Fourteen county recorders and other public officials of Arizona appeal from a judgment of a three-judge district court holding the State’s 50-day durational voter residency requirement and its 50-day voter registration requirement unconstitutional under the decision in Dunn v. Blumstein, 405 U. S. 330 (1972).1 A permanent injunction was entered against enforcement of these or any other greater-than-30-day residency and registration requirements in any election held after November 1972. Appellants do not seek review of the District Court’s judgment insofar as it enjoins application of the 50-day requirements in presidential elections. See Voting Rights Act Amendments of 1970, 84 Stat. 316, 42 U. S. C. § 1973aa-l.2 Appellants assert, however, that the re*680quirements, as applied to special, primary, or general elections involving state and local officials, are supported by sufficiently strong local interests to pass constitutional muster. We agree and reverse.

In Dunn v. Blumstein, we struck down Tennessee’s durational voter residency requirement of one year in the State and three months in the county. We recognized that a person does not have a federal constitutional right to walk up to a voting place on election day and demand a ballot. States have valid and sufficient interests in providing for some period of time — prior to an election — in order to prepare adequate voter records and protect its electoral processes from possible frauds. A year, or even three months, was found too long, particularly in the context of “the judgment of the Tennessee lawmakers,” who had set “the cutoff point for registration [at] 30 days before an election . . . .” 405 U. S., at 349. The Arizona scheme, however, stands in a different light. The durational residency requirement is only 50 days, not a year or even three months. Moreover, unlike Tennessee’s, the Arizona requirement is tied to the closing of the State’s registration process at 50 days prior to elections and reflects a state legislative judgment that the period is necessary to achieve the State’s legitimate, goals.

We accept that judgment, particularly in light of the realities of Arizona’s registration and voting procedures. Those procedures, apparently first adopted during the Populist Era, rely on a “massive” volunteer deputy registrar system. See Ariz. Rev. Stat. Ann. § 16-141. According to appellants’ testimony, although these volunteers make registration convenient for voters, they average 1.13 mistakes per voter registration and the county recorder must correct those mistakes before certifying to *681the “completeness and correctness” of each precinct register. Ariz. Rev. Stat. Ann. § 16-155. The District Court itself noted that there were estimates that “in Maricopa County alone, some 4,400 registered voters might be denied the right to vote if the county voter list is in error by only one percent.”

An additional complicating factor in Arizona registration procedures is the State’s fall primary system. The uncontradicted testimony demonstrates that in the weeks preceding the deadline for registration in general elections — a period marked by a curve toward the “peak” in terms of the registration affidavits received — county recorders and their staffs are unable to process the incoming affidavits because of their work in the fall primaries. It is only after the primaries are over that the officials can return to the accumulated backlog of registration affidavits and undertake to process them in accordance with applicable statutory requirements.

On the basis of the evidence before the District Court, it is clear that the State has demonstrated that the 50-day voter registration cutoff (for election of state and local officials) is necessary to permit preparation of accurate voter lists. We said in Dunn v. Blumstein that “[f]ixing a constitutionally acceptable period is surely a matter of degree. It is sufficient to note here that 30 days appears to be an ample period of time for the State to complete whatever administrative tasks are necessary to prevent fraud — and a year, or three months, too much.” 405 U. S., at 348. In the present case, we are confronted with a recent and amply justifiable legislative judgment that 50 days rather than 30 is necessary to promote the State’s important interest in accurate voter lists. The Constitution is not so rigid that that determination and others like it may not stand.

*682The judgment of the District Court, insofar as it has been appealed from, is

Reversed.

dissenting.

In Dunn v. Blumstein, 405 U. S. 330, 348 (1972), just last Term, we held that a 30-day residency requirement provided the State with “an ample period of time . . . to complete whatever administrative tasks are necessary to prevent fraud” in the process of voter registration. We made that judgment in light of the facts that Congress had made a similar judgment as to presidential and vice-presidential elections, 42 U. S. C. § 1973aa-l (a) (6), that roughly half the States had periods of similar length, 1972-1973 Book of the States 36-37 (as of time of decision), and that the evidence needed to determine residency was relatively easy to find. The District Court, after hearing evidence about the administrative burdens in Arizona, found that appellants needed no longer than 30 days to complete the same tasks. I find nothing in the record that leads me to conclude that this judgment was erroneous.

The Court relies on two factors to justify the longer period. First, Arizona’s volunteer registrar system is said to result in so many errors that their correction requires 45 days. But these errors occur only because the deputy registrars are inadequately trained and the central supervision of the data-control process is not well organized. The District Court found that “under present conditions, at least forty-five days are required to make a voter list as free from error as possible” (emphasis added). This justified its refusal to enjoin the operation of the statute as to the election held in November 1972. But appellant Marston’s testimony was directed almost exclusively to what can only be considered readily *683solvable problems caused by untrained personnel in a relatively small office. Appellants presented no evidence that improvements in the administration of the deputy registrar system, including earlier recruitment and better training of deputy registrars and of data-processing personnel in the central offices, could not be adopted before the next election. If, as we held in Dunn, the State “cannot choose means which unnecessarily burden or restrict constitutionally protected activity,” and if the State must carry “a heavy burden of justification,” 405 U. S., at 343, surely it must show that it cannot, by better administration, eliminate the errors that justified a 50-day period in 1972. The District Court, in my view, correctly concluded that “the State has presented no facts demonstrating a compelling interest” in its 50-day requirement.

The second “complicating factor” is said to be the burden on county recorders caused by the need to interrupt the processing of affidavits filed by new registrants in order for them to work on the fall primaries. Here too the appellants showed no need to use small staffs. It is by no means obvious that the recorders’ staffs could not be increased temporarily to deal with this “complication.” Certainly that is a method of processing affidavits which less seriously burdens the right to vote. “And if there are other, reasonable ways to achieve those goals with a lesser burden on constitutionally protected activity, a State may not choose the way of greater interference.” Dunn v. Blumstein, supra, at 343.

In addition, appellants have established a system to register voters for presidential and vice-presidential elections, in compliance with the requirement of 42 U. S. C. § 1973aa-l (d), that no State may impose a residency requirement of greater than 30 days for such elections. In Arizona, those voters who qualify for presidential and vice-presidential elections, but not for state elections, *684are given absentee ballots. This eliminates the necessity to prepare a separate list of registration lists. Any administrative problems caused by the inability to correct misspellings, to alphabetize the lists, and to determine in which precinct the voter lived — the only difficulties which appellants mentioned in their testimony* — could be eliminated by similar treatment of late registrants for all elections. And if these voters did not have to appear at the polls, the fears of deterring other voters by delays at the polling places would disappear.

Even if the evidence below established that the administrative burdens of a 30-day limitation on general registration could not possibly be removed, that would not itself justify the same limitation on registration of newly arrived voters. General registration requirements affect every voter in the State. Durational residency requirements affect a much smaller class of potential voters, and the burdens of registering the members of that class will therefore be significantly smaller. Further, general registration requirements, with which any otherwise eligible voter may comply if he acts with sufficient diligence, might be thought to impair less substan*685tially the right to vote than do durational residency requirements, which bar a newly arrived voter from any participation in the elections. Serious administrative problems might justify the less severe impairment, but a total bar to participation can be justified only by administrative problems of the highest order.

In short, the evidence produced below abundantly supports the District Court’s conclusion that appellants had failed to carry the heavy burden of justifying the 50-day limitation period in light of reasonably available and less restrictive alternatives. If this Court has drawn a line beyond which reliance on administrative inconvenience is extremely questionable, as we did in Dunn, we can avoid an unprincipled numbers game only if we insist that any deviations from the line we have drawn, after mature consideration, be justified by far more substantial evidence than that produced in the District Court by appellants. I would therefore affirm the judgment of the District Court.