NOTICE: This opinion is subject to motions for rehearing under Rule 22 as well as formal revision before publication in the New Hampshire Reports. Readers are requested to notify the Reporter, Supreme Court of New Hampshire, Supreme Court Building, Concord, New Hampshire 03301, of any editorial errors in order that corrections may be made before the opinion goes to press. Opinions are available on the Internet by 9:00 a.m. on the morning of their release. The direct address of the court's home page is:




No. 2002-0210




Argued: June 11, 2002

Opinion Issued: July 26, 2002

Hatem & Donovan, P.C., of Salem (Michael D. Hatem and Bonnie J. Boulanger on the memorandum, and Mr. Hatem orally), for the petitioners.

Hinckley, Allen & Snyder, LLP, of Concord (Michael J. Connolly and Christopher H.M. Carter on the memorandum, and Mr. Carter orally) for the Speaker of the New Hampshire House of Representatives.

Betsy B. Miller, house legal counsel, by memorandum and orally, for the New Hampshire House of Representatives.

John M. Pratt & a., by memorandum, and Mr. Pratt orally, pro se, as amici curiae.

Barry J. Glennon, staff attorney, of Concord, filed no memorandum, for the Secretary of State.


Per Curiam. Daniel Webster once said, "[T]he right to choose a representative is every man’s portion of sovereign power." Luther v. Borden, 48 U.S. (7 How.) 1, 30 (1849) (statement of counsel).

For the first time in the history of this State, the supreme court is required to scrutinize the process of apportioning the people’s right to vote in the election of representatives. That scrutiny has revealed significant anomalies, perpetuated for many years in the legislative redistricting process, which have undermined the principles of equality upon which the New Hampshire House of Representatives was founded. See N.H. CONST. pt. II, art. 9. Rather than protecting the people’s constitutional right to "one person/one vote," a system has evolved that falls far short of that ideal. We hold, therefore, that the current method of creating districts fails to insure that "every voter is equal to every other voter" in this State. Gray v. Sanders, 372 U.S. 368, 380 (1963).

This court has been drawn reluctantly into what is primarily a legislative task. "It is not our function to decide the peculiarly political questions involved in reapportionment, but it is our duty to insure the electorate equal protection of the laws." Silver v. Brown, 405 P.2d 132, 140 (Cal. 1965). Therefore, when the legislature has failed to act, it is the judiciary’s duty to devise a constitutionally valid reapportionment plan. See Scott v. Germano, 381 U.S. 407, 409 (1965) (per curiam).

In furtherance of that duty, we establish a plan for new house districts. Accordingly, RSA 662:5 (1996) is no longer applicable. This plan corrects the constitutional deficiencies in the existing districts and eliminates the present inequities. We are primarily governed by the constitutional requirement of "one person/one vote." In addition, in this case, we are able to adhere to other New Hampshire constitutional requirements and traditional State redistricting policies. We are indifferent to political considerations, such as incumbency or party affiliation. The plan we establish restores as nearly equal weight as possible to the votes of the people of New Hampshire. We do this by eliminating floterials and creating as many single-member districts as possible, with as few multi-member districts as necessary.


I. Background and Procedural History

The New Hampshire Constitution requires the legislature to redraw each representative district "as equal as circumstances will admit" every ten years, based upon the decennial census. N.H. CONST. pt. II, art. 9; see N.H. CONST. pt. II, art. 11. In anticipation of the results of the 2000 census, the house began the reapportionment process in January 2001 with the introduction of House Bill (HB) 420.

According to the 2000 census, between 1990 and 2000, New Hampshire’s population grew more than 10%, increasing from 1,109,252 citizens in 1990 to 1,235,786 citizens in 2000. This growth was unevenly distributed between the northern and southern portions of the State, however, with the largest population growth occurring in the south. As a result, it is undisputed that following the 2000 census, the existing representative districts, established in 1992 pursuant to the 1990 census, violate both the State and Federal Constitutions. See N.H. CONST. pt. I, art. 11; N.H. CONST. pt. II, art. 9; U.S. CONST. amend. XIV; RSA 662:5.

In the winter of 2002, after a series of public hearings on proposed reapportionment plans, HB 420, containing a new apportionment plan for the house was passed by both the house and the senate along party lines. The Governor vetoed the bill, however, on April 3, 2002. The house considered the Governor’s veto on May 22, 2002, but was unable to achieve the two-thirds vote necessary to override it. As a result, HB 420 did not become law. See N.H. CONST. pt. II, art. 44.

In April 2002, the eleven petitioners, all incumbents, filed a petition for original jurisdiction requesting the court to declare the existing representative districts unconstitutional and to impose a deadline for the legislature to enact a valid reapportionment plan for the house. Given the imperative to establish a redistricting plan consistent with constitutional requisites before the 2002 elections, the court accepted jurisdiction. See Monier v. Gallen, 122 N.H. 474, 476 (1982).

Because the senate and house recessed on May 22, 2002, without enacting a house reapportionment plan, the court determined that it must establish a constitutional reapportionment plan for the house before a 2002 election could be held. See Reynolds v. Sims, 377 U.S. 533, 585 (1964); Connor v. Finch, 431 U.S. 407, 415 (1977).

The court has endeavored to reapportion the house as fairly, as efficiently and as quickly as possible. It ordered the parties to submit constitutional reapportionment proposals by June 6, 2002. The court further required that any proposal submitted be based upon the 2000 census data and comply with the constitutional principle of one person/one vote. Oral argument was held on June 11, 2002. It was not until July 16, 2002, that the parties finally provided the court with all necessary information. This decision follows ten days later.

The court informed the parties of its intent to appoint Bobby Bowers, Director of the South Carolina Budget and Control Board Office of Research and Statistics, as its technical advisor in this case because it is an "extraordinary [one] where the introduction of outside skills and expertise, not possessed by the judge, will hasten the just adjudication of a dispute without dislodging the delicate balance of the juristic role." Reilly v. United States, 863 F.2d 149, 156 (1st Cir. 1988). Without objection, the court appointed him pursuant to its inherent authority. See id. See generally State v. Coon, 974 P.2d 386, 395-96 (Alaska 1999) (discussing authority of courts to appoint expert technical advisors). Bowers was appointed to serve the same role in this case as he was appointed to serve in Below v. Gardner, 148 N.H. ___ (decided June 24, 2002).

We have reviewed, in detail, each plan filed in accordance with court deadlines and have also considered the written and oral submissions of the parties.


II. Governing Principles

The New Hampshire Constitution is the supreme law of this State. See Merrill v. Sherburne, 1 N.H. 199, 217 (1818). The oath we took to honor that constitution makes it our duty to apply the State Constitution when it does not conflict with the United States Constitution. See State v. LaFrance, 124 N.H. 171, 177 (1983).

A. One Person/One Vote

1. History of Part II, Articles 9 and 11

We begin with a discussion of the one person/one vote standard under our own constitution. The New Hampshire Constitution guarantees that each citizen’s vote will have equal weight. N.H. CONST. pt. I, art. 11. With respect to the house of representatives, this right is assured by Part II, Articles 9 and 11 of the State Constitution.

Part II, Article 9, as amended in 1964, requires that the house of representatives be "founded on principles of equality" and that representation in the house of representatives "be as equal as circumstances will admit." N.H. CONST. pt. II, art. 9. Part II, Article 11, as amended in 1964, states, in pertinent part:

When any town, ward, or unincorporated place, according to the last federal decennial census, has less than the number of inhabitants necessary to entitle it to one representative, the legislature shall form those towns, wards, or unincorporated places into representative districts which contain a sufficient number of inhabitants to entitle each district so formed to one or more representatives for the entire district. In forming the districts, the boundaries of towns, wards and unincorporated places shall be preserved and the towns, wards and unincorporated places forming one district shall be reasonably proximate to one another.

N.H. CONST. pt. II, art. 11.

Both articles were last amended as a result of the Constitutional Convention in 1964. See Journal of Constitutional Convention 334, 358 (1964). Before 1964, Part II, Article 9 required the legislature to reapportion the house of representatives every ten years, following the taking of the national census. See Levitt v. Stark, 233 F. Supp. 566, 567 (D.N.H. 1964). It further provided:

The number of inhabitants necessary to entitle any town or ward to representatives additional to the first shall be for each additional representative twice the number of inhabitants required for the first representative, so that the mean increasing number for every additional representative shall be twice the number required for the first or one representative.

Id. (quotation omitted).

Before 1964, Part II, Article 11 required the legislature to provide representation for towns or wards having fewer than the number of inhabitants to entitle them to a representative "in at least one session in every ten years." Id. at 567-68 (quotation omitted). These towns or wards elected a representative "such proportionate part of the time as the number of its inhabitants shall bear to the requisite number established for one representative." Id. at 568 (quotation omitted).

In 1961, the legislature set the number of inhabitants entitling a town to one representative at 822. Id.; see Laws 1961, ch. 275. It also established the election years in the decade to follow in which towns having fewer than 822 people were to elect a representative. Levitt, 233 F. Supp. at 568; see Laws 1961, ch. 275. Under the 1961 law, the smallest of these towns elected a representative to only one out of the five legislatures to be called in the decade, and the largest of these towns elected a representative to four out of the five legislatures to be called in the decade. Levitt, 233 F. Supp. at 568. Inhabitants of unincorporated places had no representation. Id. Thus, before 1964, representation in the house was essentially based upon a principle of one town/one vote, not one person/one vote.

At the 1964 Constitutional Convention, resolutions were introduced to amend both Part II, Article 9 and Part II, Article 11 to comply with recent United States Supreme Court decisions. The Governor supported amending Part II, Article 9 because he was "convinced that our present requirement that a town or ward have twice as much population for each additional representative as it needs for the first, would be declared unconstitutional." Journal of Constitutional Convention, supra at 48. He urged delegates to amend Part II, Article 11 so as to "provide a system of full-time representation in the House of every citizen in New Hampshire" and, thus, "forestall action by the courts." Id. The Governor warned that if the convention failed to amend these articles, "precedents in other states show that the courts will take action in your place." Id.

Approximately one week after the convention began, the United States District Court for the District of New Hampshire issued its opinion in Levitt. In Levitt, the federal court stated that it "entertain[ed] serious doubt of the federal constitutional validity of the New Hampshire method for selecting the members of the legislature." Levitt, 233 F. Supp. at 569. The court noted, however, that the United States Supreme Court had not yet held that both houses of a bicameral state legislature had to be apportioned on the basis of population, and intimated that if only one of the houses of the New Hampshire Legislature were apportioned on the basis of population, the other house might survive federal court scrutiny. Id.

Before adjourning on June 10, 1964, the convention successfully passed a resolution to amend Part II, Article 11. The resolution to amend Part II, Article 11 "was intended to grant to the General Court the power to create districts where there are towns, wards and unincorporated places which are too small to be entitled to one full-time representative." Journal of Constitutional Convention, supra at 231. As one delegate noted, "At the present time, such towns, wards and places send representatives on a part-time basis only" and the delegates "have been warned in [Levitt] that this provision for part-time representation is probably unconstitutional." Id.

The amendment was also intended to give the legislature "a large measure of flexibility in forming districts." Id. Thus, the legislature was not confined to drawing single-member districts, but was authorized to form districts that were represented by "one or more representatives." Id.; see Opinion of the Justices, 111 N.H. 146, 150-51 (1971). Nor was the legislature required to form these multi-member districts only from the towns, wards and places that formerly had only part-time representation. The reference to "those towns, wards or unincorporated places" was not intended to limit the legislature’s discretion as to how to form multi-member districts. Journal of Constitutional Convention, supra at 231. As one delegate observed, "[W]e cannot here draw up a districting. That is a matter which will have to be taken up with the Legislature." Id. at 220.

The requirement that the towns, wards and places within a district be "reasonably proximate" to one another was also intended to give the legislature flexibility in drawing house districts. The convention delegates "felt that the Legislature should join towns and places which are close together wherever possible, but that there might be some instances where it would be just that two or more towns in the same county be put into a district though not adjacent or very close to each other." Id. at 231. The amendment the delegates passed, however, did not require the legislature to maintain county boundaries. Id.

The convention was unable to pass a resolution to amend Part II, Article 9 before adjourning on June 10, 1964. See id. at 399-402 (history of resolution nos. 4, 19 and 29); Levitt v. Maynard, 105 N.H. 447, 450-51 (1964). The convention reconvened on July 8, 1964, however, following the United States Supreme Court decision in Reynolds. In Reynolds, 377 U.S. at 576, the Court held, for the first time, that the Equal Protection Clause of the Federal Constitution requires both houses of a bicameral state legislature to be apportioned on the basis of population. See Levitt, 105 N.H. at 450-51. When it reconvened, the convention resolved to amend Part II, Article 9 to state that the house of representatives was "founded on principles of equality" and to require that representation in the house be "as equal as circumstances will admit." Journal of Constitutional Convention, supra at 351-53, 355, 358; N.H. CONST. pt. II, art. 9.

In light of the history of the 1964 amendments to Part II, Articles 9 and 11, we hold that these provisions are at least as protective of a citizen’s right to vote as the federal constitutional standard of one person/one vote. Accordingly, we need not undertake a separate federal analysis and we base this decision upon our State Constitution. See State v. Ball, 124 N.H. 226, 233 (1983). We rely upon federal cases interpreting the Federal Constitution only to aid in our analysis. See id.

2. Substantive Requirement of One Person/One Vote

"[T]he overriding objective [of apportionment] must be substantial equality of population among the various [legislative] districts, so that the vote of any citizen is approximately equal in weight to that of any other citizen in the State." Reynolds, 377 U.S. at 579. Although "absolute population equality" need not be achieved, Karcher v. Daggett, 462 U.S. 725, 732-33 (1983), a court devising a remedial apportionment plan for a state legislature "must ordinarily achieve the goal of population equality with little more than de minimis variation." Chapman v. Meier, 420 U.S. 1, 26-27 (1975). "[A]ny deviation from approximate population equality must be supported by enunciation of historically significant state policy or unique features." Id. at 26. In devising a court-ordered remedial apportionment plan, we must also act "circumspectly, and in a manner free from any taint of arbitrariness or discrimination." Connor, 431 U.S. at 415 (quotation omitted).

a. Single- and Multi-Member Districts

To achieve "substantial equality" in an apportionment plan, States generally use two types of districts: single-member districts consisting of one representative elected by the district’s voters, and multi-member districts from which more than one representative are elected. See Burns v. Richardson, 384 U.S. 73, 88 (1966); see also G. Moncrief & R. Joula, When the Courts Don’t Compute: Mathematics and Floterial Districts in Legislative Reapportionment Cases, 4 J.L. & Pol. 737, 741 (1988).

Single-member districts are preferred. See Connor, 431 U.S. at 415; Chapman, 420 U.S. at 19. Use of multi-member districts is constitutionally permissible, however, unless the districts are designed to or would "minimize or cancel out the voting strength of racial or political elements of the voting population," Fortson v. Dorsey, 379 U.S. 433, 439 (1965); see Opinion of the Justices, 111 N.H. at 150-51, or their use "produces deviations from substantial equality beyond the range of constitutional tolerance," Opinion of the Justices, 307 A.2d 198, 209 (Me. 1973) (quotation omitted).

b. Floterials

A third method of representation used in a very few States, including New Hampshire within the last few decades, is a "floterial." See G. Moncrief & R. Joula, supra at 742. A floterial has been described as a district that "floats above" several distinct districts. See id. at 738 (quotation omitted). Floterials, as constructed in New Hampshire, have led to unusual results and voting right inequities.

For example, in Carroll County, the 1992 house reapportionment plan included three floterials: Districts 3, 8 and 10. A map of the 1992 plan for Carroll County is attached to this opinion as Appendix A. Each floterial consisted of some locations that were part of the floterial only and other locations that were part of either a single-member or multi-member district, in addition to the floterial.

Carroll County District 10 was a floterial with two representatives covering four towns, Moultonborough, Sandwich, Tamworth and Tuftonboro, with a total population of 8,029, according to the 1990 census. Moultonborough was also a single-member district with one representative for the town’s total population of 2,956. Thus, voters in Moultonborough voted for three representatives and voters in the other three towns voted for two representatives.

Also in the 1992 plan, Carroll County District 8 consisted of one floterial representative for Wakefield, Wolfeboro and Brookfield. In addition, Wakefield was a single-member district with one representative for the town’s population of 3,057 and Wolfeboro was a single-member district with one representative for the town’s population of 4,807. Because of the floterial, voters in both Wakefield and Wolfeboro voted for two representatives, despite the difference in the population of these towns.

Finally, Carroll County District 3, a floterial, consisted of six communities: Bartlett (population 2,290), Chatham (population 268), Conway (population 7,940), Hale’s Location (population 0), Hart’s Location (population 36), and Jackson (population 678). Using 1990 census figures, the combined population of the six communities in District 3 was 11,212. The floterial representative covered all six communities in District 3. Additionally, four of the six communities in District 3 (Bartlett, Chatham, Hart’s Location and Jackson with a combined total population of 3,272) were also in District 1, a single-member district; the other two communities in District 3 (Conway and Hale’s Location with a combined total population of 7,940) were in District 2, a multi-member district with two representatives. The result of this configuration was that voters in Hart’s Location, population 36, voted for the same number of representatives as the voters in Bartlett, population 2,290.

We select Carroll County as an example because it best illustrates that floterials are usually complicated and often confusing. By contrast, the court plan for Carroll County, attached as Appendix B, simply has one single-member district and four multi-member districts, each with four or fewer representatives. Moreover, as explained more fully in Sections III and IV below, when the towns within a floterial have vastly different populations, the use of the floterial can cause substantial deviations from the one person/one vote principle. See G. Moncrief & R. Joula, supra at 745. While the New Hampshire Constitution specifically contemplates the use of multi-member districts, see N.H. CONST. pt. II, art. 11, it is silent as to floterials.

B. Other State Constitutional Principles

In addition to requiring that representative districts be drawn "as equal as circumstances will admit," the New Hampshire Constitution directs that any apportionment be based upon "the last general census of the inhabitants of the state taken by authority of the United States or of this state." McGovern v. Secretary of State, 138 N.H. 128, 131 (1993) (quotation omitted); N.H. CONST. pt. II, art. 9. The State Constitution also mandates that: (1) the house of representatives be comprised of no fewer than 375 and no more than 400 members; (2) no town, ward or place be divided unless it requests to be divided by referendum; and (3) the boundaries of towns, wards and places be preserved. N.H. CONST. pt. II, arts. 9, 11, 11-a. As previously discussed, the New Hampshire Constitution also requires that the towns, wards and places in a district be "reasonably proximate to one another." N.H. CONST. pt. II, art. 11.

C. Traditional Reapportionment Principles

We also consider the State’s traditional redistricting policy of maintaining county boundaries. See Boyer v. Gardner, 540 F. Supp. 624, 629-30 (D.N.H. 1982). Preserving county boundaries has been important historically because "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." Id. at 630 n.10; RSA 24:1 (2000), :13 (2000), :13-a (2000).

A second consideration is that representative districts have traditionally been comprised of contiguous territories. "A State may legitimately desire to maintain the integrity of various political subdivisions, insofar as possible, and provide for compact districts of contiguous territory in designing a legislative apportionment scheme." Reynolds, 377 U.S. at 578.


III. Determining Whether a Plan Complies with One Person/One Vote

The established method to determine whether a reapportionment plan affords citizens an equal right to vote is to calculate the extent to which the plan deviates from the ideal district population. See New York City Bd. of Estimate v. Morris, 489 U.S. 688, 700, 700-01 n.7 (1989). The first step is to determine the ideal population. To calculate the ideal population of a single-member district, the state population is divided by the total number of state representatives. In New Hampshire, assuming that the house contains 400 members, the ideal population for a single-member district is 3,089 (1,235,786 people divided by 400 representatives). The ideal population for a multi-member district is expressed as a multiple of the ideal population for a single-member district. In New Hampshire, the ideal population for a district with three representatives is 3,089 multiplied by 3, or 9,267.

Once the ideal population is calculated, it is then possible to determine the extent to which a given district population deviates from the ideal. Relative deviation is the most commonly used measure and is derived by dividing the difference between the district’s population and the ideal population by the ideal population.

For example, the relative deviation for a single-member district in New Hampshire with a population of 4,000 is calculated by subtracting 3,089 from 4,000 and dividing the difference (+911) by 3,089. The relative deviation is 29%. For a multi-member district, the relative deviation is calculated using the "aggregate method," which aggregates the total number of representatives and the total population in the district to calculate deviation. Thus, for a district with a population of 8,000 and three representatives, the difference between 8,000 and 3 x 3,089 (9,267) is divided by 9,267, and the relative deviation is –14%.

Using the relative deviation, one can calculate the range of deviation by adding the largest positive deviation and the largest negative deviation without regard to algebraic sign. See Abrams v. Johnson, 521 U.S. 74, 98 (1997). Thus, in the example above, 29% and –14% yields a range of deviation of 43%.


IV. Plans Submitted by the Parties

We have carefully reviewed each plan against the neutral principles set forth above. All of the plans submitted for our consideration suffer from the same flaws. None is appropriate for wholesale adoption by the court. Moreover, none is appropriate for use as the court’s starting point.

First, all of the plans contain erroneous population figures. All of the submitted plans were based upon ward boundaries drawn after the 2000 federal census was conducted. None of the plans identified which boundaries had changed, their location, or the data from which the changes were derived.

Despite repeated requests, the parties did not forward this information to the court until July 1, 2002. After the court received the information, it discovered that there were discrepancies between the ward populations as reported by various cities and as reported in federal census data. This fact alone would require us to reject the parties’ plans. However, the court notified the parties of the discrepancies, and on July 16, 2002, the parties filed a joint stipulation to the accuracy of most of the federal census data. The only portion of the data the parties disputed concerned the federal census data for Manchester wards 5, 6 and 7 as reconfigured after the 2000 census. After reviewing the boundaries used by the city of Manchester for wards 5, 6 and 7, the court learned that the city did not use census block features as ward lines in two areas in these wards. Accordingly, the court used the census block features that were closest to the ward lines set by the city.

Although the parties stipulated to the federal census figures for all other cities with changed wards, the cities of Dover and Claremont filed separate partial objections to the figures for their cities. The court was able to verify the information submitted by Dover, but was not able to verify the information submitted by Claremont. Thus, the court used the information submitted by Dover in constructing its plan, but did not use the information submitted by Claremont.

More importantly, all of the plans miscalculate the extent to which they deviate from the one person/one vote principle. All of the plans rely upon floterials and use the aggregate method to calculate the deviation of the floterials. The aggregate method is appropriate for multi-member districts, but is not appropriate for the floterials in the parties’ plans because it masks substantial deviation from the one person/one vote principle.

For example, in the plan submitted by the house, the towns of Epping (population 5,476) and Fremont (population 3,510) are combined in a floterial with one representative. Each town also constitutes a single-member district and thus each town has its own representative. The plan calculates the deviation as if all three representatives represents both towns together (-3.03%). In fact, each town is represented by one representative as well as a floterial representative. Thus, treating this floterial as if it were simply a three-member district is misleading.

Similarly, in the plans submitted by the speaker of the house, Brentwood, Epping and Fremont are single-member districts and all three are also part of a floterial. The plans calculate the deviation as if the representative for Epping, the representative for Fremont, the representative for Brentwood, and the floterial representative represent all four towns. They do not. In fact, each town is represented by its own representative and also the floterial representative. Again, treating this floterial as if it were simply a four-member district distorts the actual deviation.

The parties rely upon Boyer, 540 F. Supp. 624, as support for their use of the aggregate method. In Boyer, the United States District Court for the District of New Hampshire reviewed the constitutionality of the seventeen floterials included in the 1982 New Hampshire legislative apportionment plan under the Federal Constitution. Id. at 625-26. In assessing the validity of the floterials, the court ruled that it was proper to apply the aggregate method of calculating the range of deviation. Id. at 627-28. Boyer, however, is not binding on this court when we are construing our own constitution. Moreover, we believe that, particularly as used by the parties in this case, the aggregate method obscures substantial deviations from the one person/one vote principle. See G. Moncrief & R. Joula, supra at 745.

No party has argued that, to the extent a plan relies upon floterials, the deviations for the floterials should be calculated using the component method. See Appendix C. Nor has any party relied upon Morris, 489 U.S. at 700-02, 702 n.9, in which the court apparently relied upon a version of the component method to calculate total deviation in districts that had both single-member and at-large representatives. Here, unlike Morris, we have a record devoid of application of the component method to floterials.

Nonetheless, our own calculations indicate that even using the component method, the range of deviation produced by the floterials in the plans submitted is unacceptably high. For example, in the speaker’s plans, the deviation created by the Rockingham District 25 floterial (Brentwood, Epping and Fremont) is -18.1% for Brentwood and +22.3% for Epping, yielding an overall deviation of 40.4%. In the house plan, the deviation created by the Rockingham District 24 floterial (Fremont and Epping) is -18.3% for Fremont and +10.1% for Epping, yielding a total deviation of 28.4%. In the petitioners’ plan, the deviation created by the Rockingham District 9 floterial (Epping, Newfields, Newmarket and Nottingham) for Epping alone is 37.2%. Even without analysis of all of the floterials in each plan, these few examples demonstrate impermissible deviations, which are also far in excess of the deviation in the plan the court establishes today.

Another method for calculating the deviation for a floterial is the same as the method for calculating the deviation for a single-member district; this method results in exceptionally high deviations. Under this method, the ideal population (3,089) is subtracted from the floterial population and the result is divided by the ideal population. For example, a floterial that has a population of 10,000 -- and there are many this size or larger in the plans submitted --would have a deviation of 223%.

Because each plan miscalculates the deviation for floterials, the plans necessarily miscalculate their range of deviation. The properly calculated ranges of deviation for all of the submitted plans significantly exceed "the range of constitutional tolerance." Opinion of the Justices, 307 A.2d at 209. All of the submitted plans thus deviate substantially and impermissibly from the one person/one vote principle.

Further, all of the submitted plans openly embrace political agendas. For instance, in the plan submitted by the speaker, he asserts, over the minority leader’s objection, that one of the districts was created "despite a high deviation and a subsequently necessary floterial, at the request of the Minority Leader." Similarly, in the supporting memorandum submitted by the house, the house notes that certain districts have been apportioned to preserve incumbent seats, the apportionment of one district in Merrimack County "was part of a bi-partisan agreement," and the apportionment of a district in Sullivan County was also "a political agreement." At oral argument, the parties accused each other of crafting apportionment plans to achieve partisan advantage. While political considerations are tolerated in legislatively-implemented redistricting plans, they have no place in a court-ordered plan. See Wilson v. Eu, 823 P.2d 545, 576-77 (Cal. 1992); see also Wyche v. Madison Parish Police Jury, 769 F.2d 265, 268 (5th Cir. 1985) (per curiam).

The degree to which the submitted plans may reflect political considerations is perhaps best illustrated by how each plan treats the same cities and towns differently. For example, in the speaker’s plans, the city of Berlin (population 10,331) constitutes a multi-member district with three representatives and is also part of a two-representative floterial that includes twenty-two other towns (from Bean’s Grant to Whitefield).

By contrast, in the house plan, Berlin has no dedicated representatives. Rather, it is part of two multi-member districts, each with two representatives. The first multi-member district has thirteen other towns and the second has nineteen other towns. The towns in the first multi-member district are also part of the second multi-member district.

In the petitioners’ plan, Berlin, Jefferson, Milan and Randolph are in one multi-member district that has four representatives. And, in the amici plan, Berlin, along with eight other towns, is part of a multi-member district that has four representatives.

Based upon our review of the submitted plans, we conclude that none can be adopted by the court. Each plan relies upon incorrect population data. Each plan miscalculates the overall range of population deviation. Each plan has "calculated partisan political consequences (the details of which are unknown). . . . We have no principled way to choose [among] the plans, especially knowing that we would be endorsing an unknown but intended political consequence by the choice we make." Wilson, 823 P.2d at 576-77.

Accordingly, the court has devised a reapportionment plan consistent with neutral State and federal constitutional principles.


V. Court’s Plan

The court’s plan, which is attached as Appendix D, retains the same number of representatives (400) as in the 1992 house plan. The court’s plan creates eighty-eight representative districts, none of which is a floterial. Five are single-member districts, fourteen are two-member districts, twenty-four are three-member districts, fourteen are four-member districts and thirty-one are districts with more than four representatives. Thus, 65% of the districts have four or fewer representatives.

All of the districts are comprised of contiguous territories. No town, ward or place was divided unless it had requested division by referendum. See N.H. CONST. pt. II, arts. 11, 11-a. Nor were county boundaries crossed in creating the districts.

The court’s plan has a range of deviation of 9.26%, which is dramatically lower than the range of deviation in any of the submitted plans. The plan’s deviation range was derived by adding the deviations of the highest relative positive deviation (Nashua ward 2 at +4.72%) and the highest negative relative deviation (Manchester ward 9 at -4.54%). These deviations were calculated by using the traditional method to calculate the deviations of single-member and multi-member districts. See Morris, 489 U.S. at 700, 700-01 n.7. Given the small population of this State, the unusually large size of its house of representatives, and our State Constitution and traditional redistricting policies, we hold that a deviation range of approximately 9% achieves "substantial equality." Reynolds, 377 U.S. at 579.

New Hampshire has the largest state house of representatives in the country. See Council of State Governments, 33 The Book of the States at 70 (2000). New Hampshire also has one of the smallest state populations in the country. According to the 2000 federal census, New Hampshire ranks 41st in population. See U.S. Census Bureau, Statistical Abstract of the United States: 2001 at 21 (121st ed. 2001). Because New Hampshire has such a large house of representatives (400 members) and such a small population (1,235,786), it takes very few people to affect deviation substantially. For instance, a 10% deviation represents only 309 people, and a 1% deviation represents a mere 31 people.

By contrast, Pennsylvania, with the next largest house of representatives (203) has a much larger population (12,281,054, according to the 2000 census). See Council of State Governments, supra at 70; U.S. Census Bureau, State and County QuickFacts Pennsylvania, available at The ideal district population in Pennsylvania is 60,498 -- about twenty times the size of the ideal district population in New Hampshire. A 10% deviation from the ideal district population in Pennsylvania represents 6,050 individuals -- about twenty times the number of individuals represented by a 10% deviation from the ideal district population in New Hampshire (309 people).

Even Maine, with a population that is similar in size to New Hampshire’s (1,274,923, according to the 2000 census), has a larger ideal district population than does New Hampshire. U.S. Census Bureau, State and County QuickFacts Maine, available at The ideal district population in Maine is 8,443. This is because the size of the Maine House of Representatives is only 151 representatives, compared to New Hampshire’s 400 representatives. See Council of State Governments, supra at 70. A 10% deviation in Maine represents more than 800 people -- almost three times the number of individuals represented by the same deviation from the ideal district in New Hampshire (309 persons).

The court did not use the 1992 house districting plan as its starting point because it was of dubious constitutionality at the time it was passed. The range of deviation for the 1992 plan, using the 1990 census figures, was at least 49.7%. Deviations in this range are too high to be justified by any state interest. See Gaffney v. Cummings, 412 U.S. 735, 744 (1973); Morris, 489 U.S. at 702.

Moreover, the 1992 plan relied heavily upon floterials. Although it ostensibly contained seventy-two single-member districts, only nine of these districts were true single-member districts. The rest were created by floterials. For the reasons stated above, we reject floterials as an unsound redistricting device. The range of deviation calculated above did not include the floterials. Had they been included in the calculation using the component method, the range of deviation would likely have been higher.

The court attempted to create as many true single-member districts as possible, but the mathematical reality is that only a handful of towns have a population that is close to the ideal district population of 3,089.

Although the court’s plan achieves a range of deviation that complies with the one person/one vote principle, the court considered taking steps to reduce the deviation range even further. The court discovered, however, that reducing the range of deviation further required dividing wards into single-member districts. Dividing wards would violate Part II, Article 11-a of the New Hampshire Constitution. Indeed, if the court divided one ward, it would have no principled basis for keeping the boundaries of any other ward intact. In other words, if the court were to ignore the boundaries of one town, ward or place to create a single-member district, it would have no valid reason not to create single-member districts statewide.

Although creating 400 single-member districts statewide would have resulted in little or no deviation, such a radical restructuring of the house was not only not required by the one person/one vote principle, but would have contravened other State constitutional imperatives.

The court endeavored to create multi-member districts that had as few representatives as possible because of its concern that large multi-member districts may tend to dilute the voting strength of racial or political elements of the voting population. See Opinion of the Justices, 111 N.H. at 151. Although we are not called upon, today, to determine the effect of RSA 656:5 (Supp. 2001), requiring majority party candidates for the house to be listed first on all ballots, the number and size of multi-member districts in this plan may justify the concern about the statute raised at oral argument by the amici.

Large multi-member districts exist in the court’s plan for three reasons: (1) the city or town did not request division, see N.H. CONST. pt. II, art. 11-a; (2) a city divided into wards had not properly drawn its ward boundaries; or (3) joining contiguous towns and/or joining contiguous wards was required to produce deviations within the 5% range.

For instance, the court discovered that in the city of Rochester, city officials used features unrecognized by the United States Census Bureau as ward boundaries and used its own enumeration, instead of federal census data, to determine ward populations. "An actual census taken by an individual city . . . [is] not . . . a general census taken by the authority of the United States or of this State," as required by Part II, Article 9 of the New Hampshire Constitution, and may "not be used as a basis for apportionment." Opinion of the Justices, 111 N.H. at 150 (quotation omitted). Accordingly, the court was required to consider the city of Rochester as a whole and did not use the city’s ward boundaries. It is now the largest multi-member district.

Similarly, the court was forced to consider the city of Claremont as a whole, rather than to use its ward boundaries, because the court was not able to verify the populations totals provided by the city for its adjusted wards.

One of the largest multi-member districts, Salem/Windham, has 13 representatives, the same number which the two communities had between them under the 1992 plan.

In Manchester, wards 2, 3, 10 and 11 are combined into an eleven representative multi-member district because it was the only combination of wards that considered in light of all other towns, wards and places, would produce a deviation of less than 5%.

In devising the reapportionment plan, the court did not consider the impact upon either political parties or incumbency.


VI. Conclusion

The court recognizes that its redistricting plan changes house districts significantly. These changes were unavoidable because past house districting plans have not given the fundamental democratic principle of one person/one vote the attention and weight to which it is entitled. The court’s plan reinstates the primacy of this principle and ensures that "the vote of any citizen is approximately equal in weight to that of any other citizen in the State." Reynolds, 377 U.S. at 579.

This plan is effective immediately and the injunction against the house filing period is dissolved as of 12:01 a.m. July 31, 2002. Unless otherwise ordered by the court, the filing of any motion to reconsider shall not stay the effective date of the plan.

So ordered.

BROCK, C.J., and NADEAU, DALIANIS and DUGGAN, JJ., concurred.



Carroll County N.H. House Districts 1992 Appendix A
Carroll County House Districts Court Plan 2002 Appendix B
Component Method Appendix C
N.H. House Districts Court Plan 2002 Appendix D
Manchester Area Map Court Plan 2002 Appendix E
Nashua Area Map Court Plan 2002 Appendix F
Concord Area Map Court Plan 2002 Appendix G
Dover Area Map Court Plan 2002 Appendix H
Population Summary Report Court Plan 2002 Appendix I
Full Geography Report Court Plan 2002 Appendix J