Justice Souter, with whom Justice Ginsburg joins, dissenting.

The Constitution guarantees both formal and substantial equality among voters. For 40 years, we have recognized that lines dividing a State into voting districts must produce divisions with equal populations: one person, one vote. Reynolds v. Sims, 377 U. S. 533, 568 (1964). Otherwise, a vote in a less populous district than others carries more clout.

Creating unequally populous districts is not, however, the only way to skew political results by setting district lines. The choice to draw a district line one way, not another always carries some consequence for politics, save in a mythical State with voters of every political identity distributed in an absolutely gray uniformity. The spectrum of opportunity runs from cracking a group into impotent fractions, to packing its members into one district for the sake of marginalizing them in another. However equal districts may be in population as a formal matter, the consequence of a vote cast can be minimized or maximized, Karcher v. Daggett, 462 U. S. 725, 734, n. 6 (1983), and if unfairness is sufficiently demonstrable, the guarantee of equal protection condemns it as a denial of substantial equality. Davis v. Bandemer, 478 U. S. 109, 129-134 (1986) (plurality opinion).

I

The notion of fairness assumed to be denied in these cases has been described as "each political group in a State [having] the same chance to elect representatives of its choice as any other political group," id., at 124, and as a "right to 'fair and effective representation,' " id., at 162 (Powell, J., concurring in part and dissenting in part). Cf. Wells v. Rockefeller, 394 U. S. 542, 551 (1969) (Harlan, J., dissenting) (describing the need for "a structure which will in fact as well as theory be responsive to the sentiments of the community"). It is undeniable that political sophisticates understand such fairness and how to go about destroying it, see App. to Juris. Statement 134a, although it cannot possibly be described with the hard edge of one person, one vote. The difficulty has been to translate these notions of fairness into workable criteria, as distinct from mere opportunities for reviewing courts to make episodic judgments that things have gone too far, the sources of difficulty being in the facts that some intent to gain political advantage is inescapable whenever political bodies devise a district plan, and some effect results from the intent. Wells, supra, at 554-555 (White, J., dissenting) ("In reality, of course, districting is itself a gerrymandering in the sense that it represents a complex blend of political, economic, regional, and historical considerations"). Thus, the issue is one of how much is too much, and we can be no more exact in stating a verbal test for too much partisanship than we can be in defining too much race consciousness when some is inevitable and legitimate. See Bush v. Vera, 517 U. S. 952, 1057-1062 (1996) (Souter, J., dissenting). Instead of coming up with a verbal formula for too much, then, the Court's job must be to identify clues, as objective as we can make them, indicating that partisan competition has reached an extremity of unfairness.

The plurality says, in effect, that courts have been trying to devise practical criteria for political gerrymandering for nearly 20 years, without being any closer to something workable than we were when Davis was decided. Ante, at 11.1 While this is true enough, I do not accept it as sound counsel of despair. For I take it that the principal reason we have not gone from theoretical justiciability to practical administrability in political gerrymandering cases is the Davis plurality's specification that any criterion of forbidden gerrymandering must require a showing that members of the plaintiff's group had "essentially been shut out of the political process," 478 U. S., at 139. See, e.g., Badham v. Eu, 694 F. Supp. 664, 670-671 (ND Cal. 1988) (three-judge court). That is, in order to avoid a threshold for relief so low that almost any electoral defeat (let alone failure to achieve proportionate results) would support a gerrymandering claim, the Davis plurality required a demonstration of such pervasive devaluation over such a period of time as to raise real doubt that a case could ever be made out. Davis suggested that plaintiffs might need to show even that their efforts to deliberate, register, and vote had been impeded. 478 U. S., at 133. This standard, which it is difficult to imagine a major party meeting, combined a very demanding burden with significant vagueness; and if appellants have not been able to propose a practical test for a Davis violation, the fault belongs less to them than to our predecessors. As Judge Higginbotham recently put it, "[i]t is now painfully clear that Justice Powell's concern that [Davis] offered a ' "constitutional green light" to would-be gerrymanderers' has been realized." Session v. Perry, 298 F. Supp. 2d 451, 474 (ED Tex. 2004) (quoting Davis, supra, at 173 (Powell, J., concurring in part and dissenting in part)).

II

Since this Court has created the problem no one else has been able to solve, it is up to us to make a fresh start. There are a good many voices saying it is high time that we did, for in the years since Davis, the increasing efficiency of partisan redistricting has damaged the democratic process to a degree that our predecessors only began to imagine. E.g., Issacharoff, Gerrymandering and Political Cartels, 116 Harv. L. Rev. 593, 624 (2002) (The "pattern of incumbent entrenchment has gotten worse as the computer technology for more exquisite gerrymandering has improved"); Karlan, The Fire Next Time: Reapportionment After the 2000 Census, 50 Stan. L. Rev. 731, 736 (1998) ("Finer-grained census data, better predictive methods, and more powerful computers allow for increasingly sophisticated equipopulous gerrymanders"); Pildes, Principled Limitations on Racial and Partisan Restricting, 106 Yale L. J. 2505, 2553-2554 (1997) ("Recent cases now document in microscopic detail the astonishing precision with which redistricters can carve up individual precincts and distribute them between districts with confidence concerning the racial and partisan consequences"). See also Morrill, A Geographer's Perspective, in Political Gerrymandering and the Courts 213-214 (B. Grofman ed. 1990) (noting that gerrymandering can produce "high proportions of very safe seats"); Brief for Bernard Grofman et al. as Amici Curiae 5-8 (decline of competitive seats). Cf. Wells, supra, at 551 (Harlan, J., dissenting) ("A computer may grind out district lines which can totally frustrate the popular will on an overwhelming number of critical issues").

I would therefore preserve Davis's holding that political gerrymandering is a justiciable issue, but otherwise start anew. I would adopt a political gerrymandering test analogous to the summary judgment standard crafted in McDonnell Douglas Corp. v. Green, 411 U. S. 792 (1973), calling for a plaintiff to satisfy elements of a prima facie cause of action, at which point the State would have the opportunity not only to rebut the evidence supporting the plaintiff's case, but to offer an affirmative justification for the districting choices, even assuming the proof of the plaintiff's allegations. My own judgment is that we would have better luck at devising a workable prima facie case if we concentrated as much as possible on suspect characteristics of individual districts instead of state-wide patterns. It is not that a statewide view of districting is somehow less important; the usual point of gerrymandering, after all, is to control the greatest number of seats overall. But, as will be seen, we would be able to call more readily on some existing law when we defined what is suspect at the district level, and for now I would conceive of a statewide challenge as itself a function of claims that individual districts are illegitimately drawn. Finally, in the same interest of threshold simplicity, I would stick to problems of single-member districts; if we could not devise a workable scheme for dealing with claims about these, we would have to forget the complications posed by multimember districts.

III

A

For a claim based on a specific single-member district, I would require the plaintiff to make out a prima facie case with five elements. First, the resident plaintiff would identify a cohesive political group to which he belonged, which would normally be a major party, as in this case and in Davis. There is no reason in principle, however, to rule out a claimant from a minor political party (which might, if it showed strength, become the target of vigorous hostility from one or both major parties in a State) or from a different but politically coherent group whose members engaged in bloc voting, as a large labor union might do. The point is that it must make sense to speak of a candidate of the group's choice, easy to do in the case of a large or small political party, though more difficult when the organization is not defined by politics as such.2

Second, a plaintiff would need to show that the district of his residence, see United States v. Hays, 515 U. S. 737 (1995) (requiring residence in a challenged district for standing), paid little or no heed to those traditional districting principles whose disregard can be shown straightforwardly: contiguity, compactness, respect for political subdivisions, and conformity with geographic features like rivers and mountains. Because such considerations are already relevant to justifying small deviations from absolute population equality, Karcher, 462 U. S., at 740, and because compactness in particular is relevant to demonstrating possible majority-minority districts under the Voting Rights Act of 1965, Johnson v. De Grandy, 512 U. S. 997, 1008 (1994), there is no doubt that a test relying on these standards would fall within judicial competence.

Indeed, although compactness is at first blush the least likely of these principles to yield precision, it can be measured quantitatively in terms of dispersion, perimeter, and population ratios, and the development of standards would thus be possible. See generally Pildes & Niemi, Expressive Harms, "Bizarre Districts," and Voting Rights: Evaluating Election-District Appearances After Shaw v. Reno, 92 Mich. L. Rev. 483 (1993); see also Bush v. Vera, 517 U. S., at 1057 (Souter, J., dissenting) (suggesting that such measuring formulas might have been applied to salvage Shaw v. Reno, 509 U. S. 630 (1993)).3 It is not necessary now to say exactly how a district court would balance a good showing on one of these indices against a poor showing on another, for that sort of detail is best worked out case by case.

Third, the plaintiff would need to establish specific correlations between the district's deviations from traditional districting principles and the distribution of the population of his group. For example, one of the districts to which appellants object most strongly in this case is District 6, which they say "looms like a dragon descending on Philadelphia from the west, splitting up towns and communities throughout Montgomery and Berks Counties." App. to Juris. Statement 136a. To make their claim stick, they would need to point to specific protuberances on the draconian shape that reach out to include Democrats, or fissures in it that squirm away from Republicans. They would need to show that when towns and communities were split, Democrats tended to fall on one side and Republicans on the other. Although some counterexamples would no doubt be present in any complex plan, the plaintiff's showing as a whole would need to provide reasonable support for, if not compel, an inference that the district took the shape it did because of the distribution of the plaintiff's group. That would begin, but not complete, the plaintiff's case that the defendant had chosen either to pack the group (drawn a district in order to include a uselessly high number of the group) or to crack it (drawn it so as to include fatally few), the ordinary methods of vote dilution in single-member district systems. Ante, at 17, n. 7.

Fourth, a plaintiff would need to present the court with a hypothetical district including his residence, one in which the proportion of the plaintiff's group was lower (in a packing claim) or higher (in a cracking one) and which at the same time deviated less from traditional districting principles than the actual district. Cf. Thornburg v. Gingles, 478 U. S. 30, 50 (1986) (requiring a similar showing to demonstrate that a multimember district is "responsible for minority voters' inability to elect [their preferred] candidates"). This hypothetical district would allow the plaintiff to claim credibly that the deviations from traditional districting principles were not only correlated with, but also caused by, the packing or cracking of his group. Drawing the hypothetical district would, of course, necessarily involve redrawing at least one contiguous district,4 and a plaintiff would have to show that this could be done subject to traditional districting principles without packing or cracking his group (or another) worse than in the district being challenged.

Fifth, and finally, the plaintiff would have to show that the defendants acted intentionally to manipulate the shape of the district in order to pack or crack his group. See Washington v. Davis, 426 U. S. 229 (1976). In substantiating claims of political gerrymandering under a plan devised by a single major party, proving intent should not be hard, once the third and fourth (correlation and cause) elements are established, politicians not being politically disinterested or characteristically naļ ;ve. Davis v. Bandemer, 478 U. S., at 128 ("[W]e think it most likely that whenever a legislature redistricts, those responsible for the legislation will know the likely political composition of the new districts"). I would, however, treat any showing of intent in a major-party case as too equivocal to count unless the entire legislature were controlled by the governor's party (or the dominant legislative party were vetoproof).5

If the affected group were not a major party, proof of intent could, admittedly, be difficult. It would be possible that a legislature might not even have had the plaintiff's group in mind, and a plaintiff would naturally have a hard time showing requisite intent behind a plan produced by a bipartisan commission.

B

A plaintiff who got this far would have shown that his State intentionally acted to dilute his vote, having ignored reasonable alternatives consistent with traditional districting principles. I would then shift the burden to the defendants to justify their decision by reference to objectives other than naked partisan advantage. They might show by rebuttal evidence that districting objectives could not be served by the plaintiff's hypothetical district better than by the district as drawn, or they might affirmatively establish legitimate objectives better served by the lines drawn than by the plaintiff's hypothetical.

The State might, for example, posit the need to avoid racial vote dilution. Cf. Vera, 517 U. S., at 990 (O'Connor, J., concurring) (compliance with §2 of the Voting Rights Act of 1965 is a compelling state interest). It might plead one person, one vote, a standard compatible with gerrymandering but in some places perhaps unattainable without some lopsided proportions. The State might adopt the object of proportional representation among its political parties through its districting process. Gaffney v. Cummings, 412 U. S. 735, 754 (1973); 6 cf. Johnson v. De Grandy, 512 U. S., at 1024 (totality of the circumstances did not support finding of vote dilution where "minority groups constitute[d] effective voting majorities in a number of state Senate districts substantially proportional to their share in the population").7

This is not, however, the time or place for a comprehensive list of legitimate objectives a State might present. The point here is simply that the Constitution should not petrify traditional districting objectives as exclusive, and it is enough to say that the State would be required to explain itself, to demonstrate that whatever reasons it gave were more than a mere pretext for an old-fashioned gerrymander.

C

As for a statewide claim, I would not attempt an ambitious definition without the benefit of experience with individual district claims, and for now I would limit consideration of a statewide claim to one built upon a number of district-specific ones. Each successful district-specific challenge would necessarily entail redrawing at least one contiguous district, and the more the successful claims, the more surrounding districts to be redefined. At a certain point, the ripples would reach the state boundary, and it would no longer make any sense for a district court to consider the problems piecemeal.

D

The plurality says that my proposed standard would not solve the essential problem of unworkability. It says that "[i]t does not solve the problem [of determining when gerrymandering has gone too far] to break down the original unanswerable question . . . into four more discrete but unanswerable questions." Ante, at 27-28. It is common sense, however, to break down a large and intractable issue into discrete fragments as a way to get a handle on the larger one, and the elements I propose are not only tractable in theory, but the very subjects that judges already deal with in practice. The plurality asks, for example, "[w]hat ... a lower court [is] to do when, as will often be the case, the district adheres to some traditional criteria but not others?" Ibid. This question already arises in cases under §2 of the Voting Rights Act of 1965, and the district courts have not had the same sort of difficulty answering it as they have in applying the Davis v. Bandemer plurality. See, e.g., Johnson v. Hamrick, 155 F. Supp. 2d 1355, 1362-1363 (ND Ga. 2001) (noncontiguity of a plaintiff's Gingles districts was not fatal to a §2 claim against a municipal districting scheme because "the city's boundaries are rough and asymmetrical . . . [and] the non-contiguous portions [of the proposed districts] are separated by unincorporated areas and are relatively near the districts to which they are joined"). The enquiries I am proposing are not, to be sure, as hard-edged as I wish they could be, but neither do they have a degree of subjectivity inconsistent with the judicial function.

The plurality also says that my standard is destined to fail because I have not given a precise enough account of the extreme unfairness I would prevent. Ante, at 28-30. But this objection is more the reliable expression of the plurality's own discouragement than the description of an Achilles heel in my suggestion. The harm from partisan gerrymandering is (as I have said, supra, at 1-2, 8, 10) a species of vote dilution: the point of the gerrymander is to capture seats by manipulating district lines to diminish the weight of the other party's votes in elections. To devise a judicial remedy for that harm, however, it is not necessary to adopt a full-blown theory of fairness, furnishing a precise measure of harm caused by divergence from the ideal in each case. It is sufficient instead to agree that gerrymandering is, indeed, unfair, as the plurality does not dispute; to observe the traditional methods of the gerrymanderer, which the plurality summarizes, ante, at 4-6; and to adopt a test aimed at detecting and preventing the use of those methods, which, I think, mine is. If those methods are unnecessary to effective gerrymandering, as the plurality implies, ante, at 28-29, it is hard to explain why they have been so popular down through the ages of American politics. My test would no doubt leave substantial room for a party in power to seek advantage through its control of the districting process; the only way to prevent all opportunism would be to remove districting wholly from legislative control, which I am not prepared to say the Constitution requires. But that does not make it impossible for courts to identify at least the worst cases of gerrymandering, and to provide a remedy. The most the plurality can show is that my approach would not catch them all. Cf. Scalia, The Rule of Law as a Law of Rules, 56 U. Chi. L. Rev. 1175, 1178 (1989) ("To achieve what is, from the standpoint of the substantive policies involved, the 'perfect' answer is nice--but it is just one of a number of competing values").

IV

In drafting the complaint for this case, appellants' counsel naturally proceeded on the assumption that they had to satisfy the Davis v. Bandemer plurality, or some revision in light of Shaw, but not the prima facie case I have in mind. Richard and Norma Jean Vieth make only statewide claims, for which the single district claim brought by Susan Furey provides insufficient grounding. As for Furey's own claim, her allegations fall short, for example, on the feasibility of an alternative district superior to her own, as I would require. But she might well be able to allege what I would require, if given leave to amend. I would grant her that leave, and therefore would vacate the judgment of the District Court and remand for further proceedings. From the Court's judgment denying her that opportunity, I respectfully dissent.