concurring in the result.
Because judicial responsibility requires me, as I see things, to bow to the authority of Reynolds v. Sims, 377 U. S. 533, despite my original and continuing belief that the decision was constitutionally wrong (see my dissenting opinion, 377 U. S., at 589 et seq.), I feel compelled to concur in the Court’s disposition of this case. Even under Reynolds, however, I cannot agree with the rationale, elaborated in Part III of the Court’s opinion, by which Hawaii’s registered voter base is sustained. As I read today’s opinion, registered voter figures are an acceptable basis for apportionment only so long as they *99substantially approximate the results that would be reached under some other type of population-based scheme of apportionment.
Many difficult questions of judgment, relating both to policy and to administrative convenience, must be resolved by a State in determining what statistics to use in establishing its apportionment plan. I would not read Reynolds as precluding a State from apportioning its legislature on any rational basis consistent with Reynolds’ philosophy that “people,” not other interests, must be the basis of state legislative apportionment. I think apportionment on the basis of registered voters is a rational system of this type, and that it is therefore permissible under Reynolds regardless of whether in the particular case it approximates some other kind of a population apportionment.