with whom Mr. Justice Black joins, concurring.
The complaint in this case was filed on April 27, 1964. The District Court stayed all proceedings on June 25, 1964, until after the next regular session of the legislature and, when nothing was achieved, stayed them again until after a special session. A reapportionment plan produced by that legislature was held unconstitutional. 250 F. Supp. 537.
Thereupon the District Court drew a “temporary and provisional” plan for the general elections of 1966 and 1968. See 254 F. Supp. 997; 289 F. Supp. 827; 303 F. Supp. 224. In 1967 the legislature produced another plan which was approved by the voters and became effective January 17, 1969. This plan was also declared unconstitutional by the'District Court on July 22, 1969. The legislature then adopted a new plan effective January 22, 1970. The District Court allowed' this plan to be used for- the 1970 general election, although it considered the plan to be unconstitutional. The District Court in its decree provided:
“The court, having been advised that detailed ■ population figures for the State of Arizona will be available from the official 1970 census by the summer of 1971, assumes that the Arizona Legislature will by November 1, 1971, enact a valid plan of reapportionment for both houses of the Arizona Legislature and a valid plan of redistricting the con*117gressional districts of Arizona. Upon failure of the Legislature so to do, any party to this action may apply to the court for appropriate relief.”
The District Court also retained jurisdiction of the cause. 313 F. Supp. 148.
■ Since Reynolds v. Sims, 377 U. S. 533, Arizona has not had a constitutionally valid apportionment plan. Members of the Arizona Legislature who were elected in the 1970 election were elected under a plan the District Court held unconstitutional. Under that plan a computer was instructed to redistrict the State and to accomplish, in order, the following objectives: (1) to make the districts as equal in population as possible; (2) to circumscribe the districts in such a way that each included one incumbent senator and two incumbent representatives; (3) to make the districts compact; and (4) to make districts politically homogeneous.
Even assuming the legislative districts were of eqiial population the plan would have several practical deficiencies as far as minority representation goes. The 1970 plan insured that no incumbent would be funning against another incumbent, as often may happen under a reapportionment plan. ^ Thus the opportunity for preserving the status quo was assisted'.
An effort to make each district politically homogeneous compounded this problem. The record provides a new definition of gerrymandering. A gerrymandered district in Arizona is not one where a “natural” majority finds its power erased by either moving lines to increase the numbers of the opposition in the district or by moving the lines so that a majority is dispersed. ■ In Arizona a gerrymandered district came to be one that is overwhelmingly either Republican or Democratic. Thus when the second and fourth factors are combined an incumbent had not only the natural benefits of mourn-*118bency, but also the benefits (where possible) of a one-party district, his own fiefdom.
The record reveals that the 1970 plan heavily favored incumbents even if we assumed equal population districts. Such an assumption, of course, is contrary to the facts; deviations in Arizona ranged from about 24% above the median to about .52% below the median.
The basic unit for a district was the local political precinct. Unfortunately, there were no population figures for the basic unit, thus making it difficult to build the districts. Such figures were created by programming the computer to assume that a precinct population was that part of the 1960 county population which the number of registered voters in the precinct in 1968 bore to the number of registered voters in the county in 1968.
If all segments of society were equally likely to register to vote, then the Arizona method of computing population would be unobjectionable. But all members of a community are not equally likely to register. For example, only two counties out of eight with Spanish surname populations in excess of 15% showed a voter registration equal to the statewide average.1 Not Only are the poor, the blacks, the Chicanos, and the Indians less likely to register in the first place, they are also likely to have a higher rate of illiteracy among their members. Arizona law at the time of the decision below required a literacy test for voter registration. Ariz. Rev. Stat. Ann. §§ 16-101 (A) (4), 16-101 (A) (5). Naturally this compounded the problem of underregistration of minority groups.2 *119While the present record lacks some basic statistics, we do know that in 1965 the Bureau of the Census determined that less than 50% of the residents of voting age were registered or voted in the 1964 presidential election in Apache County, Navajo County, and Coconino County. 30 Fed. Reg. 9897, 14505. Under § 4 (a) of the Voting Rights Act of 1965, 79 Stat. 438, the application of the literacy tests was suspended by the publication. of the statistics in the Federal Register, but the suspension was lifted a year later on the showing that the literacy tests had not been used in a discriminatory manner. Apache County v. United States, 256 F. Supp. 903. As of last fall Yuma County was subject to the literacy test ban of the Voting Rights Act of 1965. See Oregon v. Mitchell, 400 U. S. 112, 131 n. 12.
The 1970 plan adversely affected minorities. Because of the registration statistics used, one district in the Phoenix ghetto had approximately 70,000 residents while an .affluent all-white district in another area of Phoenix had only '27,000 residents. The Indian reservation area in northeastern Arizona fared little better. While it had sufficient numbers of Indians to justify a separate district which could undoubtedly elect Indian representatives in the state legislature, the Indians were done in. At the time of this suit there were no Indians elected to either the State House or Senate. But just to the south of the area two state senators lived 10 miles apart. Hence, the incumbency rule was invoked to split the Indian area so as to accommodate the two white senators:
The Arizona Legislature has yet to develop, a reapportionment plan which can pass constitutional muster. The incumbents who now have the opportunity to draft *120the plan come from districts which are malapportioned and overrepresent the white vote. A valid apportion-ment plan will seemingly mean the defeat of several incumbents. The new efforts to gerrymander the State for the members of the current legislature will doubtless be prodigious. Members of the 1970 legislature had the twin advantages of running as single incumbents and in politically homogeneous districts. Members of minority groups had the disadvantage of underrepresentation. That invidious discrimination still exists.
On oral argument it was said that there is no point in initiating the design of a reapportionment plan now because the 1970 census figures- are- not available. That argument is difficult to comprehend, for it appears3 that in March 1971 New Jersey completed a comprehensive reapportionment plan based on the 1970 census. The District Court has shown great patience and has been persevering. It probably - is the first to realize that the Gordian knot must be cut if there is to be a plan that satisfies constitutional requirements.
It has indicated it will wait until November 1, 1971, before it initiates a constitutional plan. The hearings on such a plan will doubtless be long drawn out and extensive. The prize is great, for if the present incumbents can prolong matters, the 1972 election may come and go with the existing invalid 1970 plan in effect. It is not difficult to imagine how easy that strategy might be. The 1972 primaries in Arizona are in September.4
Primaries apart, there is always the problem of review by this Court. We are plagued with election cases coming here on the eve of elections, with the remaining *121time so short we do not have the days needed for oral argument and for reflection on the serious problems that are usually presented. If an election case is filed in our summer recess, we will not consider it until the first weék in October; and our effort to.note the appeal, hear the case, and decide it before November without disrupting the state election machinery is virtually impossible. The time needed is lacking.5 .
*122If a case is to'be heard and decided on these important issues it must be here by February so that we .can work it into our spring calendar of argued cases and decide it before July. If the District Court waits until November to hold hearings and put a reapportionment plan in operation, it is unlikely that any such schedule can. be met.
It is, therefore, essential that the judicial machinery be put into motion soon, so that a resolution of a matter *123that has defied solution for seven years be no longer delayed. I write these words not in criticism of the District Court but in support of its steadfast efforts to bring this stubborn litigation to an early end.
Mr. Justice Harlan concurs in the result upon the premises set forth in his separate opinions in Whitcomb v. Chavis, post, p. 165; Oregon v. Mitchell, 400 U. S. 112, 152 (1970); and Reynolds v. Sims, 377 U. S. 533, 589 (1964).
Hearings on S. 818, S. 2456, S. 2507, and Title IV of S. 2029 before the Subcommittee on Constitutional 'Rights of the Senate Committee on the Judiciary, 91st Cong., 1st and 2d Sess., 406 (1969-1970).
Because of the Voting Rights Act Amendments of 1970, literacy tests will not be a factor in future elections. Section 201, 84 Stat. 315, bars a State from denying the right to vote in any federal, *119state, or local election because of “any test or device” which is defined, inter alia, to include literacy. This part of the Act was upheld in Oregon v. Mitchell, 400 U. S. 112.
N. Y. Times, March 24, 1971, p. 47.
The primary election in Arizona in 1972 will' be held on Sept. 13. See Ariz. Rev. Stat. Ann. § 16-702.
Williams v. Rhodes, 393 U. S. 23, was an exceptional case. There Mr. Justice Stewart actihg as Circuit Justice and in consultation with available members of this Court granted injunctive relief ordering the election ballots printed in such a way as . to include the American Independent Party, the losing party in the District Court. This was to insure that, if it prevailed here,' relief would be available. 21 L. Ed. 2d 69, 89 S. Ct. 1. An expeditéd briefing schedule was authorized and we heard oral argument as soon as the Term commenced. Eight days later our opinion was handed down modifying the judgment of the District Court. Had not Mr. Justice Stewart granted the injunction in September the appellants’ victory would have been a hollow one.
A challenge to Colorado’s durational residency requirement prior to the 1968 election did not fare as well. The District Court upheld the requirement and we heard oral argument after the election was over. The case-was dismissed as moot. Hall v. Beals, 396 U. S. 45.
Durational residency requirements have' come before the Court several times' this Term. In Hayes v. Lieutenant Governor of Hawaii, there was a challenge to the Hawaii durational residency-requirement for candidates. The Hawaii Supreme Court upheld the law in late August. An application for an injunction was denied. When the appeal finally came up for consideration on the merits, again after the election, it ■was, dismissed as moot, 401 U. S. 968. In Sirak v. Brown a state durational residency requirement for voters was upheld and, when this Court denied an injunction, 400 U. S. 809, the plaintiff chose not to docket his appeal, probably on the basis of Hall v. Beals, supra. A similar issue was present in Fitzpatrick v. Board of Election Comm’rs of Chicago, where we denied a motion to expedite the appeal, 401 U. S. 905. Had all the lower courts followed Drueding v. Devlin, 234 F. Supp. 721 (Md. 1964), aff’d, 380 U. S. 125, then mootness might have prevented any plenary review of the issue. Hut several district courts have con-*122eluded that- subsequent decisions have undermined Drueding and thus have invalidated durational residency requirements. This avoids the mootness issue and we have noted probable jurisdiction in one such case, Ellington v. Blumstein, 401 U. S. 934.
In Beller v. Kirk there was a challenge to the Florida requirement demanding an independent candidate obtain 5% of the registered voters to sign a petition so that he could get on the ballot. Injunc-tive relief was denied by individual Justices early in October, but the case has subsequently been docketed sub nom. Beller v. Askew, No. 1360. We have heard oral argument on the same issue in Jenness v. Fortson, No. 5714.
The Ohio laws are involved in several cases pending this Term. In one, the District Court handed down its decision late in July 1970. By that decision several sections of the Ohio laws were invalidated and we noted probable jurisdiction. Gilligan v. Sweetenham, 401 U. S. 991. A loyalty oath was upheld and we noted probable jurisdiction in that case. Socialist Labor Party v. Gilligan, 401 U. S. 991. • The court also upheld a provision requiring independent candidates to file at the same time as major party candidates. Sweetenham v. Gilligan, No. 790. A similar issue is also presented in Pratt v. Begley, No. 1044, where the District Court for the Eastern District of Kentucky made its ruling in early October.
The then-forthcoming Chicago election in April 1971 also presented cases where one of the parties needed immediate action. In Jackson v. Ogilvie, the issue was the requirement that an independent obtain' 5% of the registered voters on a nominating petition. We denied a stay on February 22, 1971, 401 U. S. 904, and there was no way the case could be heard prior- to the-election.
Through aE these cases Williams v. Rhodes stands out as exceptional, because both the necessary preargument injunctive relief and expedited oral argument were obtained.