concurring in part and dissenting in part.
The Court’s opinion seeks to do justice by granting each side half of what it requests. The majority first grants appellants all they could hope for, by adopting an overly broad construction of § 5 of the Voting Rights Act. As if to compensate for its generosity, the Court then denies some of the same appellants the relief that they deserve. Section 5 is thereby reduced to a dead letter in a very substantial number of situations in which it was intended to have its full effect.1
*583I.
I shall first consider the Court’s extremely broad construction of § 5. It is best to begin by delineating the precise area of difference between the position the majority adopts and the one which I consider represents the better view of the statute. We are in agreement that in requiring federal review of changes in any “standard, practice, or procedure with respect to voting,” Congress intended to include all state laws that changed the process by which voters were registered and had their ballots counted. The Court, however, goes further to hold that a State covered by the Act must submit for federal approval all those laws that could arguably have an impact on Negro voting power, even though the manner in which the election is conducted remains unchanged. I believe that this reading of the statute should be rejected on several grounds. It ignores the place of § 5 in the larger structure of the Act; it is untrue to the statute’s language; and it is unsupported by the legislative history.
A.
First, and most important, the Court’s construction ignores the structure of the complex regulatory scheme *584created by the Voting Rights Act. The Court's opinion assumes that § 5 may be considered apart from the rest of the Act. In fact, however, the provision is clearly designed to march in lock-step with § 4 — the two sections cannot be understood apart from one another. Section 4 is one of the Act’s central provisions, suspending the operation of all literacy tests and similar “devices” 2 for at least five years in States whose low voter turnout indicated that these “tests” and “devices” had been used to exclude Negroes from the suffrage in the past. Section 5, moreover, reveals that it was not designed to implement new substantive policies but that it was structured to assure the effectiveness of the dramatic step that Congress had taken in § 4. The federal approval procedure found in § 5 only applies to those States whose literacy tests or similar “devices” have been suspended by § 4. As soon as a State regains the right to apply a literacy test or similar “device” under § 4, it also escapes the commands of § 5.
The statutory scheme contains even more striking characteristics which indicate that § 5’s federal review procedure is ancillary to § 4’s substantive commands. A State may escape § 5, even though it has consistently violated this provision, so long as it has complied with § 4, and has suspended the operation of literacy tests and other “devices” for five years. On the other hand, no matter how faithfully a State complies with § 5, it *585remains subject to its commands so long as it has not consistently obeyed § 4.3
As soon as it is recognized that § 5 was designed solely to implement the policies of § 4, it becomes apparent that the Court’s decision today permits the tail to wag the dog. For the Court has now construed § 5 to require a revolutionary innovation in American government that goes far beyond that which was accomplished by § 4. The fourth section of the Act had the profoundly important purpose of permitting the Negro people to gain access to the voting booths of the South once and for all. But the action taken by Congress in § 4 proceeded on the premise that once Negroes had gained free access to the ballot box, state governments would then be suitably responsive to their voice, and federal intervention would not be justified. In moving against “tests and devices” in § 4, Congress moved only against those techniques that prevented Negroes from voting at all. Congress did not attempt to restructure state governments. The Court now reads § 5, however, as vastly increasing the sphere of federal intervention beyond that contemplated by § 4, despite the fact that the two pro*586visions were designed simply to interlock. The District Court for the District of Columbia is no longer limited to examining any new state statute that may tend to deny Negroes their right to vote, as the “tests and devices” suspended by § 4 had done. The decision today also requires the special District Court to determine whether various systems of representation favor or disfavor the Negro voter — an area well beyond the scope of § 4. Section 4, for example, does not apply to States and localities which have in the past permitted Negroes to vote freely, but which arguably have limited minority voting power by adopting a system in which various legislative bodies are elected on an at-large basis. And yet, in Fairley v. Patterson, No. 25, the Court holds that a statute permitting the at-large election of county boards of supervisors must be reviewed by federal authorities under § 5. Moreover, it is not clear to me how a court would go about deciding whether an at-large system is to be preferred over a district system. Under one system, Negroes have some influence in the election of all officers; under the other, minority groups have more influence in the selection of fewer officers. If courts cannot intelligently compare such alternatives, it should not be readily inferred that Congress has required them to undertake the task.
The Court’s construction of § 5 is even more surprising in light of the Act’s regional application. For the statute, as the Court now construes it, deals with a problem that is national in scope. I find it especially difficult to believe that Congress would single out a handful of States as requiring stricter federal supervision concerning their treatment of a problem that may well be just as serious in parts of the North as it is in the South.4
*587The difficulties with the Court’s construction increase even further when the language of the statute is considered closely. When standing alone, the statutory formula requiring federal approval for changes in any “standard, practice, or procedure with respect to voting” can be read to support either the broad construction adopted by the majority or the one which I have advanced. But the critical formula does not stand alone. Immediately following the statute’s description of the federal approval procedure, § 5 proceeds to describe the type of relief an aggrieved voter may obtain if a State enforces a new statute without obtaining the consent of the appropriate federal authorities: “no person shall be denied the right to vote for failure to comply with such qualification, prerequisite, standard, practice, or procedure.” (Emphasis supplied.) This remedy serves to delimit the meaning of the formula in question. Congress was clearly concerned with changes in procedure with which voters could comply. But a law, like that in Fairley v. Patterson, No. 25, which permits all members of the County Board of Supervisors to run in the entire county and not in smaller districts, does not require a voter to comply with anything at all, and so does not come within the scope of the language used by Congress. While the Court’s opinion entirely ignores the obvious implications of this portion of the statute, the Solicitor General’s amicus brief candidly admits that this provision is flatly inconsistent with the broad reading the Government has advanced and this Court has adopted. The Government’s brief simply suggests that Congress’ choice of the verb “comply” was merely the result of an oversight. I cannot accept such a suggestion, however, when Congress’ choice of language seems to me to be consistent with the general statutory framework as I understand it.
*588B.
While the Court’s opinion does not confront the factors I have just canvassed, it does attempt to justify its holding on the basis of its understanding “of the legislative history and an analysis of the basic purposes of the Act.” Ante, at 569. Turning first to consider the Act’s basic purposes, the Court suggests that Congress intended to adopt the concept of voting articulated in Reynolds v. Sims, 377 U. S. 533 (1964), and protect Negroes against a dilution of their voting power. See ante, at 565-566, 569. It is clear, of course, that the Court’s reapportionment decisions do not apply of their own force to the problem before us. This is a statute we are interpreting, not a broad constitutional provision whose contours must be defined by this Court. The States are required to submit certain kinds of legislation for federal approval only if Congress, acting within its powers, so provided. And the fact is that Congress consciously refused to base § 5 of the Voting Rights Act on its powers under the Fourteenth Amendment, upon which the reapportionment cases are grounded. The Act’s preamble states that it is intended “[t]o enforce the fifteenth amendment to the Constitution of the United States, and for other purposes.” When Senator Fong of Hawaii suggested that the preamble include a citation to the Fourteenth Amendment as well, the Attorney General explained that he “would have quite a strong preference not to,” because “I believe that S. 1564 as drafted can be squarely based on the 15th amendment.” Hearings on S. 1564 before the Senate Committee on the Judiciary, 89th Cong., 1st Sess., pt. 1, p. 193. Attorney General Katz-enbach’s position was restated repeatedly,5 and any men*589tion of the Fourteenth Amendment is absent from this portion of the statute.6
As the reapportionment cases rest upon the Equal Protection Clause, they cannot be cited to support the claim that Congress, in passing this Act, intended to proceed against state statutes regulating the nature of the constituencies legislators could properly represent. If Congress intended, as it clearly did, to ground § 5 on the Fifteenth Amendment, the leading voting case is not Reynolds v. Sims, but Gomillion v. Lightfoot, 364 U. S. 339 (1960). While that case establishes the proposition that redistricting done with the purpose of excluding Negroes from a municipality violates the Fifteenth Amendment, it also maintains the distinction between an attempt to exclude Negroes totally from the relevant constituency, and a statute that permits Negroes to vote but which uses the gerrymander to contain the impact of Negro suffrage.
It is unnecessary, of course, to decide whether Gomillion v. Lightfoot marks the limit of the Fifteenth Amendment. It is enough to recognize that Congress did not in any way adopt the reapportionment cases’ expansive concept of voting when it enacted the Voting Rights Act of 1965. Once it is determined that Reynolds v. Sims holds no magic key to the “basic purposes” of this statute, one is obliged to determine the Act’s purposes in more traditional ways. And it is here where the Court’s opinion fails to convince. As I have already suggested, the Act’s structure assigns to § 5 a role that is a good deal more modest than the one which the majority gives it.7
*590The majority is left, then, relying on its understanding of the legislative history. With all deference, I find that the history the Court has garnered undermines its case, insofar as it is entitled to any weight at all. I refer not only to the unequivocal statement of Assistant Attorney General Burke Marshall, ante, at 564, which the Court concedes to be diametrically opposed to the construction it adopts. For the lengthy testimony of Attorney General Katzenbach, upon which the Court seems to rely, actually provides little more support for its position. Mr. Katzenbach, unlike his principal assistant, was never directly confronted with the question raised here, and we are left to guess as to his views. If guesses are to be made, however, surely it is important to note that though the Attorney General used many examples to illustrate the operation of § 5, each of them concerned statutes that had an immediate impact on voter qualifications or which altered the manner in which the election was conducted.8 One would imagine that if the *591Attorney General believed that § 5 had the remarkable sweep the majority has now given it, one of his hypothetical would have betrayed that fact.9
C.
Section 5, then, should properly be read to require federal approval only of those state laws that change either voter qualifications or the manner in which elections are conducted. This does not mean, however, that *592the District Courts in the four cases before us were right in unanimously concluding that the Voting Rights Act did not apply. Rather, it seems to me that only the judgment in Fairley v. Patterson, No. 25, should be affirmed, as that case involves a state statute which simply gives each county the right to elect its Board of Supervisors on an at-large basis.
In Whitley v. Williams, No. 36, however, Mississippi’s new statute both imposes new qualifications on independent voters who wish to nominate a candidate by petition and alters the manner in which such nominations are made.10 Since the Voting Rights Act explicitly covers “primary” elections, see §14 (c)(1), the only significant question presented is whether a petitioning procedure should be considered a “primary” within the meaning of the Act. As the nominating petition is the functional equivalent of the political primary, I can perceive no good reason why it should not be included within the ambit of the Act.
The statute involved in Bunion v. Patterson, No. 26, raises a somewhat more difficult problem of statutory interpretation. If one looks to its impact on the voters, the State’s law making the office of school superintendent appointive enacts a “voting qualification” of the most drastic kind. While under the old regime all registered voters could cast a ballot, now none are qualified. On the other hand, one can argue that the concept of a “voting qualification” presupposes that there will be a vote. On balance, I would hold that the statute comes *593within § 5. Cf. Gomillion v. Lightfoot, supra. Such a holding would not, of course, disable the State from adopting an appointive system after the force of § 5 has spent itself.
Finally, Virginia has quite obviously altered the manner in which an election is conducted when for the first time it has been obliged to issue regulations concerning the way in which illiterate voters shall be processed at the polls. Consequently, I would reverse the lower court’s decision in the Allen case, No. 3.
II.
After straining to expand the scope of § 5 beyond its proper limits, the majority surprisingly refuses to grant appellants in the Mississippi cases11 the only relief that will effectively implement the Act’s purposes. As the Court recognizes, ante, at 572, the Voting Rights Act only applies to the States for a limited period of time— Mississippi may free itself from § 5’s requirements in 1970.12 And yet the Court affords appellants in the Mississippi cases only declaratory relief, permitting state *594officials selected in violation of § 5 to hold office until their four-year terms expire in 1971.13 An election for these offices may never be held in compliance with Congress’ commands. And of course, the Court’s decision respecting relief does not only control these particular cases. There may have been hundreds of officials throughout the South who began serving long terms in office this November under procedures that have not been federally approved. As a result of this part of the Court’s decision, the Voting Rights Act may never play the full role that Congress intended for it.
It seems clear to me that we should issue a conditional injunction in the Mississippi cases along the lines suggested by the Solicitor General, except of course in the Fairley case which I think should be affirmed. Unless Mississippi promptly submits its laws to either the Attorney General or the District Court for the District of Columbia, new elections under the pre-existing law should be ordered. Of course, if the laws are promptly submitted for approval, a new election should be required only if the District Court determines that the statute in question is discriminatory either in its purpose or in its effect.
1 concur in the Court’s disposition of the complex jurisdictional issues these cases present. While I consider the question whether § 5 authorizes a three-judge court a close one, it is clear to me *583that we would not avoid very many three-judge courts whatever we decide. I would suspect that generally a plaintiff attacking a state statute because it has not been federally approved under § 5 could also make at least a substantial constitutional claim that the state statute is discriminatory in its purpose or effect. Consequently, in the usual case a three-judge court would always be convened under 28 U. S. C. § 2281. Once convened, the Court would, of course, first consider the plaintiff’s § 5 argument in the name of avoiding a constitutional question. Therefore, it appears to me that there is no good reason to invoke the normal rule that three-judge court statutes should be construed as narrowly as possible. As the Court suggests, the more natural reading of the statute confers jurisdiction on three-judge courts even in an action brought by private parties.
Section 4(c) reads:
“The phrase ‘test or device’ shall mean any requirement that a person as a prerequisite for voting or registration for voting (1) demonstrate the ability to read, write, understand, or interpret any matter, (2) demonstrate any educational achievement or his knowledge of any particular subject, (3) possess good moral character, or (4) prove his qualifications by the voucher of registered voters or members of any other class.”
The Solicitor General expressly adopts this construction of the statute in his supplemental amicus brief. In any event, the Act is clear: §4 (a) permits a State to free itself from §4 by proving to a District Court in the District of Columbia that no “test or device has been used during the five years preceding the filing of the action for the purpose or with the effect of denying or abridging the right to vote on account of race or color.” (Emphasis supplied.) As already noted, see n. 2, supra, the phrase "test or device” is a term of art including a class of statutes much narrower than those included under § 5. However, since § 5 applies by its own terms only to “a State or political subdivision with respect to which the prohibitions set forth in section 4 (a) are in effect,” a State that escapes from § 4, escapes from § 5 as well, even though it has not complied with that section.
Indeed, I would have very substantial constitutional difficulties with the statute if I were to accept such a construction.
See, e. g., Senate Hearings, supra, at 35, 141; Hearings on H. R. 6400 before Subcommittee No. 5 of the House Committee on the Judiciary, 89th Cong., 1st Sess., ser. 2, p. 102.
When, in § 10 of the Act, Congress moved against the imposition of poll taxes, it expressly invoked the Fourteenth Amendment as providing an additional basis for its action in this specific area. See §10 (b).
The Court seeks to strengthen its case by looking to the language of one of the definitional sections of the Act. Ante, at 565-566. Section 14(c)(1) defines the term “vote” or “voting” to “include *590all action necessary to make a vote effective in any primary, special, or general election, including, but not limited to, registration, listing pursuant to this Act, or other action required by law prerequisite to voting, casting a ballot, and having such ballot counted properly and included in the appropriate totals of votes cast with respect to candidates for public or party office and propositions for which votes are received in an election.” (Emphasis supplied.) All of the aspects of voting that are enumerated in this definition concern the procedures by which voters are processed. When the statute cautions that its enumeration of stages in the election process is not exclusive, it merely indicates that the change of any other procedure that prevents the voter from having his ballot finally counted is also included within the range of the Act’s concern. Surely the Court is entirely ignoring the textual context when it seeks to read the italicized phrases as embracing all electoral laws that affect the amount of political power Negroes will derive from the exercise of the franchise, even when the way in which voters are processed remains unchanged.
The examples given by the Attorney General concerned changes in a State’s voting age, residence, or property requirements; changes *591in the frequency that registrars’ offices are open; and changes from paper ballots to machines or vice versa. See House Hearings, supra. n. 5, at 60-62, 95; Senate Hearings, supra, at 191-192, 237.
The Court emphasizes three specific colloquies in which Mr. Katzenbach participated to support its understanding of the legislative history. In the most important one, see ante, at 566-567, n. 31, Senator Fong expressed concern that § 5, which at that time merely required federal review of changes in state “procedures,” would not encompass a state regulation which would radically limit the hours during which new voters could register. The Attorney General agreed that the statute should be elaborated to more clearly include such a change. Since such a law alters the manner in which voters are processed, I fail to see how this colloquy undermines my construction of the section — which clearly requires federal review in cases of the sort Mr. Katzenbach and Senator Fong were discussing. Similarly, a second extract highlighted by the Court, ante, at 567-568, is one in which the Attorney General emphasizes that § 5 is intended to prevent the States from evading the requirements of § 4 — a point I believe to count strongly in favor of the interpretation I deem the correct one. Finally, it is quite true that the Attorney General opposed carving out exceptions from § 5 that would permit the State to switch from paper ballots to voting machines without federal approval. See ante, at 568. But this fact hardly indicates that he or anyone else was of the opinion that the section required review of statutes that did not concern themselves with voting procedures. In fact, on the one occasion that Mr. Katzenbach discussed the reapportionment cases in connection with § 5, he indicated no awareness whatever that § 5 could be construed to apply to cases involving laws that change the voting power of various groups. See House Hearings, supra, at 93-94.
The statute requires supporters of a candidate to write their own names on the nominating petition, together with their polling district. Moreover, petitions must be filed by an earlier date and must contain many more signatures. The Act also imposes a “voting qualification” on those who wish to vote in a party primary, by providing that they may not subsequently compete with the primary victor by running as an independent candidate.
In the Allen case, coming from Virginia, the term of the Congressman who gained his seat under procedures that have not been approved under § 5 has already expired. Consequently, only a grant of declaratory relief is appropriate in this case, as the appellants themselves recognize.
Since the Voting Rights Act became effective in Mississippi in August 1965, the State will be able to escape the requirements of § 5 in 1970 by proving that it has not imposed a “test or device” in violation of § 4 for a five-year period. See text, at n. 3, supra. Section 5 will only continue to apply after 1970 if Mississippi is found to have continued imposing “tests or devices” after 1965. The Court’s decision today, however, does not consider whether any of the statutes involved in these cases impose a “test” or “device” within the meaning of §4, see n. 2, supra. It simply holds that the statutes fall into the much broader class of laws that modify a “standard, practice, or procedure with respect to voting” under § 5.
The state senator, state representative, county supervisor, justice of the peace, and constable involved in Whitley v. Williams, No. 36, were all elected for four-year terms ending in 1971. See Mississippi Code §3238 (1942). Similarly, the affected county superintendents of education in Bunton v. Patterson, No. 26, were appointed to four-year terms, expiring in 1971.
While I would affirm in Fairley v. Patterson, No. 25, the incumbents in that case also will serve until 1971.