Plaintiffs, seven black residents of Leflore County, Mississippi, brought this action in 1971, challenging the decision of the county’s Board of Supervisors to hold at-large, rather than district, elections for Board positions. A three-judge court allowed the election to be held on an at-large basis but ordered the Board to develop a redistricting plan *623which would cure the county’s unconstitutional malapportionment.1 The three-judge court then dissolved and remanded the case to a single district judge. Moore v. Leflore County Board of Election Commissioners, 351 F.Supp. 848 (D.C. 1971).

On December 20, 1972, the district judge filed a memorandum opinion holding the plan developed and submitted by the Board (hereafter “Kellum Plan”) unconstitutional because both its purpose and effect “. . . [were] to divide the black population and dilute the black vote in Leflore County. . . .” Moore v. Leflore County Board of Election Commissioners, 361 F.Supp. 603, 607 (D.C. 1972). The court also appointed a special master, Hoyt T. Holland, Jr., to formulate an acceptable redistricting plan (hereafter “Holland Plan”). Finally, on June 4, 1973, the court held that the Holland Plan satisfied constitutional requirements and ordered it implemented immediately in district elections. Moore v. Leflore County Board of Election Commissioners, 361 F.Supp. 609 (D.C. 1973). A majority of the Board appealed, and plaintiffs cross appealed. A two member minority of the Board are appellees here. We affirm the order of the district court.

As is the case in many areas of the South, the population of Leflore County is concentrated in and around a single urban area, here the city of Greenwood; the rest of the county is largely undeveloped, sparsely populated, farming country. This uneven population distribution is a complicating factor in any reapportionment equation, as we shall see. Further complications arise from the racial composition of the population. Of the county’s 1970 population (41,923), 58% (24,373) is black and 42% (17,550) is white. Greenwood’s population, however, is divided almost equally between black (11,130) and white (11,118). For reapportionment purposes, the most significant statistic is that 10,763 of Greenwood’s blacks live in southeast Greenwood. Thus, close to half of the county’s blacks, one-fourth of its entire population, is concentrated in a five square mile quadrant of the county seat ; the heart of this controversy is located within those five square miles.

The Kellum Plan, which appellant supervisors contend should have been accepted by the court, divided the county into five districts or “beats” of practically equal population.2 No party has contended that this plan does not satisfy the arithmetical aspect of the one man-one vote standard established by Reynolds v. Sims, 377 U.S. 533, 84 S.Ct. 1362, 12 L.Ed.2d 506 (1964); the .33% variance between largest and smallest districts is well within the judicially established guidelines for meeting that test.3 Rather, the court rejected the plan because it diluted black voting strength and because it failed to take into consideration legitimate planning objectives, such as equality of road mileage and land area. In this determination, we think the court was clearly correct.

A reapportionment plan is unconstitutional if it is a racially motivated gerrymander or if it is a plan drawn along racial lines which, “. . . designedly or otherwise . . . would operate to minimize or cancel out the *624voting strength of racial or political elements of the voting population.” Fortson v. Dorsey, 379 U.S. 433, 439, 85 S.Ct. 498, 501, 13 L.Ed.2d 401 (1965). If he cannot prove a racial gerrymander, a plaintiff must show

. . . that the political processes leading to nomination and election [would not be] equally open to participation by the group in question — that its members [would have] less opportunity than [would] other residents in the district to participate in the political process and to elect legislators of their choice. White v. Regester, 412 U.S. 755, 93 S.Ct. 2332, 2339, 37 L.Ed.2d 314, 324 (1973) (citation omitted).

As a three-judge court recently said in connection with a New Orleans, Louisiana, redistricting proposal:

The measure of the plan’s validity is equality of opportunity, and the crucial inquiry is whether the plan leaves black citizens at liberty to participate in the electoral process on the same plane with white citizens. Beer v. United States, 374 F.Supp. 363, 384 (D.C. 1974) (footnote omitted).

Under the Kellum Plan, the majority of each district’s residents are black. The extent of each majority, however, is diluted in all but one of the districts when compared to pre-redistricting figures.4 Significantly, it also appears that in terms of registered voters, blacks would have exceedingly slim majorities in some of these districts and minorities in others.5

The mere existence of a black population majority does not preclude a finding of dilution. Zimmer v. McKeithen, 485 F.2d 1297 (5th Cir. 1973) (en banc). As we said in Zimmer, p. 1305, with reference to a black majority, in East Carroll Parish, Louisiana, if a group of voters can show

. a lack of access to the process of slating candidates, the unresponsiveness of legislators to their particularized interests, a tenuous, state policy underlying the preference for multi-member or at-large districting, or that the existence of past discrimination in general precludes the effective participation in the election system, a strong case [supporting a finding of dilution] is made.

We agree with the district judge that such a showing was made here. Prior to the passage of the 1965 Voting Rights Act, most of Leflore County’s blacks were prevented from registering to vote by a variety of discriminatory, unconstitutional, state law devices. See United States v. Mississippi, 380 U.S. 128, 85 S.Ct. 808, 13 L.Ed.2d 717 (1965). As the district court found, against this “. . . background of fear and civil rights repression, blacks have minimally engaged in political activity. . . .” Moore v. Leflore County Board of Election Commissioners, 361 F.Supp. 603, 605 (D.C. 1972). No black has ever been elected to public office in the county; a black’s near victory over Supervisor Kellum in 1966 immediately preceded, the Board’s decision to switch to at-large elections. No black has since run for the office. The point is that against this background, the Kellum Plan diluted the black vote; by retaining the barest of black, population majorities, it enhanced the possibility of continued black political impotence.

The district court found several other infirmities in the plan. In drawing district lines, Kellum completely ignored *625land area and road mileage as planning criteria, and drew the boundaries in such a way as to avoid pitting incumbent supervisors against each other in future elections. Although none of these errors may be fatal in themselves, the district court was correct in considering them additional factors requiring rejection of the plan. Equalization of land area and road mileage is extremely important here; although each district is allotted the same amount of public funds for road and bridge maintenance, under the Kellum Plan there are gross discrepancies among the facilities to be maintained by those funds. District Three, for example, would encompass only 3.9% of the county and would contain only 4.1% of its roads, while District One would cover 33.2% of the area and would contain 29.6% of the roads. As we said in Howard v. Adams County Board of Supervisors, 453 F.2d 455, 456 (5th Cir. 1972), the equalization of road mileage and land area in such a situation are “legitimate planning objectives.”

We turn now to the Holland Plan, which both appellant supervisors and plaintiffs contend is unconstitutional, although for stunningly different reasons: the supervisors contend it discriminates against whites while plaintiffs make the identical argument as to blacks. We hold the plan constitutional because it neither discriminates against either race nor dilutes its voting strength.

The racial distribution of the entire population under the plan is:

DIST. NO. WHITE % BLACK % TOTAL

1 6,331 75 2,139 25 8,470

2 3,306 39 5,186 61 8,492

3 2,742 33 5,642 67 8,384

4 3,043 36 5,372 64 8,415

5 2,128 25 6,222 75 8,350

17,550 42 24,561 58 42,111

Thus, in the four districts with black majorities, the extent of the majorities is far greater than under the Kellum Plan. Black voters are therefore far more likely to be able to exercise their franchise in a full and meaningful way.

The racial distribution of registered voters under the plan is:

DIST. NO. WHITE % BLACK % TOTAL

1 4,170 88 569 12 4,739

2 1,779 49 1,833 51 3,612

3 1,551 42 2,156 58 3,707

4 1,713 50 1,736 50 3,449

5 1,172 34 2,276 66 3,448

10,385 55 8,570 45 18,955

Plaintiffs argue that such a distribution, leaving blacks with a majority in only three districts, constitutes a dilution of their voting strength. Such a discrepancy between the number of black citizens and black registered voters is only to be expected, however, given the history of discrimination and re*626pression outlined above. The point is, no realignment of districts could eliminate this discrepancy; blacks have simply not registered in numbers commensurate with their population strength. Hopefully, blacks will now and in the future register in numbers approaching their full potential. If they do, the Holland Plan allows them full access to the reigns of government. The racial distribution of the voting age population under the plan is:

DIST. NO. WHITE %_ BLACK %_ TOTAL

1 4,278 81 994 19 5,272

2 2,191 45 2,667 55 4,858

3 1,930 38 3,099 62 5,029

4 2,019 41 2,927 59 4,946

5 1,399 30 3,298 70 4,681

11,817 48 12,969 52 24.786

Thus, blacks have substantial voting age majorities in four of the five districts. Although this fact alone does not negate the possibility of dilution, we are convinced that, under the totality of circumstances here dilution will not result.

Shifting the focus of their attack from statistics to geography, plaintiffs contend that dilution results from the drawing of the district lines in such a way as to preserve a “safe” white district in north Greenwood while segmenting the black population of south Greenwood into four separate districts. We reject this contention for two reasons. First, any district of approximately 8,500 persons encompassing north Greenwood would have a white majority since the area’s residents are almost exclusively white. It is eminently logical to structure north Greenwood as a separate district since it is separated from the rest of the town on the south and the east by the Yazoo River. Acknowledging this, plaintiffs nevertheless contend it was unconstitutional for the master to draw each of the districts with long corridors reaching into and segmenting Greenwood. As to this argument, we reaffirm what we said in Howard v. Adams County Board of Supervisors, 453 F.2d 455, 456 (5th Cir. 1972). That case involved a redistricting plan for a county strikingly similar, geographically and demographically, to Lef-lore County: Natchez, with 64% of the county’s population, is, like Greenwood, located near a county line. The racial distribution of the county is 48% black, 52% white. In drawing his plan the special master was told to equate the population, road mileage, and land area of the districts as nearly as possible. “With its land area largely rural, and its population concentrated in one urban area” we said,

[t]he realization of these legitimate planning objectives dictated a plan which would consolidate urban and rural areas into each district. . therefore, each district converged in spokelike fashion from a broad rural base into the city of Natchez.

A more accurate description of the Lef-lore County situation can hardly be imagined.

The Holland Plan withstands appellant supervisors’ attack as well. As noted above, the concentration of whites in District One is a natural product of the geography and the residential patterns *627of Greenwood. It is not the result of a racial gerrymander and must stand since no claim or showing of white dilution has been made.

Finally, the district court was clearly correct in ordering district elections rather than the at-large election for which appellant supervisors contend. Single-member districts are, of course, preferable to large, multi-member districts. Connor v. Johnson, 402 U.S. 690, 91 S.Ct. 1760, 29 L.Ed.2d 268 (1971). A district court may override this preference only if it “. . . determines that significant interests would be advanced by the use of multi-member districts and the use of single-member districts would jeopardize constitutional requirements . . . [or] multi-mem-ber districts afford minorities a greater opportunity for participation in the political processes than do single-member districts.” Zimmer, supra, 485 F.2d p. 1308. The record amply supports the district court’s conclusion that an at-large scheme, far from meeting this test, would rather diffuse and dilute black voting strength and work to deny Leflore County blacks equal access to the political process. As noted above, far fewer blacks are of voting age or are registered to vote than whites. When this fact is added to the “background of fear and civil rights [suppression in the county]”6 the majority vote requirement for democratic primary nomination, the anti-single shot voting provision of Mississippi law, and the concentration of black residents in south Greenwood, there is ample support for the conclusion that at-large districting far from being required here, would, in fact, be constitutionally infirm. Cf. White v. Regester, supra, 412 U.S. at 766, 93 S.Ct. at 2339, 37 L.Ed.2d at 324.

For the above reasons the judgment of the district court is

Affirmed.