Martin v. Venables

401 F. Supp. 611 (1975)

David J. MARTIN et al.
v.
Joseph W. VENABLES et al.

Civ. No. B-75-227.

United States District Court, D. Connecticut.

September 8, 1975.

William J. Kupinse, Jr., Bridgeport, Conn., for plaintiffs.

Alexander A. Goldfarb, Hartford, Conn., Isaac J. Murov, Anthony P. Copertino, Jr., Richard P. Gilardi, Charles S. Weidman, Bridgeport, Conn., for defendants; Robert F. Frankel, pro se.

MEMORANDUM OF DECISION

NEWMAN, District Judge.

This suit, challenging the apportionment of councilmen within the town of Stratford, presents initially the question of whether apportionment on the basis of registered voters, rather than total population, violates the one person-one vote requirements of the Fourteenth Amendment.

Plaintiffs are residents and qualified voters in each of the ten voting districts of the Town. Defendants are the town's councilmen, town manager, town clerk, and a political party's town chairman, all of whom have responsibilities for the conduct of elections. The suit is brought under 42 U.S.C. § 1983, with jurisdiction based on 28 U.S.C. § 1343(3). After a motion for a temporary restraining order was denied, the Court held a hearing on August 6, 1975, on plaintiffs' motion for a preliminary injunction. At that hearing, plaintiffs urged that the injunction hearing be combined with a hearing on the merits, Fed.R.Civ.P. 65 (a). Defendants indicated that some factual disputes could be resolved by stipulation, but asserted the right to advise the Court by affidavit of factual matters they wished to contest. The hearing was concluded on the understanding that a decision would be rendered on the merits if appropriate, based on the undisputed facts, but if the Court encountered factual disputes material to resolution of any of the issues in the case, an opportunity for a further hearing on the merits would be afforded. After examining the stipulation of facts and the affidavits submitted by both sides, the Court concludes that the undisputed facts provide an adequate basis for decision, and the case can be decided as if submitted on cross motions for summary judgment.

The Charter of the Town provides that there shall be a town council consisting of eleven members, one to be elected at large, and one from each of ten districts within the Town. Stratford Town Charter, § 2.1.1. The council is required to appoint an election district revision commission within ninety days after the adoption of the Charter, which occurred on July 16, 1964, and to appoint another commission every ten years thereafter. § 7.1.4. The commission is obligated to rearrange the boundaries of the ten election districts "in such manner as to follow most natural geographic divisions" and each district is to include "not more than twelve per centum and not less than eight per centum of the total number of electors who were entitled to cast their votes at the last preceding regular town election." § 7.1.5.

An election district revision commission was appointed and conducted meetings in 1974. At a special meeting of the town council held on December 16, 1974, the commission's report, including a map of ten recommended voting districts, was accepted by a divided vote. An ordinance specifying the election districts in accordance with the council's action became effective June 3, 1975. This suit followed on July 21.

To substantiate their challenge to the new election districts, plaintiffs submitted a map and description of the new districts to Robert W. Marx, the director of a division of the U. S. Bureau of the Census. By affidavit Marx reported the results of an analysis done by his division, applying the census data from the census made on April 1, 1970, to the 1975 election districts of Stratford. The Marx data reflect that an ideal district would contain a population of 4,977, and the districts actually contain populations ranging from 3,406 to 5,929. Thus, if the 1970 population figures are used, the smallest district has 1,571 less population than the average, and the largest exceeds the average by 952. In percentage terms the deviation from the average ranges from -31.6%^[1] to + 19.1%, for a total maximum deviation of 50.7%.^[2] Plaintiffs assert this deviation is far greater than allowed even by recent Supreme Court decisions, see Gaffney v. Cummings, 412 U.S. 735, 93 S. Ct. 2321, 37 L. Ed. 2d 298 (1973) (8% allowed without justification); White v. Regester, 412 U.S. 755, 93 S. Ct. 2332, 37 L. Ed. 2d 314 (1973) (9.9% allowed without justification); Mahan v. Howell, 410 U.S. 315, 93 S. Ct. 979, 35 L. Ed. 2d 320 (1973) (16.4% allowed with justification); Abate v. Mundt, 403 U.S. 182, 91 S. Ct. 1904, 29 L. Ed. 2d 399 (1971) (11.9% allowed with justification). There can be no doubt that if the 1970 census data must be used as the basis for assessing the apportionment of Stratford's voting districts in 1975, the maximum deviation that results would violate Fourteenth Amendment standards.

Defendants contend that, in the circumstances of this case, registered voters are a permissible basis on which to apportion council seats, and that the apportionment is within constitutional standards when tested by the numbers of registered voters in each district. As of June 30, 1973, there were 28,089 registered voters in Stratford; thus an ideal district based on registered voters would contain 2,809 registered voters. The comparable figure as of June 30, 1975, is 2,828. If the 1973 figures are used, the 1975 apportionment produces a deviation of registered voters from the average that ranges from -7.9% to +6.4% for a total maximum deviation of 14.3%. If the figures for registered voters as of June 30, 1975, are used, the deviation ranges from -8.1% to +7.6% for a total maximum deviation of 15.7%. Neither the decision on the merits, nor the appropriateness of a remedy turns on whether the registered voter figures from 1973 or 1975 are used, since the percentages are so similar with either set of figures.

Defendants assert that the deviation resulting from registered voter figures is within acceptable limits and that, if justification is needed, it is supplied by a "rational municipal policy" of recognizing local growth patterns and leaving growth districts with a slightly smaller number of registered voters than the norm. The issues on the merits thus become whether registered voters may be used as a basis for apportionment and if so, whether the deviation in the Stratford apportionment exceeds Fourteenth Amendment limits.

As the Supreme Court has recognized, see Burns v. Richardson, 384 U.S. 73, 91, 86 S. Ct. 1286, 16 L. Ed. 2d 376 (1966), total population figures provided the basis of comparison for apportionment plans in Reynolds v. Sims, 377 U.S. 533, 84 S. Ct. 1362, 12 L. Ed. 2d 506 (1964), and most of the early reapportionment cases. Even in Reynolds v. Sims, supra, the Court was not always consistent in expressing the ultimate value to be served by equality of population. To be sure, there is unequivocal language stating that a legislature "must be apportioned on a population basis." 377 U.S. at 568, 84 S.Ct. at 1385. If what is being equalized pursuant to the Equal Protection Clause is the representation of each elected official, then districts of equal population would be required. And in Reynolds, the Court said the Equal Protection Clause demands "substantially equal state legislative representation for all citizens . . . ." Ibid. However, the opinion later observed that "the overriding objective must be substantial equality of population among the various districts, so that the vote of any citizen is approximately equal in weight to that of any other citizen in the State." 377 U.S. at 579, 84 S.Ct. at 1390 (emphasis added). If the objective is to equalize the weight of every vote, that can be accomplished by establishing districts with substantially equal numbers of voters. If 100 votes are cast in one district and 100 in another, each person voting has 1/100th of a say in the outcome, i. e., the weight of all their votes is equal, regardless of how many people live in those two districts. Apportioning on the basis of votes cast is obviously impractical, since a plan must be adopted before an election. But a count of registered voters is available and generally correlates well with actual votes cast within the subdivisions of a particular state, and certainly within those of a particular town. Reynolds v. Sims really did not choose between equal units of population or equal weight of each vote as the ultimate objective. In fact, the Court at one point recognized the ambiguity by stating, "We realize that it is a practical impossibility to arrange legislative districts so that each one has an identical number of residents, or citizens, or voters." 377 U.S. at 577, 84 S. Ct. at 1390 (emphasis added). In Burns the Court for the first time approved an apportionment plan based on registered voters. In upholding the plan, the Court first observed that the Equal Protection Clause does not require use of total population figures. 384 U.S. at 91, 86 S. Ct. 1286. Citizen population figures were explicitly stated to be a permissible population basis. Burns approved registered voters as a basis but "only" because of a lower court finding that the resulting apportionment was not "substantially different from that which would have resulted from the use of a permissible population [i. e., citizen] basis." 384 U.S. at 93, 86 S.Ct. at 1297.

Whether use of registered voters is permissible only when it produces a plan not substantially different from what a permissible population basis would produce is not entirely clear. Admittedly, the Court in Burns used the word "only." However, the correlation between the registered voter apportionment and one based on citizen population, which is alleged to be required, is not demonstrated in either the Supreme Court opinion, nor in the lower court opinion that made the "finding" that the Supreme Court relied upon. Holt v. Richardson, 238 F. Supp. 468 (D.Hawaii 1965). The District Court did find that one of the State's four counties, Honolulu, would have the same number of senators and nearly the same number of representatives under a plan based on either registered voters or total citizen population, but there was no district by district comparison throughout the State. Indeed, as the Supreme Court recognized, since citizen population figures were "hard to obtain or extrapolate," comparison of the registered voter plan with results from citizen population figures was "difficult." 384 U.S. at 95, 86 S. Ct. 1286. It is not easy to understand how a requirement of correlation between plans based on registered voters and a permissible population basis can be extracted from a case where such correlation was characterized by the Supreme Court as "difficult" and in fact was not made by the lower court.

However, it must be acknowledged that in Burns the Court expressed concerns about the use of registered voters as an apportionment basis. Because voter registration depends on the political activity of those eligible to register, there is a risk that "improper influences" might skew voter registration patterns. 384 U.S. at 92, 86 S. Ct. 1286. The Court also noted that registration figures may fluctuate because of the issues or candidates in a particular election. 384 U.S. at 93, 86 S. Ct. 1286. This last concern apparently assumed a requirement of registration prior to each election, for the Court specifically observed that permanent registration might "contribute to the stability and accuracy of the registered voters figure as an apportionment basis." 384 U.S. at 97, 86 S.Ct. at 1299.

If Burns were the Court's last word on the subject, use of registered voters as an apportionment basis would be a matter of some doubt. See also Ely v. Klahr, 403 U.S. 108, 115, n. 7, 91 S. Ct. 1803, 29 L. Ed. 2d 352 (1971); Ellis v. Mayor & City Council of Baltimore, 352 F.2d 123 (4th Cir. 1965). But the Court's opinion in Gaffney v. Cummings, supra, while not dealing explicitly with apportionment based on registered voters, evidences a significant change in approach. In considering how much of a percentage deviation is acceptable when apportionment is based on total population figures, the Court did not merely, as in Burns, make total population no longer a required basis, it strongly suggested that total population was not even an especially desirable basis. Positing that the weight of a person's vote is what matters, 412 U.S. at 746, 93 S. Ct. 2321, the Court pointed out that population figures are not necessarily a reflection of voters. Population figures, the Court observed, conceal sharp differences among localities within a state as to the number of persons eligible to register. Significantly the Court added, "nor do these figures tell anything at all about the proportion of all those otherwise eligible individuals whose vote cannot be counted or weighed because they either failed to register or failed to vote." 412 U.S. at 747-48, 93 S.Ct. at 2329. It was these shortcomings of total population figures that persuaded the Court not to insist on precise mathematical equality of districts when such figures are used. It would be ironical if registered voter figures were held to be an impermissible basis for apportionment, after the Court so recently stated they are a better reflection of the weight of a person's vote than population figures.

As a matter of democratic theory, it may well be argued that each representative ought to represent approximately the same number of people, without regard to whether those people have voted, have registered, or are even eligible to register. And every state and community is free to apportion according to that principle. But the Fourteenth Amendment, as construed in Burns, plainly does not mandate this political theory. It requires only that no person have the weight of his vote diminished by permitting the votes of others in another district to acquire added weight. When districts include substantially equal numbers of registered voters, it is difficult to see how the weight of anyone's vote has been diminished in a constitutional sense. Those who fail to register and thereby forego added representation have no valid complaint against the apportionment; they have the opportunity to register, at least where, as in Stratford, there has been no claim of any racial discrimination or other improper factor inhibiting registration. Cf. Yelverton v. Driggers, 370 F. Supp. 612, 615 (M.D.Ala.1974). And Connecticut's system of permanent registration, see Conn.Gen.Stat. § 9-35, avoids the risk of year to year fluctuations that Burns cautioned against.

Two practical considerations also support the use of registered voter figures. First, the data are kept current, thereby permitting the equality of an apportionment plan to be tested against figures available for the year in which the election is held. Population figures are available only at ten-year intervals. This factor, however, may import into the use of registered voter figures a requirement of currency. In Burns the Court cautioned against the use of registered voter figures for more than "interim" use. 384 U.S. at 97, 86 S. Ct. 1286. It may well be that an apportionment plan sought to be justified on the basis of registered voters will be vulnerable to constitutional challenge in each election year, as current registered voter figures become available.

The second practical advantage of registered voter figures over population figures is the ease of using the data. Lists of registered voters are maintained by street addresses, making it simple to draw district lines and readily determine how many registered voters are being included. On the other hand, population data from the Bureau of the Census normally disclose the boundaries and population of what are known as census blocks. These units are often too large to form election districts within a town. Stratford, for example, having chosen to form ten election districts, could use whole census blocks if they happened to be of nearly equal population and happened to come in multiples of ten, or if, equally unlikely, all the census blocks in the town could be grouped into ten compact districts of equal population. Otherwise election districts must be formed out of parts of census blocks. When that occurs, it is not clear how many people live in the fragmented blocks. In that event, one must either resort to basic census records, as Marx did, or rely on estimates, see Kapral v. Jepson, 271 F. Supp. 74, 88 (D.Conn.1967). Moreover, boundaries of census blocks may not correlate at all with boundaries of neighborhoods that are entitled to be given significance when towns are divided into election districts.

For these reasons, I conclude that registered voter figures can be used as the basis for apportionment, consistent with the Equal Protection Clause, at least for apportionment within a municipality.^[3]

In any event, even if registered voter figures may not be routinely used, they are permissible here because of the demonstrated inadequacy of available population figures. The official date of the census in Stratford was April 1, 1970, and there has been no population count since then. Yet Stratford has experienced significant population growth, which understandably has not been uniform throughout the community. The parties are in disagreement as to which election districts have grown the most and as to which figures are the most reliable measure of growth. Those disputes need not be resolved. There is no dispute that non-uniform growth has occurred. For example, from May, 1970, to December, 1975, permits for only fourteen new housing units were issued in each of two election districts, while the comparable figures in two others were 937 and 239. No matter what figures are used to estimate numbers of persons per housing unit, it is undisputed that new housing has added as few as about 50 persons in one district, while in another the increase has been at least 2,500 persons. With the average population for an ideal district, based on 1970 population figures, at 4,977, it is clear that disparate increases of this extent justify the decision not to apportion on the basis of 1970 population figures. See Kapral v. Jepson, supra, 271 F.Supp. at 87.^[4]

If registered voter figures may be used in Stratford, the next question is whether the current districts conform to constitutional standards. From the 9.9% total deviation permitted in White v. Regester, supra, without justification and the 11.9% total deviation permitted in Abate v. Mundt, supra, with justification, it appears that a 10% line has been implicitly set and that any greater total deviation requires justification. See White v. Regester, supra, 412 U.S. at 776-77, 93 S. Ct. 2332 (Brennan, J., dissenting). The 14.3% or 15.7% total maximum deviation in this case would therefore appear to require justification,^[5] unless a wider latitude is to be permitted for municipal apportionment.

Neither side has contended that the extent of permissible deviation should be greater in reapportionment litigation affecting a municipality than an entire state. Since the absolute numbers are so much smaller, it might be thought that greater percentage deviation should be allowed, since the arrival or departure of only a small number of people in one district can significantly affect the percentage deviation; if the movement of small numbers of people can significantly vary percentage deviations, there may be less cause for insisting on narrow limits of permissible deviation at the outset. However, there is a countervailing consideration. A given percentage deviation in a district of a town may pose a greater threat to the values protected by the Equal Protection Clause than the same percentage deviation in a district of a state. If, for example, both deviations are 10% below the norm, a person in a district of 900 needs to combine with fewer people to use the effect of his extra voting power than a person in a district of 90,000. On balance, it seems appropriate to apply to municipalities the same limits of percentage deviation applicable to states. Justification is therefore required for the total maximum deviation of about 15% in this case.^[6]

Plaintiffs assert the requisite justification can be found in a policy of recognizing growth patterns by deliberately leaving growth districts with a somewhat smaller number of registered voters than the norm. This contention must be rejected, wholly apart from the unresolved dispute between the parties as to which districts have in fact recently experienced the most growth. If registered voter figures are to be permitted as an apportionment basis partly because of their currency, it would be illogical to permit excessive deviations based on uncertain predictions of growth. Whenever a community chooses to reapportion, it has current voter registration figures available. If it anticipates uneven municipal growth and wishes to apportion on the basis of it, it can always reapportion as often as current registration figures reflect such growth, but it cannot deliberately create election districts with an unduly small number of registered voters and thereby diminish the weight of the votes of those in other districts until such time as the expected growth occurs. Understandably a community may be reluctant to reapportion frequently. Within some reasonable period of time, perhaps ten years, the governmental interest in maintaining constancy of election districts to promote candidate identification and lessen voter confusion may itself justify deviations that occur after a valid apportionment plan has been originally promulgated. See Reynolds v. Sims, supra, 377 U.S. at 583-84, 84 S. Ct. 1362. A town is free to choose between starting with a conforming apportionment and leaving it unchanged for a reasonable period of time despite uneven growth, or reapportioning more frequently to reflect uneven growth. The availability of these alternatives makes impermissible the deliberate creation of small districts in the expectation of predicted growth.

Even if anticipated uneven growth could justify a total maximum deviation beyond 10%, there is no indication in this case that the extent of deviation was held to the minimum necessary to accommodate such growth. In Mahan v. Howell, supra, the Court found a total deviation of 16.4% justified by the state's interest in maintaining the boundaries of political subdivisions, but this conclusion was reached because of uncontradicted evidence that the challenged apportionment plan "`produces the minimum deviation above and below the norm, keeping intact political boundaries. . . .'" 410 U.S. at 326, 93 S.Ct. at 986. Though an affidavit submitted by a member of the Stratford district revision commission asserts that growth patterns were taken into consideration in drawing district lines (a claim disputed by another member of the commission), there is no evidence in the minutes of the commission's meetings nor has any claim been made in any of the papers submitted by the defendants as to how growth projections were used in drawing district lines. Specifically, there is no indication that projections of growth in particular districts were available to, much less considered by, the commission, nor that projections of growth in particular districts were used to determine the number of registered voters by which each of those districts should be permitted to fall below the average size.^[7] A generalized concern by the commission for growth in four districts (even if undisputed) could not justify the small size of those districts without some indication that the projected growth for each district approximated the deviation below the norm.

I conclude that the total maximum deviation of approximately 15% requires justification, and that the alleged justification of allowing for population growth will not suffice in the absence of evidence or even a claim that the deviation below the norm in specific districts was held to the minimum necessary to allow for the projected growth in those districts. The apportionment plan for town council seats is therefore not consistent with the Equal Protection Clause because the weight of a person's vote in those districts with a number of registered voters above the norm has been unwarrantedly diminished.^[8]

There remains the question of the timing and extent of an appropriate remedy. In Reynolds v. Sims, supra, the Court observed that in "awarding or withholding immediate relief, a court is entitled to and should consider the proximity of a forthcoming election and the mechanics and complexities of state election laws . . .." 377 U.S. at 585, 84 S.Ct. at 1394. The election timetable for the November 4, 1975, municipal elections in Connecticut is already under way. The political parties have selected their endorsed candidates, and the time for challengers to qualify for primaries has passed. See Conn.Gen.Stat. §§ 9-391, 9-405, 9-406. While it might be possible for a court to intervene at this stage and either order an at-large election or supervise the preparation of a conforming reapportionment plan, such a course is not advisable in this case. First of all, the disruption of election machinery already in operation ought not to be lightly undertaken. Moreover, the plaintiffs' delay in filing this suit is also some reason for not fashioning a remedy that will affect the November 4 election. The apportionment plan was adopted by the town council on December 16, 1974. While plaintiffs contend that one change of substance was made before the plan became effective on June 3, 1975, it does not appear that an earlier court challenge could not have been brought. In any event, the suit could have been filed promptly after June 3. Plaintiffs justify the delay until July 18 as necessary to have the Census Bureau report on the population of the new districts. Even if that delay was unavoidable (which is doubtful since the Court is considering relief only to correct unwarranted deviations from the norm for registered voters, rather than population), it postponed consideration of this case until well into the election timetable, and the electorate of Stratford as a whole should not be disrupted on that account.

The existing deviation may warrant some relief, but not the drastic alteration of an at-large election nor a prompt judicially-required reapportionment. While the latter technique could probably be accomplished rapidly, it might move some voters into new districts in which they would be denied the opportunity to participate in the selection of candidates, and there simply is not sufficient time to reschedule a new candidate endorsement and primary timetable.

Of course, the duly constituted officials of Stratford may choose to bring their apportionment plan into conformity with federal requirements prior to the November 4 election. The 1975 registration figures disclose that only a slight adjustment would be necessary. Only the 1st and 3rd districts deviate from the norm by more than 5%, and by coincidence these districts are adjacent to each other. It appears that a shift of approximately 100 registered voters, and perhaps less, from the 3rd District into the 1st District would bring the total maximum deviation below 10%. The map discloses a section of the 3rd District that appears to be easily moved into the 1st District. If primary challenges are occurring in either district, it may be preferable to leave the districts as is for the current election. Initial discretion as to the timing of a conforming change is left to the town authorities. If a change during the coming election encounters the objection of lack of local authorizing regulations, that can be remedied by application to the Court for an order approving a proposed change to carry out federal law requirements.

In view of all the circumstances, the Court concludes that an appropriate judicial remedy is to require preparation of a new plan of apportionment, consistent with the standards outlined in this opinion, by December 31, 1976. Initial responsibility for preparing such a plan should rest with the entity contemplated by Stratford's charter, an election district revision commission designated by the town council in accordance with § 7.1.4. of the charter. Only upon failure of such a commission to adopt a conforming plan will this Court intercede more directly.

Accordingly, judgment will enter for the plaintiffs declaring that the present apportionment of councilmanic election districts in the town of Stratford is not consistent with the Equal Protection Clause of the Fourteenth Amendment, and denying all claims for injunctive relief. Jurisdiction will be retained for the implementation of the remedy set forth in this opinion, and any party may apply for further relief as may be necessary to assure implementation of that remedy.

NOTES

[1] Plaintiffs computed a larger deviation below the norm, but their computation erroneously divided the difference in population between the norm and the smallest district by the population of the smallest district, whereas percentage deviation from the norm is obtained by dividing the difference by the norm.

[2] Reapportionment cases have mentioned various measures for assessing the extent of deviation from an ideal plan in which each district has exactly the same population. Among the measures are the average deviation from the norm of all districts, the ratio of the largest district to the smallest district, the smallest percentage of the population that can elect a majority of the seats, the number of districts with a deviation greater than 3% or 5%, and the sum of the smallest and largest deviations from the norm. This last measure, referred to in the decisions as total maximum deviation, is the measure the Supreme Court appears to have used in deciding how much deviation is permissible. See, e. g., Gaffney v. Cummings, 412 U.S. 735, 93 S. Ct. 2321, 37 L. Ed. 2d 298 (1973); Mahan v. Howell, 410 U.S. 315, 93 S. Ct. 979, 35 L. Ed. 2d 320 (1973).

Use of the total maximum deviation carries with it a serious shortcoming. Combining the largest deviation above the norm with the largest below the norm permits a large deviation in one direction to be somewhat minimized by spreading the deviations in the other direction evenly among a large number of districts. For example, one district might be 19% above the norm and nineteen districts might by 1% below the norm. Such an apportionment causes a far greater degree of dilution of the weight of each person's vote in the largest district than a plan under which the largest district is 10% above the norm and the smallest is 10% below the norm, yet the total maximum deviation for both plans is 20%.

If the constitutional vice is dilution of a person's vote by placing too many people in his district, or excessively weighting a person's vote by putting too few persons in his district, then the appropriate measure of unconstitutionality ought to be the largest single deviation above or below the norm. That statistic reflects the extent to which the votes of people in that district have been over or undervalued. Examination of the total maximum deviation does, however, serve to highlight not only how much the weight of a person's vote in the largest district has been diminished compared to the proper weight that would result if all votes were weighted equally, but also how much the weight of that vote has been diminished compared to the weight of a person's vote in the smallest district, thereby reflecting the greatest extent to which one person's vote is weighted more than another person's vote. In any event, the Supreme Court has indicated that total maximum deviation is the measure to be used, and this Court is obligated to use it.

[3] One possible objection to apportionment based on voter registration is the anomalous effect that occurs when registration increases in one district to a greater extent than in other districts. The extra voters who register are diminishing the weight of their votes and the votes of other registered voters in that district. There are two responses to this concern. First, this effect is simply the parallel of what occurs in jurisdictions apportioned on a population basis when one district experiences a more rapid population growth than other districts. Second, a marked increase in the rate of registration in one district compared to other districts may well make an apportionment based on registered voters vulnerable to constitutional challenge in any year when publication of registration figures evidences such a result.

[4] Kapral approved use of registered voter figures on a finding (a) that population figures were out of date and unreliable because of recent growth, and (b) that registered voter figures correlated well with population figures, at least in the year of the census. Neither party in this case has sought to demonstrate whether or not registered voter figures correlated uniformly with population in the election districts in 1970. Plaintiffs have sought to establish that the percentage of the town's population that was registered was roughly the same in each year of the last four decennial censuses. That correlation tells nothing about the relation of registered voters to population among election districts within the town. In light of Cummings, I do not believe such a correlation need be shown. If population figures are as subject to criticism as the Court indicated, there would seem to be no virtue in demonstrating that voter registration figures correlate well with them. See also Pate v. El Paso County, Texas, 337 F. Supp. 95, 100 (W.D.Tex.), aff'd, 400 U.S. 806, 91 S. Ct. 55, 27 L. Ed. 2d 38 (1970) (registered voters used to apportion solely because of finding that presence of military personnel made use of population figures inappropriate).

[5] There is some indication in White v. Regester, supra, that total maximum deviation should not be used as the sole criterion of meeting constitutional standards. White observed that the total maximum deviation was 9.9%, that the average deviation was 1.82%, that only 23 of the 90 districts deviated above or below the norm by more than 3%, and that only three districts deviated by more than 5%. "These deviations," said the Court, did not establish a prima facie case of discrimination. 412 U.S. at 764, 93 S. Ct. 2332.

If this group of deviations are computed for the registered voter figures for Stratford, the 1973 figures disclose a maximum total deviation of 14.3%, an average deviation of 4.68%, 9 of 10 districts with deviations in excess of 3%, and 4 of 10 districts with deviations in excess of 5%. The 1975 figures disclose a maximum total deviation of 15.7%, an average deviation of 3.69%, 5 of 10 districts with deviations greater than 3%, and 2 of 10 districts with deviations greater than 5%. On balance, these alternative measures of deviation are not so slight that the total maximum deviation of approximately 15%, under either the 1973 or 1975 figures, can stand without an inquiry as to justification.

[6] There does not appear to be any reason to permit a greater total maximum deviation to stand without justification in this case than the 10% target figure implied by the Supreme Court as the outer limit when apportionment is based on population. If anything, the permissible deviation when apportionment is based on voter registration might well be smaller, since the data is generally more current and more accurate than population figures, and it was the infirmities in population data that persuaded the Court to permit deviations up to 10% without justification.

[7] The affidavit from a member of the commission asserts that the 1st, 2nd, 8th, and 9th districts are the growth districts. If the 1973 voter registration figures are examined, they disclose that the 1st, 8th, and 9th districts are each between 7% and 8% below the norm, while the 2nd district is only .03% below the norm. If the 1975 figures are used, the 1st district is below the norm by 8%, the 2nd district by 5%, the 9th district by 2%, and the 8th district is above the norm. There is no indication or even a claim that the commission considered population projections that would justify these disparate degrees of deviation below the norm.

[8] The ten election districts from which each of the district councilmen are elected also serve as the electoral base for election to the Planning & Zoning Commission and the Board of Zoning Appeals. These bodies consist of five members, one elected from each of five Planning Districts. The Charter provides that Planning District I shall consist of the First and Second Councilmanic voting districts, Planning District II shall consist of the Third and Fourth Councilmanic voting districts and so on. Stratford Town Charter, §§ 4.2.1, 4.3.1. In this suit plaintiffs also challenge the apportionment of seats on the Planning & Zoning Commission and the Board of Zoning Appeals.

The 1970 census figures disclose that the largest Planning District is 15.5% above the norm, and the smallest is 19.9% below the norm, producing a total maximum deviation of 35.4%, which is excessive if population is the only permissible basis. If registered voters are used, the 1973 figures disclose that the largest Planning District is 5.3% above the norm, and the smallest is 4.3% below the norm. With 1975 figures, the comparable results are 4.6% and 6.0%. The total maximum deviation is 9.6% using 1973 figures and 10.6% using 1975 figures. The closeness of these figures to 10% precludes a finding that the apportionment of seats on either the Planning & Zoning Commission or the Zoning Board of Appeals violates constitutional standards. Cf. White v. Regester, supra.