Full opinion text
MEMORANDUM OPINION AND ORDER BENSON, Chief District Judge. On May 18, 1972, this three judge district court panel heard evidence and oral argument on the constitutionality of the North Dakota legislative apportionment plan then in effect. On May 22, 1972, we entered an order requiring reapportionment of the state legislative districts to conform to the “one man, one vote” requirement of the Equal Protection Clause of the Fourteenth Amendment of the United States Constitution. Thereafter, on June 30, 1972, this Court filed its Memorandum Opinion and Order, Judge Benson dissenting in part, adopting an interim apportionment plan effective only for the impending 1972 general election. The plan provided for thirty-nine senatorial districts, five of which were multi-member. It increased the size of the state senate by two and increased the state house of representatives by four. It decreased the number of legislative districts from thirty-nine to thirty-eight. The plan recognized that the interests of those persons residing on the air bases at Grand Forks and Minot were more closely aligned with urban than with rural interests, and included those populations within the nearby urban districts. The multi-member districts retained are located in the cities of Fargo, Grand Forks, Minot, Bismarck and Jamestown. The Court retained jurisdiction over the cause, and directed its commission of three special masters to study and report upon a more permanent plan. Subsequently, on motion by Defendant Meier, the Court, on November 8, 1972, ordered that further action be deferred pending the possible adoption of a new apportionment plan by the 43rd Legislative Assembly of the State of North Dakota, in its 1973 session. Over the veto of the Governor, the Legislature adopted an apportionment plan which continued the multi-member districts substantially as provided in the Court’s plan, and as had existed in the state since 1965. The Governor’s principal objection to the Legislature’s plan centered on the multi-member senate districts. The operation of the plan adopted by the Legislature was suspended by referendum petition. By initiative petition, an amendment to the Constitution of North Dakota was proposed which would create a commission to reapportion the state and which would mandate the creation of single member senatorial districts. A statewide special election on the referred plan of the Legislature and the initiated constitutional amendment was held on December 4, 1973. Both measures were defeated. Therefore, the obligation to make a final determination on a reapportionment plan for the legislative districts of the State of North Dakota remains with this Court. The plaintiffs urge the Court to proceed in accordance with its Memorandum Opinion and Order of June 29, 1972, wherein the majority suggested that if a new, more permanent, plan had to be fashioned by the Court, it would probably establish single member districts in light of Connor v. Johnson, 402 U.S. 690, 91 S.Ct. 1760, 29 L.Ed.2d 268 (1971): “When district courts are forced to fashion apportionment plans, single-member districts are preferable to large multi-member districts as a general rule.” at 692, 91 S.Ct. at 1762. On the other hand, the Defendant argues that this Court is not compelled by the Connor decision to create single member districts and urges us, in light of the apportionment decisions of the United States Supreme Court that have been rendered since June 29, 1972, to adopt the June 29, 1972 Court plan (The Dobson Plan) as the permanent plan for the State of North Dakota. Evidence before this Court indicates that North Dakota is a sparsely settled, agricultural state with declining populations in most localities outside the urban areas. The state has fifty three counties. Most minor civil divisions in rural areas have very small populations that are becoming smaller. The 1970 Federal Census for North Dakota showed an overall state population of 617,761 persons, and a population density of 8.9 persons per square mile. The overall loss in population to North Dakota between 1960 and 1970 on a statewide basis was 2.3%. In 1960, the urban-rural population was divided 35.2% urban and 64.8% rural. In 1970, it was 55.7% urban and 44.3% rural. A total of 183 census county divisions are composed of mainly open country. Ten of these had more than 2,500 inhabitants in 1970, and thirty-nine (21.3%) had fewer than 1,000. The two smallest divisions had less than 500 inhabitants. The two largest census county divisions in this group had 1970 populations of 12,608 and 12,927, and contained the Grand Forks and Minot Air Force Bases. In hearings before the Joint Committee on Reapportionment held on January 3, 1973, State Representative Earl C. Rundle indicated that in Billings County there are 72 sections of land with no people residing on them. The total population of that county is 1,198. The population of the four largest cities in North Dakota in descending order are: Fargo, 53,365; Grand Forks, 39,008; Bismarck, 34,703; Minot, 32,290. Special Master Dobson, in presenting his plan, which with some minor amendments, was adopted by the Court on June 29, 1972, as an “interim” plan commented : “The plan observes natural geographical barriers, such as the Missouri River . . and . -. . every district is connected with good arterial roads. It should not be necessary to travel outside of one’s district in going from one part of it to another.” “In the formation of districts, parts of 10 counties are attached to an adjoining county or counties. However, in three instances (Burleigh, Ward and Williams), no real violence is done to county lines because an urban district is sealed off and the rural portion of the county is attached to a neighboring rural county. Three other counties which are divided (Barnes, Richland and Walsh) have traditionally been split into two districts. Thus, only four counties (McHenry, Cass, Morton and Stark) suffer any damage in the districting. In this connection, it should be noted that the greatest complaints about the existing apportionment were voiced over the breaking of county lines, particularly in smaller rural counties. Counties with very small populations should not be split because they are thereby rendered politically powerless. It is a form of de facto disenfranchisement.” In its Order of June 29, 1972, this Court found that reapportionment was required because of the general population shift from rural to urban centers in North Dakota which “created constitutionally impermissible variations in population among the existing legislative districts of North Dakota”. The following is a chart of the plan adopted by this Court and is inserted in this Opinion to illustrate that it cures the “constitutionally impermissible variations”, which were the basis for the court ordered reapportionment. North Dakota Population — 1970 Census 617,761 Number of Senators Provided for In Court Plan 51 Population per senator (absolute equality) 12,112 The majority of this Court has concluded that the interim plan should be adopted as the permanent plan. Judge Bright, in dissenting, is concerned with the fact that Districts 5, 18, 21, 29 and 32 are multi-member districts, and with the fact that the 11.43% overrepresentation in District 11, coupled with the 8.71% underrepresentation in District 4, creates what he feels to be an impermissible variation of 20.14%. I. MULTI-MEMBER DISTRICTS There has been no showing before this Court that multi-member districts, which have existed in North Dakota since 1965, have resulted in discrimination of any kind against any groups. Plaintiff’s counsel, in his brief filed with the Court on May 5, 1972, said': “Plaintiffs do not assume the burden of establishing that multi-member senate and house districts are violative of the United States Constitution”. Relying on Connor, the thrust of the contention appears to be that multi-member districts cannot be allowed to continue because they were initially fashioned by the Court. The situation in Mississippi, faced by the court in Connor, is obviously quite different from that which exists in North Dakota. According to the 1970 census, the State of Mississippi has a population of 2,216,912, and a population density of 49.6 persons per square mile. The white population is 1,393,293; the black population is 815,770. In North Dakota, except for the Indian Reservations, none of which are included in the multi-member senate districts, there are no identifiable minorities. Four days after its decision in Connor, the Supreme Court handed down Whitcomb v. Chavis, 403 U.S. 124, 91 S.Ct. 1858, 29 L.Ed.2d 363 (1971), wherein it reviewed the decision of a three judge court convened to consider whether two Indiana statutes had the effect of diluting the vote of Negroes and poor people living in Marion County (City of Indianapolis). The district court panel concluded that the existing multi-member district must be separated because of strong differences in minority groups, housing, income and educational levels. With respect to Marion County, the three judge panel drafted and adopted a plan said to protect the “legally cognizable racial minority group against dilution of its voting strength”, 307 F.Supp. 1362, at 1365 (S.D.Ind.1969). On appeal, the Supreme Court, in passing on the redistrieting of Marion County into single member districts, reversed, saying that while Reynolds v. Sims, 377 U.S. 533, 84 S.Ct. 1362, 12 L.Ed.2d 506 (1964), recognizes the right of every citizen to full and effective participation in the political processes of his state’s legislative bodies, that decision did not render multi-member districts impermissible. “In our view, however, experience and insight have not yet demonstrated that multi-member districts are inherently invidious and violative of the Fourteenth Amendment. Surely the findings of the District Court do not demonstrate it. Moreover, if the problems of multi-member districts are unbearable or even unconstitutional it is not at all clear that the remedy is a single-member district system with its lines carefully drawn to ensure representation to sizable racial, ethnic, economic, or religious groups and with its own capacity for overrepresenting parties and interests and even for permitting a minority of the voters to control the legislature and government of a State. The short of it is that we are unprepared to hold that district-based elections decided by plurality vote are unconstitutional in either single- or multi-member districts simply because the supporters of losing candidates have no legislative seats assigned to them. As presently advised we hold that the District Court misconceived the Equal Protection Clause in applying it to invalidate the Marion County multi-member district.” 403 U.S. at 159-160, 91 S.Ct. at 1877-1888. Citing particular problems which might render a multi-member plan ineffective, Whitcomb announced that the circumstances of each ease must be considered. In Fortson v. Dorsey, 379 U.S. 433, 85 S.Ct. 498, 13 L.Ed.2d 401 (1965), the Supreme Court specifically held that multi-member districts were not per se illegal under the Equal Protection Clause. In the 1973 case of Mahan v. Howell, 410 U.S. 315, 93 S.Ct. 979, 35 L.Ed.2d 320 (1973), the court noted that considerations of substantial malapportionment with respect to military personnel and combinations of other unique factors must be considered in preferring single-member over multi-member districts or vice versa. In the case of White v. Regester, 412 U.S. 755, 93 S.Ct. 2332, 37 L.Ed.2d 314 (1973), the Supreme Court was called upon to consider a Texas reapportionment plan. Again the Court stated that multi-member districts are not unconstitutional per se, (citing Whitcomb, Mahan and Sims) but where the claim is that such districts are being used invidiously to cancel out or minimize the voting strength of racial groups, they must be questioned. The Texas multi-member districts, it was found, did operate to exclude the black community from the electoral process. This case points up one of the basic foundations common to all eases finding multi-member districts unconstitutional; that is, invidious discrimination against some visible minority group. See also Hernandez v. Texas, 347 U.S. 475, 74 S.Ct. 667, 98 L.Ed. 866 (1954), involving Mexican-Americans. The group must be identifiable and must be effectively removed from the electoral process by reason of the districting scheme. As the Court in White notes, not every racial or political group has a constitutional right to be represented in the state legislature, 412 U.S. at 769, 93 S.Ct. at 2341, 37 L.Ed.2d at 326. Earlier in Whitcomb, the court commented that were every group entitled to representation, the result would be inane for within every district there are workers, university communities, religious and ethnic groups occupying identifiable areas. A court, in considering various reapportionment options before it, should go only as far as is necessary in order to meet the constitutional requirements of “one man, one-vote”. Minnesota State Senate v. Beens, 406 U.S. 187, 92 S.Ct. 1477, 32 L.Ed.2d 1 (1972). The tenor of the Supreme Court decisions with respect to reapportionment, clearly does not compel institution of single-member districts in North Dakota. The circumstances existing in this state contrast with those existing in states where apportionment plans were found constitutionally deficient. The complexion of the state and of individual cities within the state does not present any showing of unrepresented minorities or unresponsive representatives. The issue of multi-member districts or single-member districts is clearly a political issue to be resolved by the electorate and the legislature. In devising a plan, this Court will go no further than is required to formulate a constitutionally sufficient apportionment plan for the State of North Dakota, and will not intrude on the political realm. II. VARIANCE While we are satisfied that the Court plan adopted June 29, 1972, is not constitutionally deficient by reason of the multi-member districts, consideration must be given to the realm of permissible variance. This Court, by its order filed May 22, 1972, appointed three special masters to serve as a commission to formulate an apportionment plan. The following guidelines to the commission were offered by the Court: “a. The Commission shall try to conform new legislative districts to the existing districts. b. The Commission shall not substantially change the size of the Legislature. c. Natural geographic barriers shall be observed. d. Existing political subdivision lines should be observed, in so far as possible. e. In the event the Commission should find that it is unnecessary to substantially alter any one or more of the legislative districts presently defined, then it must consider and make recommendations relative to whether or not the incumbent senator or senators, whose term does not expire at the end of this year, must nevertheless stand for election in 1972.” One of the unique features existing in North Dakota is the Missouri River, which separates the state into two parts —two-thirds to the east and one-third to the west. The river traversing the state is crossed by only six highway bridges, four of them located in the area of Williston and Bismarck. This geographical reality, coupled with the difficulty of achieving the goals of observing geographical boundaries and existing political subdivisions, adds to the difficulty of minimizing the variance. That it was substantially accomplished, as illustrated by the chart made a part of this opinion, attests to the validity and soundness of the plan. The population variance between the districts in most cases is only a few hundred people. As a general rule, deviations from the average population decreases if many county lines are violated, and increases if few are split. The effort to preserve urban-rural identities and county lines intact produce the variations which look large percentage wise, but when applied to a sparsely populated state do not result in significant population variances. In Mahan, the court considered the validity of a Virginia apportionment plan allowing for a 16% deviation. The Court stated that neither courts nor legislatures, can extract with accuracy from the Fourteenth Amendment, the mathematical formula that establishes what range of percentage deviation is permissible and what is not. While the 16% deviation of the Virginia plan may have approached the limits, it did not exceed them, said the Court. The key, it seems, as to whether a certain deviation will be allowable is whether or not the deviation causes a sacrifice of substantial equality. A percentage of variance can only have validity when measured against the actual number of electors, communities of interest, transportation, and size of the base from which the representation is drawn. Virginia is a heavily populated state, and a 16% deviation there results in population discrepancies of thousands. That same percentage applied to a sparsely populated state like North Da-kota, would result in actual deviations of only a few hundred persons. The issue presented to the court in Mahan .was whether or not the equal protection clause permits only limited population variances which are unavoidable, despite a good faith effort to achieve absolute equality. The court said that some variation from the one man, one vote rule is unavoidable, because blind application of “absolute equality” in state redistricting might well impair the normal functioning of state and local governments. In Swann v. Adams, 385 U.S. 440, 87 S.Ct. 569, 17 L.Ed.2d 501 (1967), the Supreme Court disapproved a Florida reapportionment plan having a 26% deviation. The Court, noting that no evidence in support of the deviation had been offered, said the allowable variation for one state has little bearing on the validity of a similar variation in another state. In 1972, a three judge district court in Graves v. Barnes, 343 F.Supp. 704 (W.D.Tex.1972), declared unconstitutional a Texas reapportionment plan having a total variation between the largest and smallest district of 9.9%. On appeal, the Supreme Court in White v. Regester, 412 U.S. 755, 93 S.Ct. 2332, 37 L.Ed.2d 314 (1973), said that a 9.9% deviation does not establish a prima facie equal protection violation. Citing Mahan, the Court said that relatively minor population deviations among state legislative districts do not substantially dilute the weight of individual votes in the larger districts. Similarly, Gaffney v. Cummings, 412 U.S. 735, 93 S.Ct. 2321, 37 L.Ed.2d 298 (1973), decided at the same time as White, held that a state’s redistricting plan is not to be judged by the more stringent standards applicable to Congressional reapportionment. In Gaffney, the Court said that while fair and equal representation is the goal, its attainment does not depend on elimination of insignificant population variations. “Fair and effective representation may be destroyed by gross population variations among districts, but it is apparent that such representation does not depend solely on mathematical equality among district populations. There are other relevant factors to be taken into account and other important interests that states may be legitimately mindful of. An unrealistic overemphasis on raw population figures, a mere nose count in the districts, may submerge these other considerations and itself furnish a ready tool for ignoring factors that in a day-to-day operation are important to an acceptable representation and apportionment arrangement.” “That the Court was not deterred by the hazards of the political thicket when it undertook to adjudicate the reapportionment cases does not mean that it should become bogged down in a vast, intractable apportionment slough, particularly when there is little, if anything, to be accomplished by doing so. “This very case represents what should not happen in the federal courts. The official state functionaries proposed a plan with a maximum variation among the districts of 7.-89% in the House and 1.8% in the Senate, and with respective average variations of 1.9% and .45%. Appellees then proposed four alternative plans for the House, three of which involved slightly larger variations among districts but cut fewer town lines. The fourth cut more lines, but had a maximum variation between its largest and smallest district of only 2.6%. The District Court thought the state plan involved acceptably large variations between districts, although in the House, with districts of about 20,000 people, the average variation involved only 399 people, and the largest variations involved only 1,573 people. But neither did the District Court adopt any of the plans submitted by appellees. Instead, it appointed its own master to come up with still another scheme. That plan, we are told, involves a total maximum deviation in the House of only 1.16%. Was the master compelled, as a federal constitutional matter, to come up with a plan with smaller variations than were contained in appellees’ plans ? And what is to happen to the master’s plan if a resourceful mind hits upon a plan better than the master’s by a fraction of a percentage point? Involvements like this must end at some point, but that point constantly recedes if those who litigate need only produce a plan that is marginally ‘better’ when measured against a rigid and unyielding population equality standard.” 412 U.S. at 748-751, 93 S.Ct. at 2329-2330, 37 L.Ed.2d at 309-311. (citations omitted). Once a plan has been formulated which, in consideration of all the attendant circumstances, fairly meets the constitutional requirements, a court should not continue to sift and shuffle abstract figures solely to arrive at a mathematically perfect plan. Where there are identifiable minorities or political groups being forced out of the election process and, their voting strength is invidiously weakened, a plan proper in other respects may require a further examination. As we view the cases, deviations of 16%, 8%, and 10% have recently been allowed in heavily populated states having significant minority populations. Even greater deviations are permissible in a sparsely settled state barren of electorally victimized minorities. We adopt the Court Plan of reapportionment previously adopted as an interim plan on June 29, 1972, as set forth in Appendix A of that Order, (published at 372 F.Supp. 363) as the permanent plan for reapportionment for the State of North Dakota. It is ordered that judgment be entered accordingly. . Governor’s veto message, Defendant’s Report of the Court, filed March 30, 1973. . Richard R. Dobson, one of the Special Masters appointed by the Court. . Stanley W. Voelker and Thomas K. Ostenson, Population Changes within Census County Divisions of North Dakota, 1950-1970, March, 1971. . 1970 Censns of Population, U. S. Dept. of Commerce, Bureau of the Census, 1970. . 1973 World Almanac, Newspaper Enterprise Association.
BRIGHT, Circuit Judge (dissenting). I respectfully dissent. Today, a reconstituted majority of this court permanently adopts the stopgap apportionment scheme for North Dakota’s State Legislative Assembly known as the Dobson Plan, in the face of our earlier order which stated: We approve the Dobson Plan of reapportionment at this time for the 1972 election only. This court retains jurisdiction of this cause for the purpose of adopting a different plan of reapportionment which will not be hampered by considerations of impending elections. [Chapman v. Meier, 372 F.Supp. 363, 367 (D.N.D., filed June 30, 1972)]. My colleagues thus have wordlessly brushed past the doctrine of the "law of the case,” changed completely the course of our previous decisions, and, without further hearing of any kind, placed into effect a legally flawed plan whose overriding virtue at the time of its “temporary” adoption in 1972 was that it would cause minimum disruption in the then-pending elections and yet reduce substantially the existing inequality between legislative districts under the “one-man, one-vote” standard. In our order filed June 30, 1972, we stated: We recognize certain weaknesses in the Dobson Plan, namely, (1) some variance in population among the legislative districts, which, in a few instances, seems substantial; (2) an increase in the size of the legislature, notwithstanding that the state has lost population over the past decade; and (3) a continuation of multi-member legislative districts. [Chapman v. Meier, supra 372 F.Supp. at 363.] Simply stated, in adopting this plan today, the majority has not followed the supervisory admonition of the Supreme Court, contained in Connor v. Johnson, 402 U.S. 690, 91 S.Ct. 1760, 29 L.Ed.2d 268 (1971), and reaffirmed in Mahan v. Howell, 410 U.S. 315, 93 S.Ct. 979, 35 L.Ed.2d 320 (1973), disapproving the use of multi-member districts in court-fashioned plans. Nor has the majority adhered to the equal protection standard for permissible population variances between districts, enunciated in Mahan, supra; Gaffney v. Cummings, 412 U.S. 735, 93 S.Ct. 2321, 37 L.Ed.2d 298 (1973), and White v. Regester, 412 U.S. 755, 93 S.Ct. 2332, 37 L.Ed.2d 314 (1973). My colleagues argue that we must not “intrude on the political realm” by eliminating multi-member districts, nor should we “sift and shuffle abstract figures” to achieve smaller deviations between districts. These arguments, I fear, miss the point. The former ignores the political consequences of simply preserving the status quo, while the latter substitutes convenience for the Constitution. I agree with my colleagues that we must decide this case outside the “political realm,” but the assumption of the majority that doing nothing with the existing plan is not doing something represents a “head-in-the-sand” viewpoint. Either doing something or doing nothing may have political consequences, but we, of course, must close our eyes to those consequences. Our decision must be based upon application of proper legal and constitutional principles. The path to a proper resolution of the substantive issues in this case lies in understanding the nature of the decision that we are called upon to make. Unlike so many courts, we are not confronted with a challenge to an apportionment scheme duly enacted by a state legislature, where the threshold question is whether a constitutional violation exists to justify federal court intervention. We are well past that point in this litigation. The invalidity of the existing apportionment' scheme was adjudicated in our order of May 22, 1972. What remained, following our temporary adoption of the Dobson Plan on June 30, 1972, was a final exercise of our equity power to fashion a more permanent remedy in the form of a new apportionment plan. The general question now before us is this: what are the obligations and restrictions on the discretion of a federal district court in the exercise of this equity power? In particular, we search for an answer to these two specific questions : 1) Must the court replace multi-member districts with single-member districts where the former were never established as a matter of state policy, but were originally adopted as an expedient in a previous court-fashioned plan ? 2) How close to absolute equality must the court come in re-drawing district lines in the absence of clearly controlling state apportionment policy? Before suggesting my answers to these two points, I think that a detailed review of the history of reapportionment in North Dakota is useful. I. History shows that the North Dakota Legislature reapportioned itself infrequently and not too well. Mr. Richard Dobson, political editor of the Minor Daily News and drafter of the Dobson Plan here in question, has described the situation in these words: In eight decades of statehood, the reapportionment problem has never been dealt with satisfactorily. Much of the time the legislature simply tried to ignore the problem. When action was taken, it was usually designed to preserve the status quo and safeguard incumbent legislators. [Dobson, Reapportionment Problems, 48 N.D.L.Rev. 281 (1972).] Importantly, there has been no valid, new apportionment scheme placed into effect by the legislature since 1931. Variances between districts, which existed even then, swelled with population shifts over the intervening years. For example, Renville County had 7,263 inhabitants in 1930, while the City of Fargo had 28,619; 30 years later, the 43rd District (Renville County) had shrunk to 4,698 while the 9th District (Fargo) had expanded to 46,857. See Dobson, supra at 285. In 1961, a state reapportionment commission — created by a 1960 amendment to the North Dakota Constitution— sought to re-draw the districts, but produced a plan clearly favoring the rural areas in representation. An initial court-challenge was rejected by the state supreme court on a minor procedural point. State ex rel. Aamoth v. Sathre, 110 N.W.2d 228 (N.D.1961). The plaintiffs then turned to the federal district court which abstained, pending further action by the state supreme court. Lein v. Sathre, 201 F.Supp. 535 (D.N.D.1962) (Davies, J., dissenting). Accepting jurisdiction, the state supreme court invalidated the commission plan based on state consitutional law principles. State ex rel. Lein v. Sathre, 113 N.W.2d 679 (N.D.1962). Following the United States Supreme Court’s historic decision in Baker v. Carr, 369 U.S. 186, 82 S.Ct. 691, 7 L.Ed.2d 663 (1962), the plaintiffs returned to federal district court, but were again rebuffed. The court denied injunctive relief on the ground that there was no showing that the state legislature would not heed the state constitution’s mandate to redistrict. Lein v. Sathre, 205 F.Supp. 535 (D.N.D.1962) (Davies, J., dissenting). Rural interests in the legislature, however, pushed through a reapportionment scheme in the 1963 Session which comported with no known version of the “one-man, one-vote” principle. See Dobson, supra at 286. Based upon the then-newly-decided case of Reynolds v. Sims, 377 U.S. 533, 84 S.Ct. 1362, 12 L.Ed.2d 506 (1964), the federal district court invalidated that scheme as well as those portions of the state constitutional provisions relating to apportionment (sections 26, 29, and 35) amended in 1960. The court, however, refused to take further action until the legislature was given an adequate opportunity to reapportion itself within the Reynolds v. Sims guidelines. Paulson v. Meier, 232 F.Supp. 183 (D.N.D.1964) (Davies, J., dissenting in part) [hereinafter Paulson I]. A compromise legislative plan, enacted on the final day of the 1965 Session, became law without the signature of the governor, but made little improvement over the invalidated 1963 version. This plan was promptly nullified by the federal district court which then ordered into effect its own plan. Paulson v. Meier, 246 F.Supp. 36 (D.N.D.1965) [hereinafter Paulson II]. In a radical departure from state precedent, the court included in its plan five predominantly urban, multi-member senate districts, thereby creating the problem which is at the heart of the present dispute. This court-fashioned plan remained in effect until voided by our order of May 22, 1972, in the instant action. Early in 1971, following some inconclusive grappling with the politically hot issue of multi-member districts by the legislature, a suit was brought in state court challenging the districts on state constitutional grounds. The North Dakota Supreme Court ruled that that portion of section 29, which since statehood had required single-member senate districts, had been necessarily invalidated along with the other constitutional provisions relating to reapportionment in Paulson I, and dismissed the suit. State ex rel. Stockman v. Anderson, 184 N.W. 2d 53 (N.D.1971). This action was commenced in November 1971. Plaintiff alleged that population changes reflected in the 1970 Census had altered the urban-rural balance and that the court-fashioned plan of 1965 no longer provided representation in the legislature complying with the Equal Protection Clause of the fourteenth amendment. They further alleged that, since the legislature had failed to reapportion itself in 1971, the court was required to do so. A three-judge district court was not requested by plaintiffs until they filed an amended complaint on May 8, 1972. This panel, therefore, faced a time pressure in hearing the case and awarding relief, if appropriate, before the then-impending 1972 primary and general elections. We were acutely aware that making substantial changes in legislative district boundaries might lead to great confusion and disruption in the conduct of the elections and the operations of the political parties in North Dakota, which are required to organize and operate on the basis of legislative districts. N.D.Cent.Code § 16-17-01 et seq. (Supp.1971). This court, nevertheless, determined that our obligations as judges required us to undertake immediate consideration of positive action. After hearing the case on May 18, 1972, we acted on the merits four days later, declaring: The plaintiffs have made a legally sufficient showing that the existing legislative apportionment scheme in the State of North Dakota fails to meet Federal Constitutional standards. [Chapman v. Meier, Civ. No. 4664 (D.N.D., filed May 22, 1972) at 1.] We further said that we would attempt promptly to reapportion the state in conformity with federal constitutional standards and in time for the 1972 elections. To aid us in this task, we appointed three Special Masters. Exhibiting our real concern for avoiding disruption to the political processes on the eve of the 1972 elections, we set out specific guidelines for the Masters. On June 20-21, 1972, we met with the Masters to consider eight alternate plans of reapportionment. In our memorandum opinion and order filed June 30, 1972, we settled on a slightly amended version of a plan drafted by Master Dobson, now called the Dobson Plan, as an appropriate means of affording interim relief to the citizens of North Dakota from malapportionment. This relief was tailored to the exigencies of the pre-election situation and geared to minimizing disruption in the political and electoral processes. In their amended complaint, plaintiffs had asked that we provide for single-member districts in any reapportionment plan adopted by the court. In our opinion we recognized the validity of plaintiffs’ request, but did not grant relief at that time in order to avoid “extreme disruption in the elective processes.” However, we directed the Masters to give additional study to reapportionment and to submit an additional report. We recognized the public’s interest in the court’s proceedings by making arrangements for “interested persons” other than the named parties to express their views. Our order filed June 30, 1972, should have been the prelude to a final determination of the reapportionment controversy for the decade of the 1970’s. But this panel, much like the earlier panel in Paulson I, paid great heed to the request of the State of North Dakota to stay our hand to give the legislature another opportunity to “carry out its duties, responsibilities, and functions” to reapportion itself. See Brief in Support of Defendant’s Motion for Stay (filed October 30-1972) at 4, Chapman v. Meier, Civil No. 4664 (D.N.D.). Attached to this motion was a copy of a letter from Representative Bryce Streibel, Chairman of the North Dakota Legislative Council, creating a bipartisan Special Committee on Reapportionment, consisting of six legislators and five citizens, to study and to develop a reapportionment plan for submission to the legislature and “for use by [state and federal] courts in any judicial proceedings relating to reapportionment of the North Dakota legislature.” See Attachment to the Brief in Support of Defendant’s Motion for Stay (filed October 30, 1972), Chapman v. Meier, Civil No. 4664 (D.N.D.). Subsequently, we granted a stay of the proceedings, observing that “no prejudice is likely to follow from this court’s deferring further action on reapportionment until the 1973 legislature has had the opportunity to consider the reapportionment question.” Chapman v. Meier, Civil No. 4664 (D.N.D., filed November 8, 1972) at 3. We further directed the Attorney General to file a report to this court regarding action taken by the legislature, and, retaining jurisdiction, we stated that our order was “without prejudice to the contentions made by the plaintiffs in these proceedings.” Id. at 3-4. The report, filed by the Attorney General on March 30, 1973, discloses that, with great diligence, efficiency, and promptness, the Committee formulated a reapportionment plan which on its face at least complied with constitutional standards and appeared to be free of partisanship. Legislative Council Chairman Streibel described the plan, introduced as House Bill 1042, in these words: House Bill No. 1042 leaves 32 county lines intact. It has the smallest deviation ratio of all the plans submitted to either the Court or the Committee. The deviation ratio is 1.04-1, which is determined by taking the high percentage of the mean population (District #18-12,599 per Senator —• + 2%) and dividing it by the low (District #22-12,150 per Senator percentage of the mean population — —1.6%). House Bill No. 1042 calls for 50 Senators and 100 Representatives in 37 legislative districts. House Bill No. 1042 rigidly adheres to the one-man, one-vote requirement established by the Court. It maintains the “Communities of interest” philosophy to a great degree. Reapportionment is a legislative responsibility and not a job for the Courts. As elected representatives of the people, legislators can express the wishes of the voters regarding reapportionment better than appointed judges. House Bill No. 1042 was adopted unanimously by the Council’s Reapportionment Committee. The plan meets all the Courts’ tests for acceptability. It breaks fewer county lines than many plans submitted. It keeps legislative representation pretty much as it is. It does not drastically reduce the number of legislators which would have had the effect of reducing rural representation. The rural voice of North Dakota is fading fast enough as it is without us hastening it. The plan leaves in the current multisenator districts. Subdistricting can be done at a later date if the legislature so desires. There just isn’t enough census data from ■ Grand Forks, Minot, or Bismarck to reapportion at this time on any basis other than multi-senator districts. In its journey through the legislature, however, House Bill 1042 underwent substantial alteration. District lines were redrawn. The legislature refused to amend the bill to provide for subdistricts in the larger cities. See Senate Journal, January 23, 1973, at 241, Supplementary Return (filed January 22, 1974), Chapman v. Meier, Civil No. 4664 (D.N.D.). The altered reapportionment proposal lost its bipartisan support. Where the original Committee version avoided consideration of partisan politics, the revised plan produced a volley of charges that the district lines were gerrymandered to protect incumbents. See Senate Journal, February 15, 1973, at 582-583, Exhibit 5, Return to and Compliance with Order (filed March 30, 1973), Chapman v. Meier, Civil No. 4664 (D. N.D.). The bill in its altered form was vetoed by Governor Arthur A. Link on February 2, 1973, as “unfair,” principally because of the failure of the legislature to create single-member senatorial districts. See Veto Message, Exhibit 1, Return to and Compliance with Order, supra. In an atmosphere of politically partisan charges and countercharges, the legislature overrode the Governor’s veto by a vote of 72 to 30 in the House and 37 to 14 in the Senate. See Exhibit 4, Return to and Compliance with Order, supra. Yet this reapportionment never became law: first, because under Section 67 of the North Dakota Constitution, it did not take effect until July 1st after the close of the session; and second, because, prior to July 1st, the effectiveness of the measure was stayed by the filing of a referendum petition signed by the requisite number of electors, pursuant to Section 25 of the North Dakota Constitution. In addition, an initiated amendment to the North Dakota Constitution was proposed which would have created a bipartisan commission composed of non-legislators to accomplish reapportionment and which specifically provided for single-member senate districts and single-member house subdistricts thereof. The requisite number of electors signed the initiative petition pursuant to Section 202 of the North Dakota Constitution. Upon the motion of the plaintiffs, and after a hearing on the matter, we ordered a stay of proceedings in this action pending the outcome of the initiative and referendum elections. Chapman v. Meier, Civil No. 4664 (D.N.D., filed May 25, 1973). The elections were held on December 4, 1973. House Bill 1042 was rejected by the voters by a vote of 50,729 to 44,363, while the constitutional amendment was defeated by a vote of 53,831 to 43,178. See Certificate of the Secretary of State, Attached to Defendant’s Motion to Readopt Court-Fashioned Plan (filed December 26, 1973), Chapman v. Meier, Civil No. 4664 (D.N.D.). With the legislature’s action thus nullified and the commission device rejected, the matter is once again before our court without any clearly controlling state reapportionment policies for our guidance. Defendants have moved for an order readopting permanently our reapportionment. plan filed June 30, 1972; plaintiffs resist and move for the adoption of a new plan consistent with our order of that date. Having more clearly defined the historical perspective of the instant action, I turn now to an examination of the legal issues. II. The majority today directs that future elections for the state legislature in the major cities of North Dakota shall again be conducted on a multi-member district basis. In continuing these districts, the court has exceeded the permissible bounds of its discretion. See Connor v. Johnson, 402 U.S. 690, 692, 91 S.Ct. 1760, 29 L.Ed.2d 268 (1971). We are not here concerned with a legislative adoption of such a reapportionment plan. The only plan ever enacted by the legislature containing the multimember feature was House Bill 1042, which was defeated by the voters in a referendum election. If House Bill 1042 had withstood the voters’ attack, the issue before us would be quite a different one and its resolution more evident. For, as the majority correctly points out, the cases clearly hold that the adoption of a multi-member scheme by a legislature does not constitute a per se violation of the Constitution. See Opinion of Judge Benson, supra at 375. But the cases compel a different conclusion when a federal district court is called upon to exercise its discretion in fashioning a new apportionment plan. The Supreme Court has repeatedly noted the objectionable features of multi-member districting. In Lucas v. Forty-Fourth General Assembly of Colorado, 377 U.S. 713, 84 S.Ct. 1459, 12 L.Ed.2d 632 (1964), for example, the Court stated: One of the most undesirable features of the existing apportionment scheme was the requirement that, in counties given more than one seat in either or both of the houses of the General Assembly, all legislators must be elected at large from the county as a whole. Thus, under the existing plan, each Denver voter was required to vote for eight senators and 17 representatives. Ballots were long and cumbersome, and an intelligent choice among candidates for seats in the legislature was made quite difficult. No identifiable constituencies within the populous counties resulted, and the residents of those areas had no single member of the Senate or House elected specifically to represent them. Rather, each legislator elected from a multimember county represented the county as a whole. [377 U.S. at 731, 84 S.Ct. at 1471 (footnote omitted).] Seven years later, in the key case of Whitcomb v. Chavis, 403 U.S. 124, 91 S.Ct. 1858, 29 L.Ed.2d 363 (1971), while upholding a legislatively. created multimember scheme against a head-on constitutional attack, the Court stated: We are not insensitive to the objections long voiced to multi-member district plans. Although not as prevalent as they were in our early history, they have been with us since colonial times and were much in evidence both before and after the adoption of the Fourteenth Amendment. Criticism is rooted in their winner-take-all aspects, their tendency to submerge minorities and to overrepresent the winning party as compared with the party’s statewide electoral position, a general preference for legislatures reflecting community interests as closely as possible and disenchantment with political parties and elections as devices to settle policy differences between contending interests. The chance of winning or significantly influencing intraparty fights and issue-oriented elections has seemed to some inadequate protection to minorities, political, racial, or economic; rather, their voice, it is said, should also be heard in the legislative forum where public policy is finally fashioned. In our view, however, experience and insight have not yet demonstrated that multi-member districts are inherently invidious and violative of the Fourteenth Amendment. [403 U.S. at 157-160, 91 S.Ct. at 1876-1877 (footnotes omitted).] Thus, although multi-member districting does not amount to an Equal Protection violation, there are sufficient undesirable characteristics adhering to it to testify restricting its use in court-fashioned plans, absent special circumstances. This is precisely the rule which must control our decision in the case at bar, and it was enunciated by the Supreme Court under its supervisory power over federal courts in Connor v. Johnson, 402 U.S. 690, 91 S.Ct. 1760, 29 L.Ed.2d 268 (1971). The majority has unsuccessfully sought to distinguish away this controlling precedent by characterizing Connor v. Johnson as a racial discrimination case. This is a misreading of the fact situation, for it is a straight population variance case. On May 14, 1971, a three-judge district court in the Southern District of Mississippi invalidated the then-existing Mississippi legislative reapportionment statute, because of a total variance of 26 percent between the largest and smallest districts, and, on the eve of the forthcoming 1971 legislative elections, attempted to fashion an interim plan of reapportionment — -much as this court was called upon to do in the instant case in June 1972. Connor v. Johnson, 330 F.Supp. 506 (S.D.Miss.1971). The district court’s judgment was that single-member districting would be “ideal” for Hinds County, Mississippi, but that insufficient time remained before the candidate filing deadline, 17 days away, to perform the task of dividing this county into single-member districts. The three-judge court therefore issued a reapportionment plan calling for the at-large election of five senators and 12 representatives in Hinds County. Much like the action which we initially took in this ease on June 29-30, 1972, that court promised to appoint a special master to investigate the possibility of single-member districts for the subsequent elections there in 1975 and 1977. Plaintiffs promptly applied to the United States Supreme Court for a stay, requesting that the filing date for elections be postponed and that the district court provide single-member districts in Hinds County. The Supreme Court granted the application, stating: The district court is instructed, absent insurmountable difficulties, to devise and put into effect a single member district plan for Hinds County * * *. [402 U.S. at 692, 91 S.Ct. at 1762.] The court laid down the following standard for district courts exercising discretion in fashioning apportionment plans: We agree that when district courts are forced to fashion apportionment plans, single-member districts are preferable to large multi-member districts as a general matter. [402 U.S. at 692, 91 S.Ct. at 1762.] I am satisfied that this standard remains the law which governs our decision. In Connor v. Williams, 404 U.S. 549, 92 S.Ct. 656, 30 L.Ed.2d 704 (1972), decided after Whitcomb v. Chavis, supra, the Supreme Court reviewed another challenge to the district court’s Mississippi reapportionment plan and reemphasized its holding in the earlier Connor case, stating: The District Court retained jurisdiction over these three counties and ordered that a Special Master be appointed in January 1972 to “take testimony and make findings as to whether the Counties of Hinds, Harrison, and Jackson may feasibly be divided into districts of substantially equal numbers in population for the elections of 1975 and 1979.” 330 F.Supp., at 519. Such proceedings should go forward and be promptly concluded, for, as this Court has emphasized, “when district courts are forced to fashion apportionment plans, single-member districts are preferable to large multi-member districts as a general matter.” Connor v. Johnson, 402 U.S. 690, 692, 91 S.Ct. 1760, 29 L.Ed.2d 268 (1971). [404 U.S. at 551, 92 S.Ct. at 659.] Most recently, in Mahan v. Howell, 410 U.S. 315, 93 S.Ct. 979, 35 L.Ed.2d 320 (1973), the Court examined the state legislature’s apportionment of Virginia which had been invalidated by a three-judge district court. Although the Supreme Court reversed the district court in adopting a new standard on permissible population variances in state plans, it did affirm the trial court’s use of multi-member districts in reapportioning certain areas containing military personnel. Because of “unusual, if not unique, circumstances” involving severe time pressures, the Court permitted the multi-member plan to stand, but only on an interim basis. 410 U.S. at 331-332, 93 S.Ct. 979. In reaching this decision, the Court noted that: The [district] court conscientiously considered both the legislative policy and this court’s admonition in Connor v. Johnson, supra, that in fashioning apportionment remedies, the use of single-member districts is preferred, evidence of substantial malapportionment with respect to military personnel * * * and the fear that too much delay would have seriously disrupted the fall 1971 elections. Facing as it did this singular combination of unique factors, we cannot say that the District Court abused its discretion in fashioning the interim remedy of combining the three districts into one multimember district. [410 U.S. at 333, 93 S.Ct. at 989.] Turning to the case at bar, it should be obvious that we are long past the “interim” remedy stage of Mahan. The conditions which justified the creation of multi-member senate districts in 1965 by the court in Paulson II, and which justified the temporary continuation of these districts by our court in 1972, have faded into history. I can find no valid state legislative policy nor any practical grounds at the present time justifying the use of multi-member senate districts in North Dakota. Thus, under the language of Mahan and the holdings in the two Connor cases, the continuance of multi-member senate districts constitutes an abuse of discretion by this court. I have read the complete record in this case with care, and find no reasons advanced anywhere in that record for continuing multi-member senate districts as either furthering the art and science of polities or improving the conduct of state government. However, the record does disclose several arguments in favor of the more traditional single-member senate districts: (1) It gives a voter a chance to compare only two candidates, head to head in making a choice. (2) It prevents one political party with a heavy plurality in one or two potential districts from dominating other poential districts that might narrowly go for the candidate of the opposite party. (3) It prevents a city wide political organization from ostracizing or disciplining a legislator, who dares stray from the machine’s line. (4) It permits a citizen to identify a legislator as his senator and makes direct communication easier. (5) It makes each senator responsible for his actions and makes it difficult for a senator to fade into the ranks of “the team” to avoid being identified with specific actions taken. (6) It reduces campaign costs and “personalizes” a campaign. (7) It creates greater interest in the possibility of a citizen seeking a legislative seat without the political machine blessing. (8) It would diminish the animosity created in the legislature against multi-senate districts because of the tendency of senators elected by one political party from a city to vote as a bloc. (9) It would tend to guarantee an individual point of view if all senators are not elected as a team. (10) It would equalize the power of people in single senate districts with the people in the broken down multi-senate districts to influence the election of only one senator. From North Dakota’s earliest days, the policy of single-member senate districts was an integral part of its political tradition. Section 29 of the Constitution of 1889 required such districts and controlled all elections in the state until it was invalidated as an almost accidental by-product of the federal district court’s decision in Paulson II. See State ex rel. Stockman v. Anderson, 184 N.W.2d 53, 57-58 (N.D.1971). When a panel of this court adopted with some hesitancy a truly unprecedented multimember senate plan in Paulson II, the court said: We have exhaustively considered the plan as set forth in Senate Bill 39 [which we hereby adopt]. We find it not perfect. Five “multi-member” districts are created; county lines are violated in twelve instances. * * * Insofar as the multi-member districts are concerned, if experience proves that practical difficulties or inequities result therefrom, appropriate remedial legislation may reasonably be expected. [246 F.Supp. at 44.] In the nine years since the court in Paulson II first introduced multi-member senate districts to North Dakota, they have been the subject of unceasing political, philosophical, and legal disputes. Confronted with “practical difficulties or inequities” referred to in Paulson II, the legislature has been unable to resolve this bitterly divisive issue in a manner acceptable to the electorate. A court plan which has produced and continues to produce this kind of disruptive - results ought not to be readopted. This, unfortunately, is exactly what the majority has done. The evidence is at hand to permit the fashioning of a remedial plan which subdivides the multi-member senate districts in North Dakota’s major cities. During the hearings of the Special Committee on Reapportionment, the late Mr. R. R. Smith, one of the court’s Special Masters, advised the Committee that he had successfully formulated single-member districts for the cities of Bismarck, Grand Forks, and Minot. See Minutes of the Special Committee on Reapportionment, October 11-12, 1972, at 5, and November 8-9, 1972, at 5, Supplementary Return, supra. Mr. Ostenson, also one of our Special Masters, assured the Committee that: [I] t is not difficult to divide multimember districts into single-member districts in Fargo and West Fargo, but in other cities it may be necessary to make ground surveys and estimate the population within enumeration districts. Mr. Ostenson said that if the Committee selects a reapportionment plan, it would not be terribly difficult to adopt single-member districts. [Minutes of the Special Committee on Reapportionment, Nov. 8-9, 1972, at 8, Supplementary Return, supra;] Mr. Ostenson also assured this court that with present census data plus some added ground surveys, the existing multi-member senate districts may be subdivided into single-member senate districts. See Comments to Ostenson Plan, attached as Appendix B, Chapman v. Meier, Civil No. 4664, 372 F.Supp. 363 (D.N.D., filed June 30, 1972). Given this information, the court should proceed to comply with the Supreme Court’s admonition in Connor v. Johnson, supra, and establish single-member senate subdistricts within North Dakota’s urban areas. III. The Dobson Plan cannot stand as a permanent reapportionment plan for North Dakota for a second major reason : it contains unconstitutionally large population disparities between districts. The largest district (population 13,176) exceeds the size of the ideal district by 8.8 percent; the smallest (population 10,728) falls short of the ideal by 11.4 percent — a total deviation of more than 20 percent and a deviation ratio of 1.-23-1. See Chapman v. Meier, Civil No. 4664, 372 F.Supp. 363 at 365 n. 4 (D.N.D., filed June 30, 1972). In our original consideration of this case, the court fashioned interim relief based on the fact that the case came to us on the eve of the 1972 elections. We sought to avoid disruption of the impending elections and, with that consideration in mind, selected the Dobson Plan because it effected less change in the then-existing legislative lines compared to other plans. It is now the time to reevaluate this plan’s variances calmly in light of the Constitution and without these unusual circumstances, but the question is: what standards to apply? In Connor v. Williams, 404 U.S. 549, 92 S.Ct. 656, 30 L.Ed.2d 704 (1972), a court-fashioned plan for Mississippi’s state legislature — containing a total variance of 18.9 percent — was attacked as unconstitutional based on Kirkpatrick v. Preisler, 394 U.S. 526, 89 S.Ct. 1225, 22 L.Ed.2d 519 (1969), and Wells v. Rockefeller, 394 U.S. 542, 89 S.Ct. 1234, 22 L.Ed.2d 535 (1969). The Supreme Court, however, chose not to reach this issue, and instead vacated the district court’s order pending completion of proceedings consistent with the Court’s earlier mandate in Connor v. Johnson, 402 U.S. 690, 91 S.Ct. 1760, 29 L.Ed.2d 268 (1971). Since that time, the Court has more clearly defined the Equal Protection standards applicable to population deviations in legislative redistricting plans enacted by state legislatures. Mahan v. Howell, 410 U.S. 315, 93 S.Ct. 979, 35 L.Ed.2d 320 (1973); Gaffney v. Cummings, 412 U.S. 735, 93 S.Ct. 2321, 37 L.Ed.2d 298 (1973); White v. Regester, 412 U.S. 755, 93 S.Ct. 2332, 37 L.Ed.2d 314 (1973). However, it remains an open question whether a more stringent standard, one of “absolute equality,” applies to reapportionment schemes fashioned by federal district courts in the exercise of their equity power. But even assuming, arguendo, that Mahan’s flexible standard controls court-fashioned plans as well as plans enacted by legislatures, it seems to me that the variances in the Dobson Plan are in all likelihood too great to pass constitutional muster, particularly since no justification is shown on the record for continuing the use of this plan. The touchstone of the Mahan case is its reaffirmation of a statement from Reynolds v. Sims, 377 U.S. 533, 84 S.Ct. 1362, 12 L.Ed.2d 506 (1964), that: So long as the divergences from a strict population standard are based on legitimate considerations incident to the effectuation of a rational state policy, some deviations from the equal-population principle are constitutionally permissible with respect to the apportionment of seats in either or both of the two houses of a bicameral state legislature. [377 U.S. at 579, 84 S.Ct. at 1391.] Applying this flexible standard to the legislative reapportionment by the Virginia General Assembly, the Court upheld a 16.4 percent variation between the largest and smallest districts on the grounds that the devia