Boyer v. Gardner, 540 F. Supp. 624 (D.N.H. 1982)
June 4, 1982
William GARDNER, Secretary of State for the State of New Hampshire.
United States District Court, D. New Hampshire.
*625 James Normand, Manchester, N. H., I. Michael Winograd, Concord, N. H., Katherine M. Hanna, Manchester, N. H., Steven M. Gordon, Concord, N. H., for plaintiff.
Deborah J. Cooper, Deputy Atty. Gen., Concord, N. H., for defendant.
Before CAMPBELL, Circuit Judge, DEVINE, Chief Judge, and LOUGHLIN, District Judge.
LEVIN H. CAMPBELL, Circuit Judge.
This action against the Secretary of State of New Hampshire was commenced on May 17, 1982, by New Hampshire citizens and registered voters. Plaintiffs complain that their equal rights to vote for members of the New Hampshire House of Representatives, and to cast votes that are as effective as the votes of every other citizen in the state, are being diluted and effectively denied by the alleged discriminatory apportionment of seats in the House of Representatives in violation of the Equal Protection Clause of the federal Constitution and Pt. 2, Art. 9 of the New Hampshire Constitution.
The challenged reapportionment plan is embodied in Chapter 29 of the Laws of 1982, amending Chapter 662, Section 6, of the New Hampshire Revised Statutes Annotated. It was enacted into law by the New Hampshire House on April 20, 1982, pursuant to the New Hampshire Constitution's requirement that a reapportionment of legislative seats take place every ten years in accordance with the latest general census (in this case the 1980 Federal Census). Pt. 2, Art. 9, N.H. Constitution. Since representatives to the New Hampshire House are to be elected in the fall of 1982, with candidates to file papers between June 2 and 16, 1982, it was and is a matter of immediate concern whether Chapter 29 constitutes a valid apportionment of legislative seats. Plaintiffs ask, inter alia, for injunctive relief barring acceptance of filings for the office of State Representative on the designated June dates, and for the creation and institution of an alternate apportionment plan in time for the fall elections.
To consider this challenge to the recent New Hampshire reapportionment plan, a three-judge court was promptly convened pursuant to 28 U.S.C. § 2284. A hearing on the merits was held on May 27, 1982, at which time the parties filed with the court a stipulation concerning most of the material facts. The parties also agreed, with some limitations, that the court could receive and consider certain legislative history documents reflecting New Hampshire House debates and the proceedings of its Special Committee on Reapportionment. A single witness, a statistician called by the *626 defendant, was heard. The court then took the case under advisement.
Aware that the candidates' filing date commenced on Tuesday, June 1, following the Memorial Day weekend, this court issued an order on May 28 denying plaintiffs relief, on the basis that Chapter 29 "does not violate any of the rights to vote afforded the citizens of New Hampshire pursuant to the provisions of the Constitution of the United States." We indicated then that a fuller opinion would follow in due course.
We begin by observing that the biennially elected New Hampshire legislature has 400 members. The state Constitution provides that, "founded on principles of equality," representation is to be "as equal as circumstances will admit." Pt. 2, Art. 9. In apportioning seats, "no town, ward or place shall be divided nor the boundaries thereof altered." Pt. 2, Art. 9. When any town, ward, or unincorporated place has fewer than the number of inhabitants necessary to entitle it to one representative, the legislature is to create representative districts with enough inhabitants to entitle the district to one or more representatives. Pt. 2, Art. 11.
After the 1980 census, it appeared that the ideal population for each state representative district was 2,300 persons. This number was determined by dividing the number of seats (400) into the state's population of approximately 920,000. Using this ideal as a point of reference, those responsible for apportionment undertook to create districts which would not violate town and county lines, and would take into account various other factors.
House Bill 2, which became Chapter 29, ended up establishing 162 electoral districts, of which 17 were so-called floterial districts. Plaintiffs' charge of unconstitutionality focuses on these 17 floterial districts.
Plaintiffs acknowledge that floterial districts are not in themselves unconstitutional, apart from inequalities in representation that they may create. See generally Mahan v. Howell, 410 U.S. 315, 93 S. Ct. 979, 35 L. Ed. 2d 320 (1973); Reynolds v. Sims, 377 U.S. 533, 579, 84 S. Ct. 1362, 1390, 12 L. Ed. 2d 506 (1964); Kilgarlin v. Hill, 386 U.S. 120, 124, 87 S. Ct. 820, 823, 17 L. Ed. 2d 771 (1967). See also Opinion of the Justices, 111 N.H. 146, 276 A.2d 825 (1971). Plaintiffs claim, however, that due to differences in the size of the components of the floterial districts here under consideration, the use of these floterial districts results in significant disparities in representation. Plaintiffs further contend that the state's method of calculation masks and minimizes these inequalities. They argue that if the correct method of calculation were used, it would reveal that the populations within floterial districts deviate to an unconstitutional extent from the appropriate norm, resulting in underrepresentation of voters in some communities and overrepresentation in others. Finally, plaintiffs argue *627 that even under the method of calculation employed by the state, House Bill 2 is unconstitutional since it contains deviations from the ideal distribution which cannot be justified by "legitimate considerations incident to effectuation of a rational state policy." Reynolds v. Sims, 377 U.S. 533, 579, 84 S. Ct. 1362, 1390-91, 12 L. Ed. 2d 506 (1964).
We are not persuaded by plaintiffs' arguments that the plan in question is unconstitutional. We begin by discussing the method to be employed in calculating deviations in floterial districts. Plaintiffs argue that the New Hampshire Secretary of State erred in claiming an overall deviation from the "ideal" of no more than 13.74 percent. They contend that this figure, which was arrived at through the use of the so-called "aggregate method," underestimates the disparities in voting power within and between floterial districts. To assess voting strength accurately, plaintiffs argue that the so-called "component method" of calculating deviations should be used. Use of this method would reveal that House Bill 2 contains a maximum deviation in excess of 70 percent. Plaintiffs contend that whatever can be said of a variation of 13.74 percent, a maximum total variation of 70 percent from the ideal population per representative is surely unconstitutional.
The Supreme Court has typically applied the aggregate method in examining the tolerable extent of variation in challenged apportionment plans. The parties have, however, brought to our attention two cases where the method of computing deviations within a floterial district by the component method has been raised in appeals to the Supreme Court. In an early reapportionment case, Kilgarlin v. Martin, 252 F. Supp. 404, 422 n.28 (S.D.Tex.1966), reversed in part on other grounds, 386 U.S. 120, 87 S. Ct. 820, 17 L. Ed. 2d 771 (1967), a three-judge panel found unconstitutional a Texas reapportionment plan containing floterial districts. In examining the floterial districts contained within the state plan, the panel applied the component method of calculation. The Supreme Court, in a per curiam decision, reversed the district court's ruling as to the non-floterial districts, without discussing the method of calculation that *628 should be used in examining floterial districts. Kilgarlin v. Hill, 386 U.S. 120, 87 S. Ct. 820, 17 L. Ed. 2d 771 (1967). While plaintiffs rely heavily on the lower court's decision to apply the component method in Kilgarlin and the Supreme Court's failure to repudiate that method of calculation in its review of the panel's decision, the Court's silence in Kilgarlin, without more, cannot be interpreted as general approval of the component method of calculation.
In a more recent case, Mahan v. Howell, 410 U.S. 315, 93 S. Ct. 979, 35 L. Ed. 2d 320 (1973), the Court declined to enter the debate over which method of computation should be used in examining deviations within a floterial district. In Mahan, the three-judge panel, while recognizing that other methods of computation could be used, employed the aggregate method in examining Virginia's floterial districts. On appeal, it was argued that the component method of calculation would reveal greater deviations than those found by the district court. In refusing to use figures other than those applied by the district court, the Court stated that it "decline[d] to enter this imbroglio of mathematical manipulation ...." Mahan v. Howell, 410 U.S. at 319 n.6, 93 S. Ct. at 982 n.6.
While the Court's rulings in this area are not conclusive, we believe that its statements in Mahan indicate a reluctance to involve the courts in statistical complexities, at least absent reasons to indicate that a more complicated approach is required in given circumstances. We believe for the reasons stated below that the aggregate method may properly be applied in examining New Hampshire's reapportionment plan.
The Supreme Court has traditionally applied the aggregate formula in examining challenged reapportionment plans. To insert a figure reached by a totally different basis of computation into the percentages discussed by the Court in its previous cases is to compare, as the Attorney General argued, apples with oranges. It cannot be expected that a percentage deviation that might not be considered de minimus or acceptable when arrived at by one method of computation would necessarily not be permissible if arrived at by another method. While both the aggregate and component methods give results denominated in terms of "percentage deviations," those percentages are in what might be thought of as different units of measurement, which are not directly comparable.
We also believe that at least in this case, where the political districts under analysis are very small, the aggregate method of analysis sufficiently reveals any defects in apportionment that may exist. The component method measures the inequality in voting power which arises from additions of hypothetical fractional shares in at-large representatives. The numbers involved are so smalleven a 20 percent deviation represents less than 500 votersthat the aggregate method's measurement of variation between the total population of districts and the representatives allotted to them seems a sufficiently precise method of computation for present purposes. Cf. Abate v. Mundt, 403 U.S. 182, 185, 91 S. Ct. 1904, 1907, 29 L. Ed. 2d 399 (1971) (smaller populations may permit larger percentage deviations in state than federal apportionment schemes); Burns v. Gill, 316 F. Supp. 1285, 1298 (D. Hawaii 1970) (three-judge court) (relying on small population size to validate percentage deviations that might otherwise be too large).
*629 It must be borne in mind that the floterial district is a concept devised to equalize representation while preserving political boundaries. It makes it possible for most representative districts to be drawn along conventional lines, so that the voters of each town or group of towns have their visible and identifiable representative of that homogenous community. As an equalizing device it may have its imperfections, in that it may not afford perfect equality of voting power as between inhabitants of two or more communities in certain floterial districts; but none of the other available devices are any more perfect, as is reflected, for example, in the extensive gerrymandering which often occurs when districts are drawn solely on a population basis, in disregard of established town and county boundaries. We think the statistical variations presented by the 17 floterial districts that are highlighted by using the component methodwhile showing the elusiveness of any goal of perfect mathematical equalitydo not in themselves demonstrate an unconstitutional dilution of voting power. Were there a showing of invidious or other otherwise impermissible discrimination in the drawing up of floterial districts, we might find the use of the component method or something like it more illuminating. But in the absence of these motives and there is no evidence of their presence herewe think it is appropriate to look to the aggregate method of computing variations for constitutional purposes in this case.
Having determined that the state's figure of a 13.74 percent overall deviation (arrived at using the aggregate method) is the correct standard by which the plan must be evaluated, we turn to plaintiffs' second argument that, even on this measure, the plan is constitutionally invalid.
We begin by noting that in Reynolds v. Sims, 377 U.S. 533, 84 S. Ct. 1362, 12 L. Ed. 2d 506 (1964), the Supreme Court held that the "Equal Protection Clause requires that the seats in both houses of a bicameral state legislature must be apportioned on a population basis." Id., at 568, 84 S. Ct. at 1385. Realizing that "mathematical precision [was] hardly a workable constitutional requirement," id., at 577, 84 S. Ct. at 1389, the Court ruled that deviations from the ideal were permissible if they were "based on legitimate considerations incident to the effectuation of a rational state policy." Id., at 579, 84 S. Ct. at 1390.
Supreme Court cases following Reynolds have provided guidelines for examining the constitutionality of legislatively enacted state reapportionment plans. While the Court has never enunciated specific maximum variations which will invalidate a reapportionment, it has indicated that deviations below ten percent are prima facie valid under the Constitution, Connor v. Finch, 431 U.S. 407, 418, 97 S. Ct. 1828, 1835, 52 L. Ed. 2d 465 (1977), and that deviations slightly greater than ten percent are sustainable if based on "legitimate considerations incident to effectuation of a rational state policy." Mahan v. Howell, 410 U.S. at 325, 93 S. Ct. at 985, quoting from Reynolds v. Sims, 377 U.S. at 579, 84 S. Ct. at 1390-91.
Here the Secretary has been able to articulate several justifications for the divergence from absolute equality. First, state policies regarding maintenance of county, city and town boundaries were incorporated *630 into the plan.See Pt. 2, Art. 9 and Pt. 2, Art. 11, N.H. Constitution. The legitimacy of these policies cannot be doubted. See, e.g., Mahan v. Howell, 410 U.S. 315, 93 S. Ct. 979, 35 L. Ed. 2d 320 (1973). In themselves, they created significant limitations on what shape the plan could take. Second, pursuant to Pt. 2, Art. 11 of the state Constitution, it was necessary to aggregate into single districts only towns and wards having boundaries "reasonably proximate" to one another. Third, the plan, as required by the state Constitution, could neither divide nor alter existing ward boundaries. Last, it attempted to group towns of like size in creating districts.
We have been given no reason to believe that these legitimate policies are being used to mask invidious purposes of any kind. No claim of racial, ethnic or religious discrimination is suggested. The record shows that the legislature in devising the plan sought extensive comments from county, city and town representatives and that many of the concerns thus raised were incorporated into the bill adopted by the House.
The record of the special committee meetings and the House debate on the committee's report, suggests that those formulating the reapportionment plan saw themselves as motivated by a desire to serve the best interest of New Hampshire citizens by devising a reapportionment plan which both optimized representation and recognized the physical, political, social and geographical interests of the various political subdivisions within the state.
Given the absence of proof that the plan was the product of bad faith or invidious design, we believe that the apportionment scheme reflected in the Bill adopted by the House serves "legitimate considerations incident to the effectuation of ... rational state polic[ies]." Reynolds v. Sims, 377 U.S. at 579, 84 S. Ct. at 1390-91.
We accordingly see no reason to regard the 13.74 percent variation contained in the state reapportionment bill as invalid. Since the variations contained within Chapter 29 of the Laws of 1982, amending in certain respects Chapter 662, Section 5 of the New Hampshire Revised Statutes Annotated, do not exceed constitutional norms, plaintiffs' petition must therefore be dismissed.NOTES
 Plaintiffs filed an amended complaint on May 25, 1982. Jurisdiction is claimed in the amended complaint under 42 U.S.C. §§ 1982, 1988; 28 U.S.C. §§ 1343(4), 2284.
 Plaintiffs have not argued that the New Hampshire Constitution presents a standard distinct from the federal Constitution. We therefore have no occasion to treat the state law issue separately herein.
 The 1980 census lists the statewide population as 920,610 persons.
 The parties have stipulated to the following definition of a floterial district:
A floterial district is a legislative district which includes within its boundaries several separate districts or political subdivisions which independently would not be entitled to additional representation, but whose conglomerate population entitles the area to another seat in the particular legislative body being apportioned.
Davis v. Mann, 377 U.S. 678, 686 n.2, 84 S. Ct. 1441, 1445 n.2, 12 L. Ed. 2d 609 (1964).
By using some, relatively few, floterial districts, it is possible to create many regular districts which respect the boundaries of recognized political subdivisions. Deviations accumulated in regular districts are corrected for by assigning an extra representative to a floterial district made up of two or more regular districts. For example, the Weare district, with a population of 3,232, received one representative, based on the ideal population per representative figure of 2,300. The New Boston-Francestown-Bennington district ("New Boston"), with a population of 3,648, similarly received one representative. Each district, however, has more than the 2,300 persons ideally assigned a single representative. The two districts are therefore combined into a single floterial district in order to correct for this fractional underrepresentation within Weare and New Boston individually. The floterial district is assigned a single floterial representative elected at large by the voters of both Weare and New Boston.
 The two methods of calculation operate as follows, using the Weare and New Boston districts mentioned in footnote 4 as an example. On the aggregate method, one obtains the difference between the total population of a floterial districthere 6,880and the "ideal" population for its total number of representatives, which is obtained by multiplying the ideal population per representative (2,300) by the total number of representatives (both regular and floterial)here three. This difference, 6,880 - (2,300 × 3) = -20, is divided by the ideal population (2,300) to obtain the percentage deviation, here -.28 percent.
On the component method, each component district within the floterial district is assigned a "share" of the floterial representative based on its percentage of the total floterial population. See generally Cosner v. Dalton, 522 F. Supp. 350, 355 n.8 (E.D.Va.1981) (three-judge court). Thus, Weare's share would be 3,232/6,880, or 47 percent, while New Boston's would be 3,648/6,880, or 53 percent. This share is then added to the number of regular representatives of each component district, giving Weare a total of 1.47 and New Boston a total of 1.53. These are divided into each district's population to arrive at component populations per representative of 3,232/1.47 = 2,198 for Weare and 3,648/1.53 = 2,384 for New Boston. As with the aggregate method, the percentage deviation from 2,300 is then calculated, resulting in a figure of -4.4 percent for Weare and + 3.7 percent for New Boston.
It will be seen that the component deviation figures are higher than the corresponding aggregate deviations. The aggregate figure is an index of differences from the norm between floterial districts, while the component figure points out differences resulting from differential "shares" of floterial representatives assigned to each regular district. Of course, fractions of a floterial representative are never actually given to each component district. Rather, the representative is assigned to the floterial district as a unit.
 This figure was arrived at by computing the maximum positive deviation on the component method, -49.1 percent for the Litchfield district, and adding to it the maximum negative deviation on that method, 21.2 percent for the New Ipswich district, for a combined spread away from the ideal figure of 2,300 of 70.3 percent.
To avoid any confusion, it should be emphasized that the majority of districts were within 10 percent high or low of the 2,300 figure even on the component method. The 70.3 percent figure applies only to the Litchfield/New Ipswich example.
 While neither the Supreme Court nor the lower court in Mahan specified the details of the alternative methods of computation there under consideration, it appears that they were either identical to the aggregate and component methods here used, or else reasonably similar to them.
 The mathematics of the component method virtually ensures that some amount of deviation will result from any combination of regular districts into floterial ones. The conditions for an absence of deviation are more stringent and thus more difficult to satisfythan the corresponding conditions under the aggregate method. This may lend further support to the observation that the upper limit of acceptable deviation when computed using the component method might well be expected to exceed the limits used with aggregate method calculations.
 In Mahan, the Court stated that while the 16.4 percent deviation in that case did not exceed constitutional limits on state legislative redistricting, it nevertheless considered that "this percentage may well approach tolerable limits." 410 U.S. at 329, 93 S. Ct. at 987. This statement has been interpreted to imply that percentage deviations greater than 16.4 percent are unconstitutional despite the presence of legitimate state considerations which would otherwise justify the deviations. See generally Gaffney v. Cummings, 412 U.S. 735, 744, 93 S. Ct. 2321, 2326-27, 37 L. Ed. 2d 298 (1973); Note, The Reapportionment Dilemma: Lessons from The Virginia Experience, 68 Va.L.Rev. 541 (1982).
Since the deviation in this case is within the limits established by Mahan, we believe that it should be examined under the "legitimate considerations" standard contained in Reynolds and applied in Mahan.
 We note that under New Hampshire law, the state representatives of the districts of each county comprise the County Convention, which has the power to raise county taxes, make appropriations, and authorize the purchase or sale of county real estate. N.H. RSA 24:1. See also N.H. RSA 24:13 (Supp.1981) and N.H. RSA 24:13-a. It stands to reason, therefore, that as was done here, the state will seek to preserve county lines in any reapportionment plan.
 The special bipartisan committee selected from members of the House to devise a reapportionment plan met over a period of nine months to consider various proposals. The committee began by dividing the number of representatives to be apportioned among the various counties. Special subcommittees were then selected to consider how the seats could be apportioned throughout each county.
The committee held several public meetings during which interested members of the public were invited to make suggestions and to comment on the proposed plans. Members of the committee also sought formal and informal comments from members of various county, town and city government units. These comments included suggestions that the plan consider the existing relationships of townsincluding physical accessibility, shared services and telephone exchanges, and their urban or rural naturebefore grouping them together within the same district.