I join Parts I and II of the Court’s opinion and concur in its judgment. I do not understand the Court to disapprove the District Court’s decision to use county lines as districting boundaries wherever possible, even though this policy may cause a greater variation in district population than would otherwise be appropriate for a court-ordered plan. The final plan adopted in this case appears to produce even greater population disparities than necessary to effectuate the county boundary policy. Cf. Mahan v. Howell, 410 U. S. 315, 326 (1973). This being so, the District Court should have articulated precise reasons for not adopting a more evenly apportioned plan. Chapman v. Meier, 420 U. S. 1, 27 (1975).
The appeals by the private parties and the United States in this case, however, were not primarily concerned with equal-population apportionment. Their more serious objections involved aspects of the District Court’s plan that were claimed to dilute Negro voting power.1 The two issues are quite *427distinct: Equal apportionment is a majoritarian principle, but racial representation is a question of minority rights. See Smith, The Failure of Reapportionment: The Effect of Reapportionment on the Election of Blacks to Legislative Bodies, 18 How, L. J. 639 (1975). I think the Court’s opinion does not sufficiently focus upon the potential dissonance between the one-person, one-vote ideal and a goal of fair representation for minorities.
The Court does not decide the racial dilution issue at this time, but the observations in Part III of its opinion indicate an approach that I think is not entirely appropriate. Details of districting are interrelated, and it is not helpful to look at isolated aspects of a statewide apportionment plan in order to determine whether a racial or other improperly motivated gerrymander has taken place. Districts that disfavor a minority group in one part of the State may be counterbalanced by favorable districts elsewhere. A better approach, therefore, is to examine the overall effect of the apportionment plan on the opportunity for fair representation of minority voters.
Statistics from the 1970 census reveal that the black voting-age population of Mississippi is 31.4%. Brief for United States 44 n. 40. Under the District Court’s apportionment plan, nine of the 52 Senate districts (17.3%) and 24 of the 122 House districts (19.7%) have black majorities of the voting-age population. Id., at 66. These statistics indicate that the plan would be unlikely to provide black voters with representation in the legislature equivalent to their electoral strength.2 But I do not think that the plan improperly dilutes *428black voting strength just because it fails to provide propor-, tional. representation. See Whitcomb v. Chavis, 403 U. S. 124, 149-155 (1971).
The normal system of legislative apportionment in the United States is direct territorial representation by single-member districts. Such system does not normally provide electoral minorities with proportional representation in the legislature. The extent to which electoral strength is translated into legislative representation depends on a number of factors, including (1) the size of the voting group, (2) its geographical dispersion, (3) the size of the legislative districts, and (4) the way district boundaries are drawn.3 The first three factors are probably sufficient to explain the result in the present case without raising an inference that the district boundaries were drawn so as further to minimize or dilute overall black voting strength.
Of course, the fact that a plan seems generally to provide fair representation would not preclude a showing that a particular aspect was adopted with an impermissibly discriminatory intent. But where the only claim is based on disparate effect, then piecemeal review of an apportionment plan may well be misleading. For example, the Court’s opinion suggests that the District Court may have erred in not adopting an alternative plan combining Jefferson and Claiborne Counties into a single Senate district (with Copiah County). Ante, at 424-425. But the District Court’s plan does combine Jefferson and Claiborne Counties into a single House district (number 81), with a 70% black majority of the voting-*429age population. Moreover, there is no reason to believe that the alternative Senate districting would have entailed less fragmentation of county boundaries in the overall plan. The alternative proposal would have required the formation of an additional Senate district starting with three noncontiguous areas — Simpson County, Lincoln County, and part of Adams County. A complete reshuffle of the Senate districts in southwestern Mississippi thus would be necessary to implement the alternative. One can only speculate on the effect of such a reshuffle with respect to either county boundary integrity or overall black voter representation.
The Court’s opinion also suggests that adherence to the criteria of contiguity and compactness would assure neutral districting. Ante, at 425-426. These normally are desirable characteristics of a districting plan, but I doubt that such an approach will be very effective in assuring fair representation for racial or other minority groups.4
A better constraint on potential gerrymandering is imposed by the use of established political boundaries. It is at this point that the goals of equal apportionment and minority representation may well conflict. To the extent that the attainment of precisely equal districts requires abandonment of longstanding political boundaries, gerrymandering is that much easier.5 Conversely, the requirement of equal apportionment *430places very little constraint on the possibility of a gerrymander, as the Court’s discussion of the Hinds County Senate districts illustrates. Ante, at 423-424. Those districts are almost exactly equal in population, with variances from the norm ranging only from +0.3% to + 1.3%.
None of my preceding comments are meant to suggest that intentional gerrymandering is a serious problem with court-ordered apportionment plans. But even a plan adopted with the purest of motives will have an unavoidable effect on the representation of various political groups in the legislature. Where there is an established policy of respecting political or natural boundaries in districting, then I believe that a court may best avoid any appearance of partisanship by using those boundaries as much as possible in its districting.
In fact, several of the districting alternatives proposed by these appellants as a means of improving black representation also would have involved greater population disparities than the plan adopted by the District Court. See, e. g., Brief for United States 49a (Hinds County *427Senate districts); id., at 55a (Warren County House districts); Brief for Private Appellants 45-46 (Adams County House districts).
The raciahdilution challenge in this case is predicated on the common but questionable assumption that voting will take place along racial lines, and thus that blacks receive effective representation only in districts where they compose a majority of the voting-age population. See Brief for Private Appellants 28-36; Brief for United States 33-59. Such an as*428sumption perhaps would be appropriate in situations where blacks continue to be excluded from the political process. See White v. Regester, 412 U. S. 755, 765-770 (1973). Separate representation by race, however, is certainly not an optimal solution and at best can provide only a temporary, expedient remedy.
See generally D. Rae, The Political Consequences of Electoral Laws (1967); Tufte, The Relationship between Seats and Votes in Two-Party Systems, 67 Am. Pol. Sci. Rev. 540 (1973).
It is not clear that workable standards of evaluating compactness are available, and in any event a requirement of compactness would not necessarily promote minority group representation. See R. Dixon, Democratic Representation 460-461 (1968); Mayhew, Congressional Representation: Theory and Practice in Drawing the Districts, in N. Polsby, ed., Reapportionment in the 1970s, pp. 253-255 (1971).
Reynolds v. Sims, 377 U. S. 533, 578-579 (1964); Wells v. Rockefeller, 394 U. S. 542, 551-552 (1969) (Harlan, J., dissenting); id., at 554-555 (White, J., dissenting). See Baker, Gerrymandering: Privileged Sanctuary or Next Judicial Target?, in N. Polsby, ed., Reapportionment in the 1970s, pp. 137-138 (1971); Elliott, The Political Consequences of Reapportionment, 37 U. Chi. L. Rev. 474, 481-490 (1970).