tax, the black community of Norfolk has been able to elect a single councilman. But this case involves the community’s inability to elect a second member until after the institution of this action.
The city attaches great significance to the fact that black candidates have won in seven of ten elections from 1968 to 1986. It emphasizes that since 1968, of the 13 black candidates supported by 70% of black voters, 7 (54%) have been elected.
The significance of these statistics is superficial, because they are largely based on the black community’s ability to elect a single black member of the council. The statistics ignore the fact that from 1968 until after this action was brought, the black community was unable to elect a second representative to the city council. To use the city’s analysis of the statistics to defeat this action simply places a stamp of approval on token representation.
Gingles explains that “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.” 478 U.S. at 56, 106 S.Ct. at 2769. In the context of this case it is necessary to apply Gingles’ teaching to the black community’s attempt to elect a second black councilman.
It is apparent from the record that the white majority normally voted sufficiently as a bloc to defeat the combined strength of strong minority support plus white crossover votes for minority preferred candidates who sought a second seat on the council. In 1970, Roland J. Walton (B), who received 92.4% of the black vote, received only 14% of the white vote. In 1974, *1241Bond (B), who got 73.4% of the black vote, received only 7.6% of the white vote. In 1980, Butts (B), who received 92.9% of the black vote, received only 9.5% of the white vote. In 1982, Butts (B), who received 87.6% of the black vote, received only 11.2% of the white vote.
Although black voters usually tried to elect a second black councilman, this was not always the case. In 1976, for example, no black candidate opposed Jordan (B), an incumbent. He was reelected with 97.3% of the black vote and 24.3% of the white vote. The candidate whom the black voters preferred for a second representative on the council was R. Braxton Hill, Jr., (W), who received 76.3% of the black vote and only 20.0% of the white vote. His defeat demonstrates that — regardless of the race of the candidate — the black community, pri- or to the institution of this suit, could not elect a second representative of their choice.
The statistics show that from 1968 until 1984 all of the minority-preferred candidates for a second seat on the council were defeated by candidates preferred by white voters. This is precisely the situation Gin-gles described as demonstrating “legally significant white bloc voting.” 478 U.S. at 56, 106 S.Ct. at 2769. The statistics also demonstrate that the complainants have satisfied the third requirement for proving a cause of action under the Voting Rights Act when the inhibiting electoral device is at-large voting. They show that the white majority votes sufficiently as a bloc to enable it, in the absence of special circumstances, “usually to defeat the minority’s preferred candidate.” Gingles, 478 U.S. at 51, 106 S.Ct. at 2766.
VI
The city maintains that members of the black community have no cause for complaint because in addition to its success in electing a single black member to the council from 1968 to 1984, in 1984 a second black member was elected.
Gingles, however, cautions against finding a lack of racially polarized voting where the success of a minority candidate can be attributed to special circumstances. The Court said:
[T]he success of a minority candidate in a particular election does not necessarily prove that the district did not experience polarized voting in that election; special circumstances, such as the absence of an opponent, incumbency, or the utilization of bullet voting, may explain minority electoral success in a polarized contest.
478 U.S. at 57, 106 S.Ct. at 2769. The Court noted that this list of special circumstances is not exclusive. 478 U.S. at 57 n. 26, 106 S.Ct. at 2770 n. 26.
The complainants assert that the election of a second black councilman, Rev. Foster, in 1984 was caused by special circumstances and so his election does not negate the existence of racially polarized voting. The city maintains that the district court’s finding that Rev. Foster’s election was not the result of special circumstances is not clearly erroneous.
In Zimmer v. McKeithen, 485 F.2d 1297 (5th Cir.1973) aff'd sub nom East Carroll Parish School Board v. Marshall, 424 U.S. 636, 96 S.Ct. 1083, 47 L.Ed.2d 296 (1976), the court said:
[W]e cannot endorse the view that the success of black candidates at the polls necessarily forecloses the possibility of dilution of the black vote.... [S]uch success might be attributable to political support motivated by different considerations — namely that election of a black candidate will thwart successful challenges to electoral schemes on dilution grounds.
485 F.2d at 1307.
Zimmer’s example illustrates precisely what happened in Norfolk before the election of the second black councilman in 1984 after this action had been filed. For the first time the mayor supported a second black candidate. Moreover, during the pendency of this action, in which the mayor was a defendant, he publicly stated: “After the election, the issue of black representation may become a moot point.” Exhibit Vol. I at 1472. In the same election the “west side group” with which the mayor *1242was affiliated endorsed only two candidates when three seats were vacant. One of the city’s witnesses, a member of the House of Delegates of the Virginia General Assembly, described this situation as “unique” or “surprising to say the least.” App. Vol. II at 845.
When this case was previously before us after remand from the Supreme Court, we directed the district court to ascertain whether the mayor’s remark elicited unusual white support for the black candidate. We added that the district court’s findings would not be disturbed unless they were clearly erroneous. See Collins IV, 816 F.2d at 938. The district court held that there was no unusual white voting pattern in 1984. See Collins V, 679 F.Supp. at 577-78.
Review of the record convinces us that the district court’s finding is clearly erroneous. See United States v. United States Gypsum Co., 333 U.S. 364, 395, 68 S.Ct. 525, 541, 92 L.Ed. 746 (1948); Fed.R.Civ.P. 52(a). The flaw in the district court’s finding is readily apparent. It compared elections in which only one representative of the black community was elected with the 1984 election in which a second black councilman was elected. In support of its finding, the district court emphasized that Rev. Foster, the successful black candidate for a second black seat on the council, received 26.6% of the white vote in the 1984 election. The court then compared this vote with the results in previous elections — in 1982, 1978, and 1968 — in which black candidates received 31.1%, 32.2%, and 28.7% of the white vote to fill a single minority seat on the council. Collins, 679 F.Supp. at 577-78.
The statistics show that until the 1984 election, when the mayor supported a black candidate and suggested that this suit could be mooted, white voters had cast ballots in sufficient numbers to defeat a second black councilman. Never before had white support for a candidate seeking to become the second black member of the council remotely approached 26.6%. Prior to 1984, the greatest white support for a black candidate seeking a second seat on the council occurred in 1970. In that year, Walton (B) received 92.4% of the black vote and 14.0% of the white vote. He was defeated. Rev. Foster’s unprecedented receipt of 26.6% of the white vote was 90% higher than the previous high white vote ever cast for a black candidate who sought a second seat on the council.
Unusual white support for Butts (B) also demonstrates that the 1984 election of Rev. Foster was due to the special circumstances arising out of events associated with the pendency of this action. In 1982, Butts received 11.2% of the white vote and in 1980 only 9.5%. Yet in 1984 after the mayor suggested how this action could be mooted and the “west side group” fielded only two candidates, she received 24.2% of the white vote. Her white vote in 1984 increased 116% over her white vote in 1982 and 155% over her white vote in 1980.
The statistics disclose that no black candidate seeking a second seat on the council as the chosen representative of the black community received more than slight white support until the unprecedented circumstances that occurred in 1984 after this action was filed. The figures also demonstrate that, prior to the special circumstances of 1984, white voters were able to defeat the combined strength of black voters and white crossover votes, denying the black community a second seat on the council.
We therefore conclude that the election of a second black councilman in 1984 was the result of special circumstances and should not be dispositive in determining whether Norfolk’s at-large voting for a multimember council dilutes black voting power. See Gingles, 478 U.S. at 57, 75-76, 106 S.Ct. at 2769, 2779.
VII
The complainants also assign error to the district court’s finding that the council’s appointment of Rev. Green in 1977 to fill the vacancy resulting from Jordan’s resignation was not a special circumstance. Collins V, 679 F.Supp. at 579-80. The city asserts that there is little doubt that had a black not been appointed the complainants would have called foul and used this as *1243evidence of white council members’ insensitivity.
If, as it appears from the city’s brief, Rev. Green was appointed to forestall black criticism, this would indeed be a special circumstance. Cf. Zimmer, 485 F.2d at 1307. In any event, Rev. Green’s appointment has little bearing on the black community’s ability to elect a second representative to council, except in a subsequent election after this action was filed when he was able to run with the advantage of incumbency.
Gingles mentions incumbency as a special circumstance that may explain minority success. 478 U.S. at 57, 106 S.Ct. at 2769. In this case, incumbency was a significant factor in black electoral success. No black incumbent was ever defeated. Conversely, with two exceptions, in the eleven elections from 1968 to 1988, inclusive, no black candidate who was not an incumbent was elected. The first exception was in 1968 when Jordan became the first black member of council since the at-large system was instituted in 1918. The second exception was in 1984 when Rev. Foster was elected after this action was filed and the mayor stated that the election might moot the suit. The re-election of incumbents in 1986 and 1988, coupled with the pendency of this action and the manifest desire to moot it as evidenced by the 1984 election, also demonstrates that recent re-election of black candidates is attributable to special circumstances. These elections do not negate the existence of the white bloc voting that has characterized the at-large elections in Norfolk.
VIII
The complainants also allege that the district court erred in holding that there was no racial discrimination in the slating process in the 1976 and 1980 elections. Collins V, 679 F.Supp. at 581-83.
Slating and denial of black access to the slate are mentioned as pertinent factors in the Senate Report, see supra note 3. Nevertheless, in the context of this case, whether slating existed in 1976 or 1980 is largely immaterial. Analysis of the election returns show that from 1968 until 1984 no black candidate could get elected to a second seat on the seven-member council and the single black representative was never defeated. The same pattern prevailed in the two years in which slating was charged and in the years in which it was not. White bloc voting, not slating, was the cause of the black community’s inability to elect a second councilman in 1976 and 1980.
IX
In sum, the complainants have satisfied the tripartite test set forth in Gingles, 478 U.S. at 50-51, 106 S.Ct. at 2766, for proving a violation of the Voting Rights Act. They have shown 1) that the minority group is sufficiently large and geographically compact to constitute a majority in at least two single-member districts, 2) that the minority group is politically cohesive, and 3) that the white majority votes sufficiently as a bloc to enable it, in the absence of special circumstances, usually to defeat the minority’s preferred candidate for a second seat on the council.
To be sure, after the passage of the Voting Rights Act of 1965, since 1968 the black citizens of Norfolk have attained at least token representation on the city council. But in the absence of special circumstances, they have not been able to elect a second representative of their choice. Even the candidates elected as a result of special circumstances in 1984, 1986, and 1988 do not give representation in proportion to the black population. But the complainants do not seek proportional representation, and the Act does not require it. The Act does, however, protect them from a system of at-large voting that, tested by the principles explained in Gingles, 478 U.S. 30, 106 S.Ct. 2752, gives them “less opportunity than other members of the electorate to participate in the electoral process and to elect representatives of their choice.” 42 U.S.C. § 1973(b).
The judgment of the district court is reversed. The case is remanded for further proceedings consistent with this opin*1244ion. Upon remand the district court should enjoin at-large elections for city council. The district court should afford the city a reasonable, specified time to prepare a plan that will remedy the vote dilution arising out of the city’s at-large electoral system. The city must then submit the plan for clearance under section 5 of the Voting Rights Act of 1965. 42 U.S.C. § 1973c. If the city fails to enact a legal plan, the court should prepare a single district plan for the conduct of future elections. See generally Wise v. Lipscomb, 437 U.S. 535, 98 S.Ct. 2493, 57 L.Ed.2d 411 (1978); Cosner v. Dalton, 522 F.Supp. 350, 363-64 (E.D.Va.1981) (three-judge court).
. Section 2 of the Act, 42 U.S.C. § 1973, provides:
(a) No voting qualification or prerequisite to voting or standard, practice, or procedure shall be imposed or applied by any State or political subdivision in a manner which results in a denial or abridgement of the right of any citizen of the United States to vote on account of race or color, or in contravention of the guarantees set forth in section 1973b(f)(2) of this title, as provided in subsection (b) of this section.
(b) A violation of subsection (a) of this section is established if, based on the totality of circumstances, it is shown that the political processes leading to nomination or election in the State or political subdivision are not equally open to participation by members of a class of citizens protected by subsection (a) of this section in that its members have less opportunity than other members of the electorate to participate in the political process and to elect representatives of their choice. The extent to which members of a protected class have been elected to office in the State or political subdivision is one circumstance which may be considered: Provided, That nothing in this section establishes a right to have members of a protected class elected in numbers equal to their proportion in the population.
. This action was commenced in 1983. The district court entered judgment for the city, Collins v. City of Norfolk, 605 F.Supp. 377 (E.D.Va.1984) (Collins I), and this court affirmed, one judge dissenting, 768 F.2d 572 (4th Cir.1985) (Collins II). The Supreme Court vacated this court’s judgment, 478 U.S. 1016, 106 S.Ct. 3326, 92 L.Ed.2d 733 (1986) (Collins III), and remanded for consideration in light of Thornburg v. *1235Gingles, 478 U.S. 30, 106 S.Ct. 2752, 92 L.Ed.2d 25 (1986). This court then reversed the district court’s judgment and remanded for further proceedings, 816 F.2d 932 (4th Cir.1987) (Collins IV). Upon remand, the district court entered judgment for the city. 679 F.Supp. 557 (E.D.Va. 1988) (Collins V).
.
1. the extent of any history of official discrimination in the state or political subdivision that touched the right of the members of the minority group to register, to vote, or otherwise to participate in the democratic process;
2. the extent to which voting in the elections of the state or political subdivision is racially polarized;
3. the extent to which the state or political subdivision has used unusually large election districts, majority vote requirements, anti-single shot provisions, or other voting practices or procedures that may enhance the opportunity for discrimination against the minority group;
4. if there is a candidate slating process, whether the members of the minority group have been denied access to that process;
5. the extent to which members of the minority group in the state or political subdivision bear the effects of discrimination in such areas as education, employment and health, which hinder their ability to participate effectively in the political process;
6. whether political campaigns have been characterized by overt or subtle racial appeals;
7. the extent to which members of the minority group have been elected to public office in the jurisdiction.
Additional factors that in some cases have had probative value as part of plaintiffs’ evidence to establish a violation are:
whether there is a significant lack of responsiveness on the part of elected officials to the particularized needs of the members of the minority group.
whether the policy underlying the state or political subdivision's use of such voting qualification, prerequisite to voting, or standard, practice or procedure is tenuous.
S.Rep. No. 417, 97th Cong., 2d Sess. 28-29, reprinted in 1982 U.S.Code Cong. & Admin.News 177, 206-07 (footnotes omitted).
The district court found facts pertinent to the Senate Report. See Collins V, 679 F.Supp. 557, and Collins /, 605 F.Supp. 377. Except as noted in this opinion, the complainants do not question these findings.
. Gingles uses these terms interchangeably throughout the opinion. See 478 U.S. at 52 n. 18, 106 S.Ct. at 2767 n. 18.
. See supra note 3.
. The district court found that Norfolk’s black citizens vote sufficiently as a bloc to constitute a politically cohesive unit and neither side contests this finding. We need not discuss this aspect of polarized voting further.
. In the context of polarized voting, the Supreme Court differs over whether the race of the candidate is significant. See Gingles, 478 U.S. at 67-68, 106 S.Ct. at 2774-75 (irrelevant, plurality opinion); 478 U.S. at 82-83, 106 S.Ct. at 2782-83 (significant, White, J., concurring); 478 U.S. at 100-02, 106 S.Ct. at 2792-93 (significant, O’Conner, J., concurring); see also Collins IV, 816 F.2d at 936 n. 3 (commenting on the different views). We need not pursue this issue to decide this case. Nevertheless, we will identify the race of the candidates by (W) for white and (B) for black.
. The district court found that black persons in Norfolk generally do not fare as well as white persons with respect to income, education, housing, and employment. Collins I, 605 F.Supp. at 391.