dissenting.
I join Justice White’s excellent dissenting opinion, and reaffirm my previously expressed doubt that “the Constitution — a vital and living charter after nearly two centuries because of the wise flexibility of its key provisions — could be read to require a rule of mathematical exactitude in legislative reapportionment.” White v. Weiser, 412 U. S. 783, 798 (1973) (concurring opinion). I write separately to express some additional thoughts on gerrymandering and its relation to apportionment factors that presumably were not thought relevant under Kirkpatrick v. Preisler, 394 U. S. 526 (1969).
r-H
The Court, following Kirkpatrick, today invalidates New Jersey’s redistricting plan solely because various alternative plans, principally the one proposed by Professor Reock, had what the Court views as “appreciably smaller population deviations between the largest and smallest districts.” Ante, at 728. Under all of the plans, the maximum population variances were under 1%. I view these differences as neither “appreciable” nor constitutionally significant. As JUSTICE White demonstrates, ante, at 769-772 (dissenting opinion), the Court’s insistence on precise mathematical equality is self-deluding, given the inherent inaccuracies of the census data and the other difficulties in measuring the voting population of a district that will exist for a period of 10 years. See Kirkpatrick, supra, at 538 (Fortas, J., concurring) (pursuit of precise equality “is a search for a will-o’-the-wisp”). Moreover, it has become clear that Kirkpatrick leaves no room for proper legislative consideration of other factors, such as preservation of political and geographic boundaries, that plainly are relevant to rational reapportionment decisions,1 see Gaffney *785v. Cummings, 412 U. S. 735, 749 (1973); Mahan v. Howell, 410 U. S. 315, 329 (1973). As Justice White correctly observes, ante, at 775-776, a decade of experience has confirmed the fears of the Kirkpatrick dissenters that an uncompromising emphasis on numerical equality would serve to encourage and legitimate even the most outrageously partisan gerrymandering, see 394 U. S., at 551-552 (Harlan, J., dissenting); id., at 555 (WHITE, J., dissenting). The plain fact is that in the computer age, this type of political and discriminatory gerrymandering can be accomplished entirely consistently with districts of equal population.2
*786I therefore continue to believe that the Constitution permits variations from “theoretical 'exactitude’ in recognition of the impracticality of applying the Kirkpatrick rule as well as in deference to legitimate state interests.” White v. Weiser, supra, at 798 (Powell, J., concurring). Certainly when a State has adopted a districting plan with an average population deviation of 0.1384%, and a maximum deviation of 0.6984%, it has complied with the Constitution’s mandate that population be apportioned equally among districts.
r-H
The extraordinary map of the New Jersey congressional districts, see ante, following p. 744, prompts me to comment on the separate question of gerrymandering — “the deliberate and arbitrary distortion of district boundaries and populations for partisan or personal political purposes,” Kirkpatrick, supra, at 538 (Fortas, J., concurring). I am in full agreement with Justice White’s observation more than a decade ago that gerrymandering presents “a far greater potential threat to equality of representation” than a State’s failure to achieve *787“precise adherence to admittedly inexact census figures.” Wells v. Rockefeller, 394 U. S. 542, 555 (1969) (dissenting opinion). I also believe that the injuries that result from gerrymandering may rise to constitutional dimensions. As Justice Stevens observes, if a State’s electoral rules “serve no purpose other than to favor one segment — whether racial, ethnic, religious, economic, or political — that may occupy a position of strength at a particular point in time, or to disadvantage a politically weak segment of the community, they violate the constitutional guarantee of equal protection.” Ante, at 748 (concurring opinion). Moreover, most gerrymandering produces districts “without any regard for political subdivision or natural or historical boundary lines,” Reynolds v. Sims, 377 U. S. 533, 578-579 (1964), a result that is profoundly destructive of the apportionment goal of “fair and effective representation,” id., at 565. A legislator cannot represent his constituents properly — nor can voters from a fragmented district exercise the ballot intelligently — when a voting district is nothing more than an artificial unit divorced from, and indeed often in conflict with, the various communities established in the State.3 The map attached to the Court’s opinion illustrates this far better than words can describe.
I therefore am prepared to entertain constitutional challenges to partisan gerrymandering that reaches the level of discrimination described by Justice Stevens. See ante, at 748 (concurring opinion). I do not suggest that the shape of a *788districting map itself invariably is dispositive. Some irregularity in shape is inevitable, with the degree of irregularity depending primarily on the geographic and political boundaries within the State, as well as the distribution of its population. Moreover, political considerations, even partisan ones, are inherent in a democratic system. A court, therefore, should not “attemp[t] the impossible task of extirpating politics from what are the essentially political processes of the sovereign States.” Gaffney, 412 U. S., at 754. Finally, I do not suggest that a legislative reapportionment plan is invalid whenever an alternative plan might be viewed as less partisan or more in accord with various apportionment criteria. The state legislature necessarily must have discretion to accommodate competing considerations.
I do believe, however, that the constitutional mandate of “fair and effective representation,” Reynolds, supra, at 565, proscribes apportionment plans that have the purpose and effect of substantially disenfranchising identifiable groups of voters. Generally, the presumptive existence of such unconstitutional discrimination will be indicated by a districting plan the boundaries of which appear on their face to bear little or no relationship to any legitimate state purpose. As Justice Stevens states, “dramatically irregular shapes may have sufficient probative force to call for an explanation,” ante, at 755 (concurring opinion); “drastic departures from compactness are a signal that something may be amiss,” ante, at 758; and “[e]xtensive deviation from established political boundaries is another possible basis for a prima facie showing of gerrymandering,” ibid. In such circumstances, a State should be required to provide a legitimate and nondiscriminatory explanation for the districting lines it has drawn. See Reynolds, supra, at 568 (the apportionment “presented little more than crazy quilts, completely lacking in rationality, and could be found invalid on that basis alone”).
In this case, one cannot rationally believe that the New Jersey Legislature considered factors other than the most *789partisan political goals and population equality. It hardly could be suggested, for example, that the contorted Districts 3, 5, and 7 reflect any attempt to follow natural, historical, or local political boundaries.4 Nor do these district lines reflect any consideration of the likely effect on the quality of representation when the boundaries are so artificial that they are likely to confound the Congressmen themselves. As Judge Gibbons stated eloquently in his dissent below:
“The apportionment map produced by P. L. 1982, c. 1 leaves me, as a citizen of New Jersey, disturbed. It creates several districts which are anything but compact, and at least one district which is contiguous only for yachtsmen. While municipal boundaries have been maintained, there has been little effort to create districts having a community of interests. In some districts, for example, different television and radio stations, different newspapers, and different transportation systems serve the northern and southern localities. Moreover the harshly partisan tone of Speaker Christopher Jack-man’s letter to Ernest C. Reock, Jr. is disedifying, to say the least. It is plain, as well, that partisanship produced artificial bulges or appendages of two districts so as to place the residences of Congressmen Smith and Courter in districts where they would be running against incumbents.” Daggett v. Kimmelman, 535 F. Supp. 978, 984 (NJ 1982).
This summary statement by Judge Gibbons, a resident of New Jersey, is powerful and persuasive support for a con-*790elusion that the New Jersey Legislature’s redistricting plan is an unconstitutional gerrymander. Cf. ante, at 764, n. 33 (Stevens, J., concurring). Because this precise issue was not addressed by the District Court, however, it need not be reached here. As to the issue of population equality, I dissent for the reasons set forth above and in Justice White’s dissenting opinion. 3
The Court holds that “[a]ny number of consistently applied legislative policies might justify some variance, including, for instance, making districts compact, respecting municipal boundaries, preserving the cores of *785prior districts, and avoiding contests between incumbent Representatives.” Ante, at 740. Although it is remarkable that the Court thus silently discards important features of Kirkpatrick while simultaneously invoking stare decisis to defend the remainder of that decision, see ante, at 778-780 (White, J., dissenting), I welcome this change in the law. It is to be hoped that this new standard will be applied with a strong measure of deference to the legitimate concerns of the State. See ante, at 760, n. 26 (Stevens, J., concurring) (recognizing that courts should “give greater weight to the importance of the State’s interests and the consistency with which those interests are served than to the size of the deviations”).
An illustration is the recent congressional redistricting in Illinois. After the Illinois Legislature had failed to enact a reapportionment plan, a three-judge District Court chose among four plans varying from 0.02851% to 0.14797% in maximum deviation. Following Kirkpatrick, the majority of the court chose the plan with the smallest deviation, one that was a “Democratic plan” designed to maximize Democratic voting strength at the expense of Republicans. See In re Illinois Congressional Districts Reapportionment Cases, No. 81-C-3915 (ND Ill. 1981), summarily aff’d sub nom. Ryan v. Otto, 454 U. S. 1130 (1982). A commentator noted:
“The Democratic victory was due in part to a sophisticated computer program that made possible the creation of districts having almost exactly equal population. The most populous district has only 171 more people than the least populous one. That accuracy seemed to impress the court, which expressed no concern that the new district lines divided cities and carved up counties all over the state.” Illinois Map is Unpleasant Surprise for the GOP, 40 Congressional Quarterly 573 (1982).
See also Carstens v. Lamm, 543 F. Supp. 68, 73-74, and 84, n. 39 (Colo. 1982) (three-judge District Court reviewed five major redistricting plans, *786including the Republican legislature’s plan with a difference between largest and smallest districts of seven persons, i. e., a maximum deviation of 0.0015%, and the Democratic Governor’s plan with a 15-person difference, i. e., a maximum deviation of 0.0031%); O’Sullivan v. Brier, 540 F. Supp. 1200, 1202 (Kan. 1982) (three-judge District Court asked to choose between a Democratic plan with a 0.11% maximum deviation and a Republican plan with a 0.09% maximum deviation).
These cases also illustrate an additional unfortunate side effect of Kirkpatrick: the increasing tendency of state legislators and Governors— who have learned that any redistricting plan is “vulnerable to after-the-fact attack by anyone with a complaint and a calculator,” ante, at 778 (White, J., dissenting) — to spurn compromise in favor of simply drawing up the most partisan plan that appears consistent with the population equality criterion. No longer do federal district courts merely review the constitutionality of a State’s redistricting plan. Rather, in many cases they are placed in the position of choosing a redistricting plan in the first instance.
In Carstens v. Lamm, supra, the three-judge District Court noted that preserving an entire city as one voting district facilitated “voter identity”: “Most voters know what city and county they live in, but fewer are likely to know what congressional district they live in if the districts split counties and cities. If a voter knows his congressional district, he is more likely to know who his representative is. This presumably would lead to more informed voting.” 543 F. Supp., at 98, n. 78. It also is likely to lead to a Representative who knows the needs of his district and is more responsive to them.
It may be noted, for example, that the plan adopted by New Jersey (the Feldman Plan) divided the State’s 21 counties into 55 fragments. The plan proposed by Professor Reock, introduced by Assemblyman Hardwick, created 45 county fragments, and the existing congressional districts divided the counties into 42 fragments. See App. 123 (Appendix A to Affidavit of Samuel A. Alito, Executive Director of the Office of Legislative Services of the New Jersey Legislature).