This is a vote dilution case in which the plaintiffs, asserting rights under the first, thirteenth, fourteenth and fifteenth amendments and 42 U.S.C. §§ 1971, 1973, and 1983, allege that the purpose and effect of the at-large election system for constituting the City Council in Columbia, South Carolina, is to dilute the voting rights of black citizens. The district court held against the plaintiffs on the merits, expressly determining that the at-large system did not violate the fourteenth or fifteenth amendments, and, without specifically addressing them, also necessarily rejecting the plaintiffs’ other constitutional and statutory claims. After the district court’s decision, the Supreme Court decided City of Mobile v. Bolden, 446 U.S. 55, 100 S.Ct. 1490, 64 L.Ed.2d 47 (1980). On appeal, the plaintiffs seek reversal of the court’s fourteenth and fifteenth amendment determinations and press their other constitutional and statutory claims. Alternatively, they seek a remand allowing them to present additional evidence of discriminatory purpose in light of Mobile.
We find no reversible error in the district court’s ultimate determination that the at-large system violates none of plaintiffs’ constitutional or statutory rights. We decline to order a remand for the presentation *917of further evidence, for in our view City of Mobile v. Bolden, while admittedly not altogether clear in its ultimate implications, did not change the law to the extent requiring reconsideration of the claims on a reopened record. We therefore affirm the district court’s dismissal of the plaintiffs’ individual claims, but we find reversible error in its certification of the action as a class action at the time of its dismissal of the individual claims on the merits.
I
Columbia has a council-manager form of government, adopted pursuant to the South Carolina Home Rule Act in 1975. The present plan includes four councilmen and a mayor, elected at large from the City of Columbia. This general form of government has been in existence since 1910, when Columbia adopted a commission form of government to replace a system under which the City was governed by an intendant and wardens or aldermen elected from districts.
Prior to 1975, nominations for the office of mayor and councilmen were made through political parties by use of a primary, convention or petition. In September 1975, the Columbia City Council adopted a non-partisan format for election. The present system, in part adopted in 1977, provides that the election process is open to any registered elector who is a resident of the City. Candidates may campaign on particular party platforms or for a particular party, but there is no indication of party affiliation on the ballot. Any candidate who receives a majority of the votes in the election is elected, while those who receive less than a majority must participate in a run-off between one more than the number of candidates necessary to fill the offices. Those receiving the highest number of votes in the run-off election are elected.
After this suit was filed, the City Council adopted a resolution proposing a change in the election process and in the number of councilmen. The proposal would have increased the number of council members to six, with three elected at large and three from single-member districts, plus the may- or elected at large. This plan was to be adopted if it had been approved by the voters in a referendum. There was no support either in the white community or the black community for this plan, and it was rejected. Black leaders apparently thought that it did not go far enough.
The 1970 census figures show that Columbia has a population of 113,542, of which 39,998 (35%) are black. In 1978, the date of the last city election, there were 11,464 black registered voters and 22,296 white registered voters. During the 1978 primary election for City Council, 36.8% of the blacks voted and 41.5% of whites voted. Blacks have been running regularly for the City Council for the past ten to twelve years, but no black has been elected to City Council within living memory.
As expressly found by the district court, there has been a long history of de facto and de jure discrimination in Columbia. It has touched all areas of life. Until 1948 blacks were excluded from voting by a variety of methods, including poll taxes, literacy tests and membership restrictions imposed by the Democratic Party. Schools were racially segregated from the first grade through college, and public accommodations and transportation within the city were operated on a similar basis. Although the court found this kind of overt discrimination no longer existent, it found that in the areas of “income, education, housing and employment blacks in the City of Columbia, on the average, exist at a lower socio-economic level, and there are many residential areas in the City which are racially identifiable.”
As also found by the court, however, many of the techniques which had been used in the past to discourage black voters from voting or which have been found to contribute to dilution of black persons’ votes have been eliminated. South Carolina no longer has an anti-single shot voting law. This law was found unconstitutional in 1972. Further, there is no longer a poll tax, literacy test, or any other restriction which could be applied in a discriminatory *918manner. Both major political parties are open to blacks and actively solicit their membership and support. Registration officials now visit various parts of the cities and counties in order to facilitate registration, no longer requiring voters to come to the courthouse or office of the Registration Board to register. Any registered voter may run for office by paying a small filing fee, which is waived on the filing of an affidavit declaring inability to pay. There is no primary and no requirement for party affiliation.
Voting is racially identifiable because of the racial identity of certain neighborhoods. In recent elections polarization on racial lines has remained significant, but does not approach totality. White and black candidates alike seek support from all areas of the city and receive support along non-racial lines. A black candidate has not yet been elected to City Council, but in the 1978 city election black candidates received approximately 25% of the vote in predominantly white areas and white candidates received 7% of the vote in predominantly black areas.
In terms of public employment, service on appointive governmental bodies, and receipt of essential and special governmental services, black citizens — while arguably not yet fairly served — are not on the other hand simply ignored by elected city officials. Forty-three percent of the city’s work force is black. Blacks serve by appointment on a significant number of city boards and commissions; there are two black city judges, two black city clerks, and several black supervisors and division heads in various departments. Columbia recently has applied over 90% of the Community Development Block Grant Funds to benefit low to moderate income families, including many blacks, and has employed a Community Development Director to work with HUD in an effort to provide more housing and rental units for low income groups.
Addressing directly only the fourteenth and fifteenth amendment claims, the district court made detailed findings of evidentiary and ultimate fact to which it applied the vote dilution principles set out in White v. Regester, 412 U.S. 755, 93 S.Ct. 2332, 37 L.Ed.2d 314 (1973), as articulated in Zimmer v. McKeithen, 485 F.2d 1297 (5th Cir. 1973) (en banc), aff’d on other grounds 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 concluded that the plaintiffs had failed under White v. Regester’s standards to prove a discriminatory effect traceable to the at-large voting system. Further, it concluded that there had been a failure to prove any racially discriminatory intent or purpose, as required by Village of Arlington Heights v. Metropolitan Housing Development Corporation, 429 U.S. 252, 97 S.Ct. 555, 50 L.Ed.2d 450 (1977), to establish or maintain the system. Having found no discriminatory effect or purpose in violation of the fourteenth or fifteenth amendment, and by necessary implication having rejected all other claims, the district court dismissed the action, certifying it as a class action in the process.
II
We first address the fourteenth and fifteenth amendment claims; next, the plaintiffs’ contention that the case should at least be remanded for re-trial on a reopened record in light of Mobile; then the other constitutional and statutory claims; and, finally, the action of the district court in certifying the action as a class action.
A
The controlling principles for assessing the plaintiffs’ fourteenth and fifteenth amendment claims are now to be sought in City of Mobile v. Bolden, 446 U.S. 55, 100 S.Ct. 1490, 64 L.Ed.2d 47 (1980). The parties predictably disagree about the effect of Mobile on the merits of this case and, beyond that, about the proper disposition of this appeal if we conclude that, tested by Mobile, the plaintiffs’ claims fail of proof.
Without attempting here a comprehensive exegesis of the controlling plurality *919and concurring opinions in Mobile1 and of the main lines of earlier authority upon which they were based,2 we accept as authoritative and controlling for our purposes the following principles drawn directly and by necessary implication from those opinions considered in the aggregate.
Claims of racially discriminatory vote dilution exist under both the fifteenth amendment and the Equal Protection Clause of the fourteenth amendment. They are essentially congruent since, under either, the claim can only be established by proof (a) that vote dilution, as a special form of discriminatory effect,3 exists and (b) that it results from a racially discriminatory purpose 4 chargeable to the state. Id. at 66-70, 100 S.Ct. at 1499-1502 (plurality opinion).
The first inquiry in assessing proof of a vote dilution claim is whether there is — without regard to motivating purpose — a discriminatory “effect” traceable to the challenged state action. This is the critical threshold issue: are the voting rights of the racial minority, as a matter of fact, impermissibly “diluted”? Vote dilution as a discriminatory effect or condition — whether viewed from the fourteenth amendment or fifteenth amendment perspectives — is not more specifically defined by the Mobile Court than as a condition in which the “voting potential” of a racial minority has been “minimize[d] or canceled] out,” id. at 66, 100 S.Ct. at 1499, or the “political strength” of such a group “adversely affectfed],” id. at 84, 100 S.Ct. at 1509 (Stevens, J., concurring); or its “access ... to the political process” demonstrably “restricted,” id. at 69, 100 S.Ct. at 1501 (plurality opinion).
More precise meaning may, however, be gleaned from earlier decisions specifically approved in Mobile in which dilution claims were assessed in relation to particular facts, and from the Mobile Court’s reiteration of the fundamental negative principle that disproportionate representation and consistent “electoral defeat,” though traceable to the *920challenged system, do not alone constitute impermissible dilution, id. at 75 & n.22, 76, 100 S.Ct. at 1504 & n.22,1505; id. at 86, 100 S.Ct. at 1510 (Stevens, J., concurring). Factors other than bare electoral defeat and consequent disproportionate representation that may, in the aggregate, demonstrate a condition of impermissible dilution are those synthesized from earlier Supreme Court decisions in Zimmer v. McKeithen, 485 F.2d 1297 (5th Cir. 1973) (en banc), aff’d on other grounds sub nom. East Carroll Parish School Board v. Marshall, 424 U.S. 636, 96 S.Ct. 1083, 47 L.Ed.2d 296 (1976):
“lack of access to the process of slating candidates, the unresponsiveness of legislators to [the racial minority’s] particularized interests, a tenuous state policy underlying the preference for . . . at-large districting, . . . the existence of past discrimination in general precludpng] the effective participation in the election system ... [and such ‘enhancing’ factors as] large districts, majority vote requirements, anti-single shot voting provisions and the lack of provision for at-large candidates running from particular geographical subdistricts.”
Id. at 1305 (footnotes omitted).
Proof by such means that a discriminatory effect (or disproportionate impact, or dilution of voting potential) exists does not end the inquiry; it is also necessary to establish “that the disputed plan was ‘conceived or operated as [a] purposeful devic[e] to further racial . . . discrimination,’ ” Mobile, 446 U.S. at 66, 100 S.Ct. at 1499 (quoting Whitcomb v. Chavis, 403 U.S. 124, 149, 91 S.Ct. 1858, 1872, 29 L.Ed.2d 363 (1971)).
The “Zimmer factors” showing the existence of a disproportionate impact or effect “may provide an important starting point” for proving by inference the requisite intent to cause or perpetuate it. Mobile, 446 U.S. at 70, 100 S.Ct. at 1501. But, at least where the challenged system’s existence “is readily explainable on grounds apart from race . . ., disproportionate impact alone cannot be decisive, and courts must look to other evidence to support a finding of discriminatory purpose.” Id. To the extent the Fifth Circuit’s opinion in Zimmer v. McKeithen had implied the contrary, it is expressly disapproved in Mobile.
B
Applying these principles to the record before us, we find no reversible error in the district court’s rejection on the merits of the congruent fourteenth and fifteenth amendment claims. As indicated, the court addressed both the “effect” and “purpose” elements of these claims and determined that the plaintiffs’ proof established neither. Critical to that determination were certain basic and ultimate findings of fact which we review under the clearly erroneous standard, Fed.R.Civ.P. 52(a), given special color here by the Supreme Court’s admonition that in vote dilution cases special deference is owed the trial court’s superior vantage point in making the required “intensely local appraisal of the design and impact” of a challenged voting system. White v. Regester, 412 U.S. 755, 769, 93 S.Ct. 2332, 2341, 37 L.Ed.2d 314 (1973). Applying that standard of review, we conclude that the court’s dispositive fact findings against the plaintiffs on both the “effect” and “purpose” elements are not clearly erroneous and support the legal conclusions that the claims were not established. We take these in turn.
(D
The “Zimmer criteria” for assessing a racially discriminatory effect simply recognize two essential means by which any identifiable minority group’s votes can be effectively diluted. The first and most obvious are formal rules and practices integral to an electoral system that inescapably prevent “effective participation” by the minority in the total process by which elections and referenda are conducted. These range from such direct impediments as poll taxes, literacy requirements, candidate qualifications and the like to more indirect technical forms such as majority voting and anti-single shot provisions, party primaries, and the like. Obviously these in sufficient combination can operate directly to “submerge” the minority in a racially polarized voting con*921stituency; to “minimize or cancel out its voting potential;” to affect adversely its “political strength.” See, e. g., White v. Regester, 412 U.S. 755, 93 S.Ct. 2332, 37 L.Ed.2d 314 (1973).
Even where technical impediments of this kind do not exist in sufficient combination to dilute a minority’s voting potential, there may yet exist more subtle impediments traceable to the combined effects of an at-large electoral system interacting with various forces of continuing social and economic disadvantage, also chargeable to the state, that have this effect. Direct evidence of these subtleties is not likely to be available, and the only evidence of their existence and impact may lie in the circumstance that the officials in power are demonstrably “unresponsive” — in either an absolute or relative sense — to the needs and interests of the minority. Such a lack of responsiveness may raise the fair inference that within the total social, economic and political context in which the electoral system operates, its practical effect has been to cancel out that residual political strength — based upon the continuing threat of reprisal at the polls — that is normally possessed even by currently outnumbered interest groups in the voting constituency. See Nevett v. Sides, 571 F.2d 209, 223 (5th Cir. 1978), cert. denied, 446 U.S. 951, 100 S.Ct. 2916, 64 L.Ed.2d 807 (1980).
In its general assessment of the evidence of effect under these Zimmer criteria, the district court recognized that there was some evidence supporting the plaintiffs’ claim that dilution by some means had been shown. As the court observed, the black minority in Columbia still experiences the legacy of many years of de facto and de jure racial segregation, though its more overt forms in public education, transportation and accommodations, and, to some extent, in employment, have now been substantially eradicated under compulsion of law. Vestiges remain in many forms of continuing social and economic disadvantage and in racially identifiable residential areas. These indisputable facts the court perforce found, along with the bald fact that within living memory no black person had been elected to the City Council. Within the Zimmer analysis, these facts — either conceded or indisputably shown — demonstrate a history of racial discrimination whose continuing practical effects — aside from any effects traceable directly to the at-large voting system itself — may be assumed still to inhibit to some extent “effective participation in the election system” by the black minority. Zimmer v. McKeithen, 485 F.2d at 1305.
In the Zimmer equation these facts militate in favor of the plaintiffs’ claims of a discriminatory effect traceable to the challenged system, and the district court so credited them.5 Nevertheless, the court was also entitled to reject as vote dilution per se the at-large feature itself, Mobile, 446 U.S. at 66, 100 S.Ct. at 1499, as well as the ultimate consequences arguably traceable to the a1>large system of consistent electoral defeat and consequent disproportionate representation of the black minority in the municipal government of Columbia, id. at 75, 76, 100 S.Ct. at 1504, 1505. Furthermore, the court properly weighed against the claim of discriminatory effect the fact that there no longer exists any of the technical legal impediments to voting of the kind that have weighed heavily in favor of dilution claims in other cases. Though concededly of recent origin and largely the result of direct or indirect legal compulsion, there is at present, under controlling state law, no poll tax, no prohibition against single-shot voting, no literacy test, no impediment to joining political parties or participating in their activities, no primary and no *922requirement for party affiliation or other barrier to candidacy. Cf. White v. Regester, 412 U.S. 755, 93 S.Ct. 2332, 37 L.Ed.2d 314 (1973); Lodge v. Buxton, 639 F.2d 1358, 1377-79 (5th Cir. 1981). Neither was there any evidence of practical impediments to voting in the form of veiled coercion chargeable directly or indirectly to the state. Cf. id. 639 F.2d at 1381 (“symbiotic relationship” between private and public sector). Though most assuredly excluded in the past from effective participation in the electoral process by both ruthlessly direct and more indirect forms of state action, blacks do now freely run for city offices and are registered and vote without hindrance in proportions comparable to whites in the city constituency. Recent voting patterns do not reveal even substantially total racial polarization in candidate choice, though preferences on racial lines do obviously exist, as reflected in the 1978 voting analysis earlier mentioned. Based upon these evidentiary facts, the district court found that “[t]he election process for the City Council in Columbia is the most open and accessible that can be imagined.”
Though there may be some element of hyperbole in this finding, its essence — that ho state imposed impediments, either legal or practical, to candidacies or to actual voting now exist — is unassailable as an ultimate finding of fact on the record before us.
Turning to the question whether despite this lack of technical impediments to effective participation, there might nevertheless be more subtle forms reflected in a lack of responsiveness on the part of city officials to the black minority’s “needs and interests,” the court found as an ultimate fact that “[t]he City government has been responsive to the needs of the black community.”6 Support for this ultimate finding was found in employment figures showing substantial black employment in city jobs, in significant though concededly not proportional black representation in appointive offices and supervisory positions, in the heavy allocation of federal block grant development funds to the black community, and in the substantially equal provision of essential services and improvements to black residential neighborhoods.
The plaintiffs vigorously challenge the validity of this ultimate finding of fact. Specifically they question the district court’s basic finding that blacks had been appointed “to almost all city boards and commissions,” and they point to the fact that though overall public employment of blacks was proportionately high — 43% against a general population percentage of 35% — it is concentrated in the most menial, lower paid city jobs. Their own evidence— they say — conclusively shows that, contrary to the court’s finding, city officials have been unresponsive to the black minority’s needs and interests.
We disagree with plaintiffs’ contentions respecting the proof on this issue. Responsiveness in this sense must of necessity be measured only in terms of the effective powers had by the governmental unit involved — here minority employment, minority representation in appointive offices, and the provision of essential governmental services to minority groups and individuals. But cf. Mobile, 446 U.S. at 73, 100 S.Ct. at 1503 (slight relevance of employment and services data as evidence of purposeful vote dilution). Assessing the evidence in these terms and considering that the burden of proof lay with the plaintiffs to prove unresponsiveness as an element of their dilution claim and not upon the defendants to prove responsiveness by way of rebuttal or avoidance, cf. Jones v. Pitt County Board of Education, 528 F.2d 414, 417 (4th Cir. 1975), we cannot declare the district court’s finding against the plaintiffs on this issue to be clearly erroneous.
In terms of government employment, the city’s proof that 43% of city employees were *923black against their overall population percentage of 35% was countered only by plaintiffs’ proof that black employment was largely concentrated in lower paying, more menial jobs. In terms of black representation in appointive positions, the plaintiffs’ evidence that no blacks had been appointed to some eighteen boards, commissions and committees of varying degrees of importance was countered by the city’s proof that 35 blacks did currently serve on 12 such bodies having a total membership of 157, including one as Chairman of the City Board of Adjustment, and that 2 blacks currently served as city judges, 2 as city clerks, and several as supervisors and heads of divisions in city government. In terms of the provision of essential government services, the city’s proof that the overwhelming bulk of federal block grant funds for community development recently available to the city had been applied to benefit members of the black community, and that practically all streets in the city, in both black and white residential neighborhoods, had now been paved, was opposed by evidence that one of the consequences of the city’s redevelopment efforts was dislocation of members of the black minority for whose benefit it was ostensibly intended, and by the general testimony of individual witnesses7 that the provision of governmental services in black neighborhoods had lagged behind similar services in white neighborhoods.
On total balance this evidence might be taken — if assessed most favorably to plaintiffs — to establish that the legitimate aspirations of black citizens in Columbia regarding fair public employment by the city and fair representation on its appointive boards and commissions have not yet been fully realized, and that the equal provision of essential governmental services has only been tardily and not yet perfectly achieved.8 But even so assessed it cannot be taken to establish a degree of unresponsiveness to the needs and interests of the black minority so palpable that it compels— or even fairly supports — an inference that the minority’s “voting potential” has been so effectively “cancelled out” that its residual “political strength” is presently being disregarded with confident impunity by the city’s governing body. This, we believe, is what is contemplated by the concept of “unresponsiveness” as evidence of, or indeed as the essence of,9 vote dilution by means more subtle than the direct or indirect denial of formal access to the ballot or the voting booth. Certainly, assessing the *924relevant evidence under the clearly erroneous standard, and considering that the burden of proof on the issue lay with plaintiffs, we are not “left with the definite and firm conviction that a mistake has been committed,” United States v. United States Gypsum Co., 333 U.S. 364, 395, 68 S.Ct. 525, 542, 92 L.Ed. 746 (1948), in the district court’s finding on this ultimate issue of fact.
(2)
Even were the district court’s finding of no discriminatory effect clearly erroneous, its rejection of the fourteenth and fifteenth amendment claims must still stand, for its further finding that no racially discriminatory purpose had been shown is unassailable under Mobile.
As Mobile has now established, in assessing attempted proof of discriminatory purpose in vote dilution cases where “the character of a law is readily explainable on grounds apart from race . . ., disproportionate impact alone cannot be decisive, and courts must look to other evidence to support a finding of discriminatory purpose.” Mobile, 446 U.S. at 70, 100 S.Ct. at 1501. Here, the plaintiffs’ evidence as to. both effect and purpose was essentially confined to establishing a sufficient aggregate of the Zimmer factors, on the apparent perception that by this means alone a discriminatory effect would be directly proven and a discriminatory purpose might rationally be inferred. Even had this evidence sufficed to prove the requisite effect, this would have constituted, under Mobile, merely “an important starting point,” id., and not a sufficient means standing alone for proving purpose behind effect.
It is inescapable that here both the initiation and the maintenance of the at-large voting system are readily “explainable on grounds apart from race.” The district court found as fact that the at-large system was adopted in 1910 because of “abuses of the ward system” that it replaced and that “race was not a consideration” because at the time “few, if any, blacks were voting in city elections.” These findings are amply supported by contemporaneous documents introduced into evidence by the defendants, and are generally confirmed by the Mobile Court’s intervening observation that at-large systems have come to be the norm in municipal government, reflecting a general consensus that they are less subject to corruption than are ward systems. See id. at 60 & n.7, 100 S.Ct. at 1496 & n.7, id. at 70 & n.15, 100 S.Ct. at 1501 & n.15. Faced therefore with the necessity of producing “other evidence” than the Zimmer factors alone, id. at 70, 100 S.Ct. at 1501, the plaintiffs’ attempt to prove discriminatory purpose simply fails under Mobile — indeed falls far short of that which the Mobile Court found lacking on the record in that case. See id. at 97-99, 100 S.Ct. at 1515-1517 (White, J., dissenting).10
In assessing the evidence of discriminatory purpose in the instant case, the district court of course did not have Mobile for instruction, but it found as fact, relying upon pre-Mobile authority later relied upon by the Mobile Court,11 that the at-large system had not been “maintained or operated, at any time since its adoption to promote racially discriminatory purposes.” That finding, on the record before the district court, was not clearly erroneous under Mobile.
Though the plaintiffs seek on appeal to avoid this consequence of Mobile’s substantive and procedural emphasis on the pur*925pose element,12 they contend alternatively, in obvious recognition of the difficulty they face, that the emphasis in Mobile was in critical respects so unexpected and unforeseeable that, at the least, a remand should be ordered to allow them to attempt to perfect their proof in Mobile’s light. For reasons that follow, we reject that suggestion.
First off, though it is obvious that the Mobile Court was itself sharply divided on the full implications of earlier decisions concerning the purpose element in vote dilution cases, a substantial majority of the Court agreed that those decisions established discriminatory purpose as an essential element in both Equal Protection and fifteenth amendment claims of vote denial or dilution.13 Other federal courts had also expressly recognized this at the time the instant action was tried, see, e. g., Nevett v. Sides, 571 F.2d 209, 219-21 (5th Cir. 1978), cert. denied, 446 U.S. 951, 100 S.Ct. 2916, 64 L.Ed.2d 807 (1980); and in this very case, as earlier indicated, the district court assumed, on the basis of existing authority and without discussion, that this was established constitutional law. We cannot therefore accept plaintiffs’ suggestion that they could not fairly be charged with the perception that discriminatory purpose was an essential element of these claims when they were litigated in district court.
Much less clear is the extent to which pre-Mobile decisions had presaged the limitation on proof of purpose imposed by the Mobile plurality: that where a challenged system is “explainable on grounds other than race,” discriminatory purpose cannot be inferred from discriminatory effect standing alone, but that “other evidence,” independent of effect, is required. This question, for example, divided the Mobile majority, including the plurality, that was agreed (or willing to assume) that purpose was an essential substantive element. The plurality — and perhaps only the plurality — believed that the Court was not already committed by those earlier decisions to the proposition that purpose might be inferred solely from proven effect. Other members of the Court flatly disagreed.
Assuming from this division within the Mobile Court that there was sufficient uncertainty on the point to warrant our practical inquiry into the fairness of holding plaintiffs to the deficiencies of proof now revealed, we nevertheless think that remand here is not warranted. Plaintiffs presumably adduced all the evidence available to them at the time to prove a disproportionate impact under the Zimmer criteria. Certainly they must be held to have had a full and fair opportunity to do so, since obviously they knew that such an impact or effect must be established as an essential element of these claims. They must also be held to have understood that this proof could only have inferential force in establishing purpose, and that if other inferential or direct proof of purpose were available, it too would be highly relevant and perhaps needed to buttress the inference based upon effect alone. While Mobile may have made more explicit the absolute need for other, independent evidence of *926purpose, we cannot say that the need, or certainly the helpfulness, of such evidence could not have been reasonably apparent to ordinarily prudent counsel in litigating these claims. This, we think, is determinative against the suggestion that fairness requires remand in light of Mobile. See Lodge v. Buxton, 639 F.2d at 1375-76.
We are buttressed in this position by consideration of the specific basis upon which plaintiffs, in their Reply Brief in this court, say they would be able, and should be allowed, to prove discriminatory purpose upon remand. They suggest, with a proffered forecast, that upon remand they could offer direct proof that the at-large voting system was adopted in 1910 at least in part for the specific purpose of excluding blacks from effective participation in the electoral process.
Reserving the question of the probative force or relevance of the specific evidence which the plaintiffs suggest is available to them on this issue,14 there is a fatal flaw in the premise upon which they rest the contention that fairness compels remand. They say that this type of direct evidence of purpose was of doubtful admissibility prior to Mobile; that the only clearly admissible evidence at that time was evidence of effect, from which purpose could be inferred. This contention is simply without merit. Neither general principles of procedure or evidence, nor logic, nor anything said in the relevant pre-Mobile decisions of the Supreme Court supports it. At most, as earlier noted, pre-Mobile cases may have suggested that proof of effect could alone suffice to prove purpose; none suggests that only in this way could it be proven.
When, upon trial of the instant case, the City introduced contemporaneous documents tending to show that the essential purpose of adopting the system in 1910 was to root out revealed abuses of the ward system, it clearly lay with the plaintiffs to counter that proof with the kind of evidence that they suggest should now be considered for the first time. Indeed, not only right but common prudence dictated the effort at that point.
On this basis plaintiffs have failed to demonstrate any basis for the relief they seek through remand.
Ill
Although plaintiffs’ amended complaint clearly retained the allegation that Columbia’s at-large election system violated 42 U.S.C. §§ 1971 and 1973, the district court held that these statutory claims had been abandoned. On appeal, plaintiffs pursue these claims, denying their abandonment, and argue that Congress intended that proof of impact alone is sufficient to show a violation of §§ 1971 and 1973.
Assuming without deciding that the claims were not abandoned during trial and that there is a private right of action under §§ 1971 and 1973, see Mobile, 446 U.S. at 60 & n.8, 100 S.Ct. at 1496 & n.8; Brooks v. Nacrelli, 331 F.Supp. 1350, 1352 (E.D.Pa. 1971), aff’d, 473 F.2d 955 (3d Cir. 1973), we conclude that, as a matter of law, plaintiffs may not succeed on either claim, so that remand for their consideration by the district court would be unavailing.
Section 1971(a)(1) provides that “[a]ll citizens of the United States who are otherwise qualified by law to vote . . . shall be entitled and allowed to vote at all ... elections, without distinction of race, color, or previous condition of servitude . . . .” Plaintiffs cite no authority for the proposition that the prohibitions of § 1971 encompass practices which have only an indirect effect on the worth of a citizen’s vote in addition to those which directly affect the ability to cast a vote. We find nothing in the language of § 1971(a) or in the more expansive language of § 1971(e) (“the word ‘vote’ includes all action necessary to make a vote effective”) that support it. We accordingly conclude that the allegations of *927voting dilution from the at-large election system in Columbia are not cognizable under 42 U.S.C. § 1971. See United States v. Uvalde Consolidated Independent School District, 461 F.Supp. 117, 119-20 (W.D.Tex. 1978), rev’d on other grounds, 625 F.2d 547 (5th Cir. 1980, cert. denied, 451 U.S. 1002, 101 S.Ct. 2341, 68 L.Ed.2d 858 (1981).
Section 1973 provides that “[n]o voting qualification or prerequisite to voting, or standard, practice, or procedure shall be imposed or applied by any State ... to deny or abridge the right ... to vote on account of race or color.” This section concededly appears on its face to be broader than § 1971, id. at 120, and plaintiffs argue not only that it encompasses a voting dilution claim such as this one but that the standard intended by Congress is proof only of a diluting effect, not the more stringent one of effect and invidious purpose.
We think this interpretation is precluded by Mobile. Although the plurality opinion in Mobile sought to limit the proscriptions of the fifteenth amendment and its embodiment in § 1973 to direct infringements of the right to vote, 446 U.S. at 60-65, 100 S.Ct. at 1496—1499, the rest of the Court did not concur in this limitation. See United States v. Uvalde Consolidated Independent School District, 625 F.2d at 551-52. Assuming therefore, without deciding, that vote dilution claims of this type are cognizable under § 1973, we proceed on the assumption that a majority of the Supreme Court in Mobile rejected the contention that such a claim can be established by proof alone of a disproportionate impact.15 Pending a definitive ruling by the Court on this question, we consider ourselves bound by this apparent view of a majority of the Mobile Court. Accord, Lodge v. Buxton, 639 F.2d at 1364 & n.11. On this basis, even had discriminatory effect been established here, we would nevertheless have been forced to hold as we do that the claim fails for lack of proof of discriminatory purpose.
IV
The plaintiffs also press on appeal their claims that the at-large election system violates their first and thirteenth amendment rights. We hold that to the extent either of these amendments protects the voting rights here asserted — a question we do not decide — their protections do not in any event extend beyond those more directly, and perhaps only, provided by the fourteenth and fifteenth amendments.
While Congress may arguably have some discretion in determining what kind of protective legislation to enact pursuant to the thirteenth amendment, it appears that the amendment’s independent scope is limited to the eradication of the incidents or badges of slavery and does not reach other acts of discrimination. See The Civil Rights Cases, 109 U.S. 3, 23-25, 3 S.Ct. 18, 30-32, 27 L.Ed. 835 (1883). In the realm of voting, we think the thirteenth amendment offers no protections not already provided under the fourteenth or fifteenth amendments.
In support of their first amendment claim, plaintiffs argue that “electoral victory” is at the very core of the [democratic] process, and that “[g]iven the Court’s steadfast protection of more incidental political rights under the First Amendment, it would surely be strange if the right to win elected office were not equally protected from undue infringement.” We reject this proposition in the form put. The first amendment’s protection of the freedom of association and of the rights to run for office, have one’s name on the ballot^ and *928present one’s views to the electorate do not also include entitlement to success in those endeavors. The carefully guarded right to expression does not carry with it any right to be listened to, believed or supported in one’s views. Where, as here, the only challenged governmental act is the continued use of an at-large election system, and where there is no device in use that directly inhibits participation in the political process, the first amendment, like the thirteenth, offers no protection of voting rights beyond that afforded by the fourteenth or fifteenth amendments.
V
One final issue requires attention. Plaintiffs sought at the outset to have this action certified as a class action under Fed. R.Civ.P. 23(a) and (b)(2), with a plaintiff class made up of all black citizens of the city of Columbia. In a pre-trial order the district court denied class certification on the basis that certification would serve no practical benefit to black citizens, for all would be equally affected by any action of the court. Subsequently, however, in conjunction with entry of its final judgment dismissing the action on the merits, the district court reversed its earlier order and certified the action as a class action with a plaintiff class composed of all black citizen residents of Columbia.
Plaintiffs challenge this ruling and contend that, regardless of the disposition of the appeal in other respects, this belated class certification must be set aside. We agree. The effect of this ruling, if allowed to stand, would be to make the final judgment on the merits binding upon members of the belatedly certified class as well as the individual plaintiffs, Restatement (Second) of Judgments § 85(l)(e), (2) (1980), though the “former had not been parties to, nor persons otherwise subject to being bound by, the action at any time prior to entry of the adverse judgment. Id. Aside from the serious constitutional due process questions thereby raised, see Hansberry v. Lee, 311 U.S. 32, 61 S.Ct. 115, 85 L.Ed. 22 (1940), this ruling cannot stand as an appropriate exercise of trial court discretion. It constitutes at least an abuse of discretion in application of the class action rules and must on that basis alone be reversed.
Whatever the power of courts to certify class actions after judgment on the merits favoring a class, compare Peritz v. Liberty Loan Corp., 523 F.2d 349 (7th Cir. 1975) (impermissible), with, e. g., Alexander v. Aero Lodge No. 735, Int’l Ass’n of Machinists & Aerospace Workers, AFL-CIO, 565 F.2d 1364 (6th Cir. 1977) (permissible), there can be no comparable power to certify a class action only after, or contemporaneously with, judgment on the merits against a class — at least where, as here, there has been no notice to the putative class members with consequent opportunity to be heard on the certification question, and possibly not even under those circumstances.
The bald power conferred in Fed.R.Civ.P. 23(c)(1) to alter or amend an earlier certification ruling “before the decision on the merits” (emphasis added) obviously cannot be drawn upon as a source of power for this ruling. The district court’s announced perception when it denied certification before trial that “no practical benefit would be served by certifying a class, since all citizens of Columbia would be affected equally by any action of the court” was in accord with circuit precedent, see Sandford v. R. L. Coleman Realty Co., 573 F.2d 173 (4th Cir. 1978), but it simply recognizes that so far as the interests of the putative class members in this type 23(b)(2) class action seeking only injunctive relief are concerned, noncertification as a class action is likely to be of no practical consequence. If judgment on the merits goes for the individual plaintiffs, the members of the putative class will be fully benefitted by it in practical, if not technically legal, terms; if judgment goes against the individual plaintiffs, the members of the putative class will not be legally bound by it. Obviously, however, the same irrelevance of consequences does not attend a subsequent determination to certify the class and thus bind it to an unfavorable judgment.
*929However fictive may be the assumption that class members, either actual or putative, actually rely upon judicial rulings that actions potentially affecting their interests may proceed or may not proceed as class actions, it is a necessary assumption — recognized in various ways — in the administration of the class action device. See, e. g., Susman v. Lincoln American Corp., 587 F.2d 866, 869 (7th Cir. 1978), cert. denied, 445 U.S. 942, 100 S.Ct. 1336, 63 L.Ed.2d 775 (1980). Here the putative class members were entitled to rely upon the district court’s pre-trial ruling that the action would not proceed as a class action. They are entitled to protection against detrimental reliance on that ruling. This requires setting aside the order certifying the action as a class action.
VI
The judgment dismissing the individual claims on the merits is affirmed. The order certifying the action as a class action is reversed.
IT IS SO ORDERED.
. For an example, see Lodge v. Buxton, 639 F.2d 1358, 1369-72 (5th Cir. 1981).
. Essentially, on the fourteenth amendment claim: Village of Arlington Heights v. Metropolitan Hous. Dev. Corp., 429 U.S. 252, 97 S.Ct. 555, 50 L.Ed.2d 450 (1977); Washington v. Davis, 426 U.S. 229, 96 S.Ct. 2040, 48 L.Ed.2d 597 (1976); White v. Regester, 412 U.S. 755, 93 S.Ct. 2332, 37 L.Ed.2d 314 (1973); Whitcomb v. Chavis, 403 U.S. 124, 91 S.Ct. 1858, 29 L.Ed.2d 363 (1971); on the fifteenth amendment claim: Wright v. Rockefeller, 376 U.S. 52, 84 S.Ct. 603, 11 L.Ed.2d 512 (1964); Gomillion v. Lightfoot, 364 U.S. 339, 81 S.Ct. 125, 5 L.Ed.2d 110 (1960); Terry v. Adams, 345 U.S. 461, 73 S.Ct. 809, 97 L.Ed. 1152 (1953); Smith v. Allwright, 321 U.S. 649, 64 S.Ct. 757, 88 L.Ed. 987 (1944).
. Though the four members of the Mobile plurality believed that the fifteenth amendment only protects against denials of the right to register and vote, i. e., that this is the only kind of “effect” it prohibits, Mobile, 446 U.S. at 64-65, 100 S.Ct. at 1498-1499, a majority of the Court apparently disagrees and considers that its prohibition extends as well to the vote dilution effect here in issue. See id. at 94, 102, 126-29, 100 S.Ct. at 1520, 1518, 1532-1534 (Brennan, White and Marshall, JJ., dissenting); id. at 80-83, 100 S.Ct. at 1506-1508 (Blackmun, J., concurring; semble); id. at 83-85, 100 S.Ct. at 1508-1510 (Stevens, J., concurring).
There being no apparent disagreement within the Mobile Court that the Equal Protection Clause protects against this kind of vote dilution effect, the two claims are congruent on this element in the view of a majority of the Court.
. A majority of the Mobile Court was apparently agreed that a discriminatory purpose must be proved to establish either claim. The four members of the plurality clearly so agreed. See Mobile, 446 U.S. at 62, 66, 100 S.Ct. at 1497-1499. It appears that Justices White and Blackmun concur in that view, though the latter only assumed it for purposes of his special concurrence in the result. See id. at 94-103, 100 S.Ct. at 1514-1519 (White, J., dissenting); id. at 80, 100 S.Ct. at 1507 (Blackmun, J., concurring). Justices Brennan and Marshall flatly disagreed, thinking that only a discriminatory effect need be proved under either claim. See id. at 103-05, 100 S.Ct. at 1520 (Marshall, J., dissenting); id. at 94, 100 S.Ct. at 1520 (Brennan, J., dissenting). Justice Stevens’ singular view on the matter so substantially differs from that expressed in either the plurality or opposed dissenting opinions that it must simply be counted uncertain if a clear choice between those polar positions were required. See id. at 90-94, 100 S.Ct. at 1512-1514 (Stevens, J., concurring).
. Though the Mobile plurality expressed serious reservations about the relevance and probative force of such “historical and social factors” in proving a vote dilution claim, Mobile, 446 U.S. at 75 & n.22, 100 S.Ct. at 1504 & n.22 (“gauzy sociological considerations have no constitutional basis”), we assume their continued relevance in view of the Mobile Court’s acceptance nevertheless of the continued vitality of White v. Regester, wherein their relevance was assumed and undoubtedly affected the decision. See Mobile, 446 U.S. at 69, 100 S.Ct. 1501; see also Lodge v. Buxton, 639 F.2d 1358, 1379, 1381 (5th Cir. 1981).
. The actual issue within the Zimmer analysis was of course whether the city government was “unresponsive,” an issue as to which plaintiffs had the burden of proof. Obviously, the court’s finding of “responsiveness” constitutes a finding of “not unresponsive” on the issue properly cast.
. Our characterization of this particular evidence as “general” is done advisedly and to emphasize its probative weakness. Plaintiffs rely, per their references in brief, upon the testimony of two non-expert witnesses to support their contention on this point. We have carefully reviewed that testimony. It consists essentially of impressionistic observations, unsupported by any technical data and preponderately nonspecific as to time and extent, about the provision of essential governmental services — paving, police protection, sewage, housing — to a limited number of black and white neighborhoods in and contiguous to Columbia. Without questioning the credibility of these witnesses or the essential accuracy of the general observations elicited from them, the testimony simply lacks probative force because of its nonspecificity and limited nature where specificity and comprehensiveness are both possible and critical to fair fact-finding. Cf., e. g., Lodge v. Buxton, 639 F.2d at 1377 & n.37 (specific evidence of current state of road improvements). Whether or not, as plaintiffs suggest, the district court actually “disregarded” this evidence, we are satisfied that it was properly discounted in assessing whether current unresponsiveness in the provision of services was shown.
. The Mobile plurality opinion suggests that while evidence of discrimination in employment and the provision of public services might well support other constitutional and statutory claims, it “is relevant only as the most tenuous and circumstantial evidence of the constitutional invalidity of the electoral system under which [the responsible officials] attained their offices.” Mobile, 446 U.S. at 73-74, 100 S.Ct. at 1503.
. The Fifth Circuit has now held that “unresponsiveness” is an essential element of a prima facie case of vote dilution under the fourteenth and fifteenth amendments. Lodge v. Buxton, 639 F.2d at 1375. We are prepared to say and need only say that where no direct denial of voting rights can be shown, there would seem no remaining way to prove that a racially discriminatory effect chargeable to the state nevertheless exists than by this means.
. Aside from the failure of plaintiff’s proof . there is of course defendants’ countering evidence that the current city council has recently proposed a revision of the at-large system specifically to insure black representation. While obviously not conclusive upon the issue of its possible racial motivation in “maintaining” the challenged system, it certainly bears heavily upon it.
. Specifically, Village of Arlington Heights v. Metropolitan Hous. Dev. Corp., 429 U.S. 252, 97 S.Ct. 555, 50 L.Ed.2d 450 (1977). The district court also relied upon Nevett v. Sides, 571 F.2d 209 (5th Cir. 1978), cert. denied, 446 U.S. 951, 100 S.Ct. 2916, 64 L.Ed.2d 807 (1980), in which the Fifth Circuit had held that purposeful discrimination was required under the relevant Supreme Court decisions.
. In addition to challenging directly the district court’s findings and conclusions assuming Mobile applies, the plaintiffs also suggest that it does not apply to this case. The contention is that Mobile applies only to attacks upon an entire form of government and not to an attack, as here, only upon an electoral system. In the latter type situation, it is urged, Mobiles arguably more stringent proof requirements do not apply.
We reject this contention. It is based upon a passing characterization in the Mobile plurality opinion of the action according to the remedy given by the district court, 446 U.S. at 70, 100 S.Ct. at 1501, and we ascribe to it no such substantive implication as plaintiffs urge. Justice Blackmun, indeed, thought the remedy given so much wider than the actual issue presented, that on this basis alone he thought remand required. Id. at 80, 100 S.Ct. at 1507. The Mobile Court’s introductory statement of the issue there presented identifies the claim as one identical to that presented in the instant case: “The question ... is whether this at-large system of municipal elections violates ... rights ... in contravention of federal statutory or constitutional law.” Id. at 58, 100 S.Ct. at 1495. The decision in Mobile clearly controls the decision here.
. See note 4 supra.
. Evidence of past acts of official discrimination was said by the Mobile plurality to be “of limited help” in resolving questions of present official motivation. 446 U.S. at 74, 100 S.Ct. at 1504.
. Admittedly, only the four members of the plurality can be counted firmly in such a supposed majority. See Mobile, 446 U.S. at 60-65, 100 S.Ct. at 1496-1499. Only two other Justices directly addressed the statutory question. These were dissenting Justices Brennan and Marshall, whose view that the fifteenth amendment includes no purpose requirement seems necessarily to put them in conflict with the plurality on this aspect of the congruent statutory claim. Of the remaining three Justices who did not address the statutory claim, the best estimate would appear to be that at least Justice White agreed that the statute also required proof of purpose. If it did not, this would have provided a further basis, and an even stronger one, for his disagreement with the result.