I agree that the Municipal Home Rule Law (§ 10, subd. 1, par. a, cl. [13], sub cl. [b]) permits a Town Supervisor to serve on the County Legislature and requires that he stand separately for election to that body. I cannot, however, subscribe to the court’s decision that the reapportionment plan adopted by the board in this case satisfies the demands of the Equal Protection Clause. In my view, the apportionment standards which apply to the States also apply to municipalities — e.g., counties, cities and towns—which exercise general governmental functions.
As the court’s opinion indicates, Bockland County’s proposal provides that each of its five towns shall constitute a legislative district. Using the smallest town, Stony Point, as a base, it accords each of the other four towns additional seats in the County Legislature in an amount determined by dividing their populations by that of Stony Point and rounding off the result to the nearest whole number. Following this year’s election under the plan, Orangetown’s legislators will represent 12% more people than those from Clarkstown, even though each would be accorded the same vote in the County Legislature.
There is no doubt that, in passing on a plan of apportionment, we are governed by the “ one man, one vote” principle laid down in Reynolds v. Sims (377 U. S. 533) and in the cases which followed it. The majority asserts, however, that the strictures of that principle vary and are to be applied differently depending on the “ level * * * of legislative reapportionment ” involved (opn., p. 315). More specifically, it is stated *319that, although the permissible variation in population between congressional districts is “ almost micrometric, ’ ’ there is a somewhat “ broader scope for permissible deviations ” in the apportionment of State Legislatures and a 1 ‘ still broader scope” where the plan involves “ local, intrastate legislative bodies ” (opn., p. 315).
Significantly, though, the opinion furnishes not a single reason which would justify such a differentiation in the cases of counties, and I suggest that there is no basis either in logic or precedent for the distinction sought to be drawn. On the contrary, the very same rationale which requires that all votes be accorded equal weight in elections to Congress and State Legislatures demands rigorous adherence to the one man, one vote principle in a county-wide election of its legislative body. A resident and voter is undoubtedly concerned with the actions of his county or municipal legislature, as he is with those of Congress or the State Legislature. There is no reason in either case why a citizen’s voting power—his ability to influence the outcome of the election-—■ should depend upon the district in which he happens to live. There is, as the Supreme Court observed in Avery v. Midland County (390 U. S. 474, 481), "little difference, in terms of the application of the Equal Protection Clause and the principles of Reynolds v. Sims, between the exercise of state power through [state] legislatures and its exercise by elected officials in the cities, towns, and counties.” Indeed, in Seaman v. Fedourich (16 N Y 2d 94) and in Iannucci v. Board of Supervisors (20 N Y 2d 244) —both of which antedated the Supreme Court’s decision in the Avery case — our court expressly held that the Reynolds principle applies to “ elective legislative bodies exercising general governmental powers at the municipal level ” (16 N Y 2d, at p. 101; 20 N Y 2d, at p. 249) and rejected proposals for legislative apportionment on the county and municipal levels which failed to accord equality of representation. (See, also, Honig v. Board of Supervisors, 24 N Y 2d 861.)
In support of its conclusion that a lower standard of equality applies to elections of representatives to all local legislative bodies, the court relies on two Supreme Court decisions — Sailors v. Board of Educ. (387 U. S. 105) and Dusch v. Davis *320(387 U. S. 112) — in which plans for the structuring of local governments were upheld. These cages are, however, totally inapposite. Sailors v. Board of Educ., for instance, involved a scheme for the appointment of nonlegislative officials and was not at all concerned with legislative apportionment; and, in the Busch case, all of the legislators were to be elected at large and districts (or, more precisely, boroughs) were to be used ‘ ‘ ‘ merely as the basis of residence for candidates, not for voting or representation ’ ” (387 U. S., at p. 115).1
These cases do, it is true, recognize the need for innovation and flexibility in the organisation of local governments; the one man, one vote doctrine does not preclude the development of new governmental forms to meet urgent metropolitan needs. But there is nothing in the decisions which implies that, in drawing electoral districts for county legislative office, it is permissible to accord the votes of citizens living in one area greater weight than is accorded the votes of citizens in another area. “ Diluting the weight of votes because of place of residence ”, the Supreme Court noted in Reynolds (377 U. S. 533, 566, supra), “impairs basic constitutional rights under the Fourteenth Amendment just as much as invidious discrimination based upon factors such as race [case cited] or economic status [case cited].”
In considering the plan before us, I find it difficult, if not impossible, to accept the majority’s characterization of the 12% deviation infecting the proposed plan as a “ minor population variance” (opn., p. 316). I recognize, of course, that mathematical exactness or precision is not always possible and that the test to be applied in determining the validity of a plan is not solely the size of the percentage deviation. Nevertheless, an attempt must be made to keep that deviation to a miuimnm. As I read the relevant cases, the significant and operative consideration is whether the plan under review reflects such an effort to achieve, ‘ ‘ as nearly * * * as is practicable ’ ’, equality of representation. (Reynolds v. Sims, 377 U. S. 533, 577, supra; see Kirkpatrick v. Preisler, 394 U. S. 526, 531.) As the court declared in the Kirkpatrick case, legislative districts must be *321drawn to allow “ only the limited population variances which are unavoidable despite a good-faith effort to achieve absolute equality, or for which a justification is shown.” (Kirkpatrick v. Preisler, 394 U. S. 526, 531, supra.)
It is the court’s position that, despite the deviation population which the plan creates, factors other than population are equally significant and that the proposal represents a ‘ ‘ bona fide application of the ‘ one man-one vote ’ principle to the needs and circumstances of local government ” which satisfies constitutional requirements (opn., p. 316). This, it seems to me, misconceives the meaning of the phrase, “ good-faith effort,” found in the cases. Even if one were to accept the majority’s premise that the demands of the Equal Protection Clause may be relaxed in dealing with a county apportionment plan, nevertheless, the simple fact remains that “ [population is, of necessity, the starting point for consideration and the controlling criterion for judgment in legislative apportionment controversies.” (Reynolds v. Sims, 377 U. S. 533, 567, supra.)
In the present case, population was neither “ the starting point ” nor “ the controlling criterion ”. It is undisputed and indisputable that the board’s sole consideration in fixing legislative districts was the preservation of existing town lines. It was only after the districts had been created and their lines fixed that any attempt was made to equalize the voting representation— an attempt which could not possibly succeed because of the nature of the plan.
Actually, the 12% deviation neither indicates nor reflects the potential for gross inequality of representation inherent in the scheme. It is a plan not only for the coming election but for those to be held in the future that the court is called upon to approve; in deciding whether the plan meets constitutional standards, we must not blind ourselves to the extremes of discrimination it permits. Thus, even were we convinced that a 12% variation in county government voting power were constitutionally acceptable, the plan before us would be invalid. Since the number of legislators from any town is determined by dividing its population by that of the smallest town and rounding off the result to the nearest whole number, the size of the deviation will vary arbitrarily from year to year, according to the *322size of the fraction to be rounded.2 For example, had the plan been in effect in 1950, the deviation would have been 26%; in 1957, 22%; in 1963, 27%; and in 1966, 26%. The potential for inequality inherent in the plan becomes even more apparent when one considers that an increase of only 3,000 in the population of Stony Point, assuming that the populations of the other towns remain constant, would change the deviation from the present 12% to approximately 47%. In fact, it is theoretically possible for a plan such as the one before us to result in deviations approaching 100%.3 And, since the plan rigidly fixes district lines—providing for no variation—there is no way to compensate for or minimize such extreme deviations.
This, though, does not mean that political subdivisions may not be taken into account when drawing legislative districts. I recognize, of course, that there may be historical and, perhaps, functional reasons for preserving the identity of the towns in a county legislative body—even though no such reasons have been shown in this case and even though the towns would remain independent, co-ordinate governments under any scheme. All that is required is that a- plan be sufficiently flexible to give all the voters of the county an equal voice in its government. Population, as the Reynolds case indicates, need not be the only factor to be considered but it must be the controlling factor (377 U. S., at pp. 579-580):
‘ ‘ [N] either history alone, nor economic or other sorts of group interests, are permissible factors in attempting to justify disparities from population-based representation. Citizens, not history or economic interests, cast votes. Considerations of area alone provide an insufficient justification for deviations from the equal-population principle.”
*323And, underscoring the same thought, the court went on to say (p. 581) :
“ [I]f, even as a result of a clearly rational state policy of according some legislative representation to political subdivisions, population is submerged as the controlling consideration in the apportionment of seats in the particular legislative body, then the right of all of the State’s citizens to cast an effective and adequately weighted vote would be unconstitutionally impaired.”
Here, quite obviously, population was “ submerged as the controlling consideration ” in fixing the boundaries of the legislative districts and, consequently, the right of the county’s citizens “ to cast an effective and adequately weighted vote [has been] unconstitutionally impaired.”
That it is possible to reconcile the one man, one vote principle with the desire to preserve town identity in the County Legislature is demonstrated by a proposal prepared by the League of Women Voters and put forward by several of the plaintiffs. Under this plan, all but three of the 18 members of the legislature would still represent districts located wholly within town lines and, as so apportioned, the maximum population deviation would amount to only 4%. Indeed, the plan approved in Town of Greenburgh v. Board of Supervisors (59 Misc 2d 152, affd. 32 A D 2d 892, affd. 25 N Y 2d 817)—upon which the court relies — illustrates an approach which could have been, but was not, followed here. Although it resulted in a deviation of 14%, slightly more than will result this year from the plan before us, the proposal in Greenburgh was held acceptable because town lines, while accorded weight, were not treated as inviolate. In sharp contrast to the present case, the traditional town boundaries were disregarded where necessary in order to accommodate the essential interest of equal representation.
In sum, then, in view of the fact that the legislative districts were drawn strictly in accordance with town lines and without regard to population, it is evident that there was no effort, good faith or otherwise, to achieve “ as nearly as practicable ” equality of voting power. Since under the plan the weight of a citizen’s vote varies depending on the town in which he lives and this variance is likely to increase in future years, the plan *324deprives the voters of equal protection of the laws and may not be constitutionally sustained.4 However, in view of the imminence of the 1969 election and the fact that the plan achieves far greater equality than the existing Board of Supervisors form of county government, I would approve the plan as an interim measure and require the County Legislature, to be elected next month, to formulate a further plan of apportionment consistent with constitutional standards. (See Honig v. Board of Supervisors, 24 N Y 2d 861, supra.)
The order appealed from should be modified to the extent of directing that a new apportionment plan be adopted by the County Legislature chosen in the 1969 election and, as so modified, affirmed.
Judges Scileppi, Bebgan, Breitel and Gibson concur with Judge Burke ; Chief Judge Fuld dissents in part and votes to modify in a separate opinion in which Judge Jasen concurs.
Order affirmed.
. Indeed, the court pointed out that, “ If a borough’s resident on the council represented in fact only the .borough, residence being only a front, different conclusions might follow ” (387 U. S., at p. 116).
. In other words, if the population of each of the towns happens to approximate an integral multiple of that of Stony Point — as it does under the 1969 census — then, the deviation will be relatively small — for example, as in this case, 12%. More typically, however, the formula will require a rounding off of large fractions and will result, accordingly, in very much greater deviations.
. This may be seen from the following example: if the smallest town, town A, had a population of 10,000, town B had 14,999 and town C had 15,001, then, under the plan, town C would have two representatives and town B would only have one, even though their populations were almost identical. The deviation in such a case would be 99.97%.
. Having reached this conclusion, I find it unnecessary to treat the plaintiffs’ further point that a multimember districting scheme is unconstitutional on the ground that it “ allocates greater voting power to voters in the most populous districts.” (Banzhaf, Multi-Member Electoral Districts—Do They Violate the “One Man, One Vote” Principle, 75 Yale, L. J. 1309, 1337; see Chavis v. Whitcomb, 305 F. Supp. 1364, 1391; cf. Iannucci v. Board of Supervisors, 20 N Y 2d 244, 252, supra.) I merely note that, although the Supreme Court has, on several occasions, upheld the constitutionality of multimember district plans, it has yet to pass upon the arguments raised in the present case. (See, e.g., Burns v. Richardson, 384 U. S. 73; Fortson v. Dorsey, 379 U. S. 433.)