delivered the opinion of the Court.

Appellant Carol Sosna married Michael Sosna on September 5, 1964, in Michigan. They lived together in New York between October 1967 and August 1971, after which date they separated but continued to live in New York. In August 1972, appellant moved to Iowa with her three children, and the following month she petitioned the District Court of Jackson County, Iowa, for a dissolution of her marriage. Michael Sosna, who had been personally served with notice of the action when he came to Iowa to visit his children, made a special appearance to contest the jurisdiction of the Iowa court. The Iowa court dismissed the petition for lack of jurisdiction, finding that Michael Sosna was not a resident of Iowa and appellant had not been a resident of the State of Iowa for one year preceding the filing of her petition. In so doing the Iowa court applied the provisions of Iowa Code § 598.6 (1973) requiring that the petitioner in such an action be “for the last year a resident of the state.” 1

Instead of appealing this ruling to the Iowa appellate courts, appellant filed a complaint in the United States District Court for the Northern District of Iowa asserting that Iowa’s durational residency requirement for in-*396yoking its divorce jurisdiction violated the United States Constitution. She sought both injunctive and declaratory relief against the appellees in this case, one of which is the State of Iowa,2 and the other of whom is the judge of the District Court of Jackson County, Iowa, who had previously dismissed her petition.

A three-judge court, convened pursuant to 28 U. S. C. §§ 2281, 2284, held that the Iowa durational residency requirement was constitutional. 360 F. Supp. 1182 (1973). We noted probable jurisdiction, 415 U. S. 911 (1974), and directed the parties to discuss “whether the United States District Court should have proceeded to the merits of the constitutional issue presented in light of Younger v. Harris, 401 U. S. 37 (1971) and related cases.” For reasons stated in this opinion, we decide that this case is not moot, and hold that the Iowa dura-tional residency requirement for divorce does not offend the United States Constitution.3

*397I

Appellant sought certification of her suit as a class action pursuant to Fed. Rule Civ. Proc. 23 so that she might represent the “class of those residents of the State of Iowa who have resided therein for a period of less than one year and who desire to initiate actions for dissolution of marriage or legal separation, and who are barred from doing so by the one-year durational residency requirement embodied in Sections 598.6 and 598.9 of the Code of Iowa.”4 The parties stipulated that there were in the State of Iowa “numerous people in the same situation as plaintiff,” that joinder of those persons was impracticable, that appellant’s claims were representative of the class, and that she would fairly and adequately protect the interests of the class. See Rule 23 (a). This stipulation was approved by the District *398Court in a pretrial order.5 After the submission of briefs and proposed findings of fact and conclusions of law by the parties, the three-judge court by a divided vote upheld the constitutionality of the statute.

While the parties may be permitted to waive non-jurisdictional defects, they may not by stipulation invoke the judicial power of the United States in litigation which does not present an actual “case or controversy,” Richardson v. Ramirez, 418 U. S. 24 (1974), and on the record before us we feel obliged to address the question of mootness before reaching the merits of appellant’s claim. At the time the judgment of the three-judge court was handed down, appellant had not yet resided in Iowa for one year, and that court was clearly presented with a case or controversy in every sense contemplated by Art. Ill of the Constitution.6 By the time her case reached this Court, however, appellant had long since satisfied the Iowa durational residency requirement, and Iowa Code § 598.6 (1973) no longer stood as a barrier to her attempts to secure dissolution of her marriage in the Iowa courts.7 This is not an unusual development in a case challenging the validity of a durational residency requirement, for in many cases appellate review *399will not be completed until after the plaintiff has satis- ' fied the residency requirement about which complaint was originally made.

If appellant had sued only on her own behalf, both the fact that she now satisfies the one-year residency requirement and the fact that she has obtained a divorce elsewhere would make this case moot and require dismissal. Alton v. Alton, 207 F. 2d 667 (CA3 1953), dismissed as moot, 347 U. S. 610 (1954); SEC v. Medical Committee for Human Rights, 404 U. S. 403 (1972). But appellant brought this suit as a class action and sought to litigate the constitutionality of the durational residency requirement in a representative capacity. When the District Court certified the propriety of the class action, the class of unnamed persons described in the certification acquired a legal status separate from the interest asserted by appellant.8 We are of the view that this factor significantly affects the mootness determination.

In Southern Pacific Terminal Co. v. ICC, 219 U. S. 498 (1911), where a challenged ICC order had expired, and in Moore v. Ogilvie, 394 U. S. 814 (1969), where petitioners sought to be certified as candidates in an election that had already been held, the Court expressed its concern that the defendants in those cases could be expected again to act contrary to the rights asserted by the particular named plaintiffs involved, and in each case the controversy was held not to be moot because the questions presented were “capable of repetition, yet *400evading review.” That situation is not presented in appellant’s case, for the durational residency requirement enforced by Iowa does not at this time bar her from the Iowa courts. Unless we were to speculate that she may move from Iowa, only to return and later seek a divorce within one year from her return, the concerns that prompted this Court’s holdings in Southern Pacific and Moore do not govern appellant’s situation. But even though appellees in this proceeding might not again enforce the Iowa durational residency requirement against appellant, it is clear that they will enforce it against those persons in the class that appellant sought to represent and that the District Court certified. In this sense the case before us is one in which state officials will undoubtedly continue to enforce the challenged statute and yet, because of the passage of time, no single challenger will remain subject to its restrictions for the period necessary to see such a lawsuit to its conclusion.

This problem was present in Dunn v. Blumstein, 405 U. S. 330 (1972), and was there implicitly resolved in favor of the representative of the class. Respondent Blumstein brought a class action challenging the Tennessee law which barred persons from registering to vote unless, at the time of the next election, they would have resided in the State for a year and in a particular county for three months. By the time the District Court opinion was filed, Blumstein had resided in the county for the requisite three months, and the State contended that his challenge to the county requirement was moot. The District Court rejected this argument, Blumstein v. Ellington, 337 F. Supp. 323, 324-326 (MD Tenn. 1970). Although the State did not raise a mootness argument in this Court, we observed that the District Court had been correct:

“Although appellee now can vote, the problem to voters posed by the Tennessee residence require*401ments is ' “capable of repetition, yet evading review.” ’ ” 405 XJ. S., at 333 n. 2.

Although the Court did not expressly note the fact, by the time it decided the case Blumstein had resided in Tennessee for far more than a year.

The rationale of Dunn controls the present case. Although the controversy is no longer live as to appellant Sosna, it remains very much alive for the class of persons she has been certified to represent. Like the other voters in Dunn, new residents of Iowa are aggrieved by an allegedly unconstitutional statute enforced by state officials. We believe that a case such as this, in which, as in Dunn, the issue sought to be litigated escapes full appellate review at the behest of any single challenger, does not inexorably become moot by the intervening resolution of the controversy as to the named plaintiffs.9 Dunn, supra; Rosario v. Rockefeller, 410 U. S. 752, 756 n. 5 (1973); Vaughan v. Bower, 313 F. Supp. 37, 40 (Ariz.), aff’d, 400 U. S. 884 (1970).10 We note, how*402ever, that the same exigency that justifies this doctrine serves . o identify its limits. In cases in which the alleged harm would not dissipate during the normal time required for resolution of the controversy, the general principles of Art. Ill jurisdiction require that the plaintiff's personal stake in the litigation continue throughout the entirety of the litigation.

Our conclusion that this case is not moot in no way detracts from the firmly established requirement that the judicial power of Art. Ill courts extends only to “cases and controversies” specified in that Article. There must not only be a named plaintiff who has such a case or controversy at the time the complaint is filed, and at the time the class action is certified by the District Court pursuant to Rule 23,11 but there must be a live controversy at the time this Court reviews the case.12 SEC v. Medical Committee for Human Rights, supra. The controversy may exist, however, between a named defendant and a member of the class represented by the named plaintiff, even though the claim of the named plaintiff has become moot.

In so holding, we disturb no principles established by our decisions with respect to class-action litigation. A *403named plaintiff in a class action must show that the threat of injury in a case such as this is “real and immediate,” not “conjectural” or “hypothetical.” O’Shea v. Littleton, 414 U. S. 488, 494 (1974); Golden v. Zwickler, 394 U. S. 103, 109-110 (1969). A litigant must be a member of the class which he or she seeks to represent at the time the class action is certified by the district court. Bailey v. Patterson, 369 U. S. 31 (1962); Rosario, supra; Hall v. Beals, 396 U. S. 45 (1969). Appellant Sosna satisfied these criteria.

This conclusion does not automatically establish that appellant is entitled to litigate the interests of the class she seeks to represent, but it does shift the focus of examination from the elements of justiciability to the ability of the named representative to “fairly and adequately protect the interests of the class.” Rule 23 (a). Since it is contemplated that all members of the class will be bound by the ultimate ruling on the merits, Rule 23 (c)(3), the district court must assure itself that the named representative will adequately protect the interests of the class. In the present suit, where it is unlikely that segments of the class appellant represents would have interests conflicting with those she has sought to advance,13 and where the interests of that class have been competently urged at each level of the proceeding, we believe that the test of Rule 23 (a) is met. We therefore address ourselves to the merits of appellant’s constitutional claim.

*404II

The durational residency requirement under attack in this case is a part of Iowa's comprehensive statutory regulation of domestic relations, an area that has long been regarded as a virtually exclusive province of the States. Cases decided by this Court over a period of more than a century bear witness to this historical fact. In Barber v. Barber, 21 How. 582, 584 (1859), the Court said: “We disclaim altogether any jurisdiction in the courts of the United States upon the subject of divorce . . . .” In Pennoyer v. Neff, 95 U. S. 714, 734-735 (1878), the Court said: “The State . . . has absolute right to prescribe the conditions upon which the marriage relation between its own citizens shall be created, and the causes for which it may be dissolved,” and the same view was reaffirmed in Simms v. Simms, 175 U. S. 162, 167. (1899).

The statutory scheme in Iowa, like those in other States, sets forth in considerable detail the grounds upon which a marriage may be dissolved and the circumstances in which a divorce may be obtained. Jurisdiction over a petition for dissolution is established by statute in “the county where either party resides,” Iowa Code § 598.2 (1973), and the Iowa courts have construed the term . “resident” to have much the same meaning as is ordinarily associated with the concept of domicile. Korsrud v. Korsrud, 242 Iowa 178, 45 N. W. 2d 848 (1951). Iowa has recently revised its divorce statutes, incorporating the no-fault concept,14 but it retained the one-year dura-tional residency requirement.

The imposition of a durational residency requirement for divorce is scarcely unique to Iowa, since 48 States impose such a requirement as a condition for maintaining *405an action for divorce.15 As might be expected, the periods vary among the States and range from six weeks16 to two years.17 The one-year period selected by Iowa is the most common length of time prescribed.18

Appellant contends that the Iowa requirement of one year’s residence is unconstitutional for two separate reasons: first, because it establishes two classes of persons and discriminates against those who have recently exercised their right to travel to Iowa, thereby contravening the Court’s holdings in Shapiro v. Thompson, 394 U. S. 618 (1969); Dunn v. Blumstein, 405 U. S. 330 (1972); and Memorial Hospital v. Maricopa County, 415 U. S. 250 (1974); and, second, because it denies a litigant the opportunity to make an individualized showing of bona fide residence and therefore denies such residents access to the only method of legally dissolving their marriage. Vlandis v. Kline, 412 U. S. 441 (1973); Boddie v. Connecticut, 401 U. S. 371 (1971).

*406State statutes imposing durational residency requirements were, of course, invalidated when imposed by States as a qualification for welfare payments, Shapiro, supra; for voting, Dunn, supra; and for medical care, Maricopa County, supra. But none of those cases intimated that the States might never impose durational residency requirements, and such a proposition was in fact expressly disclaimed.19 What those cases had in common was that the durational residency requirements they struck down were justified on the basis of budgetary or recordkeeping considerations which were held insufficient to outweigh the constitutional claims of the individuals. But Iowa’s divorce residency requirement is of a different stripe. Appellant was not irretrievably foreclosed from obtaining some part of what she sought, as was the case with the welfare recipients in Shapiro, the voters in Dunn, or the indigent patient in Maricopa County. She would eventually qualify for the same sort of adjudication which she demanded virtually upon her arrival in the State. Iowa’s requirement delayed her access to the courts, but, by fulfilling it, she could ultimately have obtained the same opportunity for adjudication which she asserts ought to have been hers at an earlier point in time.

Iowa’s residency requirement may reasonably be justified on grounds other than purely budgetary considerations or administrative convenience. Cf. Kahn v. Shevin, 416 U. S. 351 (1974). A decree of divorce is not a matter in which the only interested parties are the State as a sort of “grantor,” and a divorce petitioner such as appellant in the role of “grantee.” Both spouses are obviously interested in the proceedings, since it will affect their marital status and very likely their property rights. Where a married couple has minor children, a decree of *407divorce would usually include provisions for their custody and support. With consequences of such moment riding on a divorce decree issued by its courts, Iowa may insist that one seeking to initiate such a proceeding have the modicum of attachment to the State required here.

Such a requirement additionally furthers the State’s parallel interests both in avoiding officious intermeddling in matters in which another State has a paramount interest, and in minimizing the susceptibility of its own divorce decrees to collateral attack. A State such as Iowa may quite reasonably decide that it does not wish to become a divorce mill for unhappy spouses who have lived there as short a time as appellant had when she commenced her action in the state court after having long resided elsewhere. Until such time as Iowa is convinced that appellant intends to remain in the State, it lacks the “nexus between person and place of such permanence as to control the creation of legal relations and responsibilities of the utmost significance.” Williams v. North Carolina, 325 U. S. 226, 229 (1945). Perhaps even more important, Iowa’s interests extend beyond its borders and include the recognition of its divorce decrees by other States under the Full Faith and Credit Clause of the Constitution, Art. IV, § 1. For that purpose, this Court has often stated that “judicial power to grant a divorce — jurisdiction, strictly speaking — is founded on domicil.” Williams, supra; Andrews v. Andrews, 188 U. S. 14 (1903); Bell v. Bell, 181 U. S. 175 (1901). Where a divorce decree is entered after a finding of domicile in ex parte proceedings,20 this Court has held that the *408finding of domicile is not binding upon another State and may be disregarded in the face of “cogent evidence” to the contrary. Williams, supra, at 236. For that reason, the State asked to enter such a decree is entitled to insist that the putative divorce petitioner satisfy something more than the bare minimum of constitutional requirements before a divorce may be granted. The State’s decision to exact a one-year residency requirement as a matter of policy is therefore buttressed by a quite permissible inference that this requirement not only effectuates state substantive policy but likewise provides a greater safeguard against successful collateral attack than would a requirement of bona fide residence alone.21 This is precisely the *409sort of determination that a State in the exercise of its domestic relations jurisdiction is entitled to make.

We therefore hold that the state interest in requiring that those who seek a divorce from its courts be genuinely-attached to the State, as well as a desire to insulate divorce decrees from the likelihood of collateral attack, requires a different resolution of the constitutional issue presented than was the case in Shapiro, supra, Dunn, supra, and Maricopa County, supra.

Nor are we of the view that the failure to provide an individualized determination of residency violates the Due Process Clause of the Fourteenth Amendment. Vlandis v. Kline, 412 U. S. 441 (1973), relied upon by appellant, held that Connecticut might not arbitrarily invoke a permanent and irrebuttable presumption of non-residence against students who sought to obtain in-state tuition rates when that presumption was not necessarily or universally true in fact. But in Vlandis the Court warned that its decision should not “be construed to deny a State the right to impose on a student, as one element in demonstrating bona fide residence, a reasonable durational residency requirement.” Id., at 452. See Starns v. Malkerson, 326 F. Supp. 234 (Minn. 1970), aff’d, 401 U. S. 985 (1971). An individualized determination of physical presence plus the intent to remain, which appellant apparently seeks, would not entitle her to a divorce even if she could have made such a showing.22 For *410Iowa requires not merely “domicile” in that sense, but residence in the State for a year in order for its courts to exercise their divorce jurisdiction.

In Boddie v. Connecticut, supra, this Court held that Connecticut might not deny access to divorce courts to those persons who could not afford to pay the required fee. Because of the exclusive role played by the State in the termination of marriages, it was held that indigents could not be denied an opportunity to be heard “absent a countervailing state interest of overriding significance.” 401 U. S., at 377. But the gravamen of appellant Sosna’s claim is not total deprivation, as in Boddie, but only delay. The operation of the filing fee in Boddie served to exclude forever a certain segment of the population from obtaining a divorce in the courts of Connecticut. No similar total deprivation is present in appellant’s case, and the delay which attends the enforcement of the one-year durational residency requirement is, for the reasons previously stated, consistent with the provisions of the United States Constitution.

Affirmed.

dissenting.

It is axiomatic that Art. Ill of the Constitution imposes a “threshold requirement . . . that those who seek to invoke the power of federal courts must allege an actual case or controversy.” O’Shea v. Littleton, 414 U. S. 488, 493 (1974); Flast v. Cohen, 392 U. S. 83, 94-101 (1968); Jenkins v. McKeithen, 395 U. S. 411, 421-425 (1969) (opinion of Marshall, J.). To satisfy the requirement, plaintiffs must allege “some threatened or actual injury,” Linda R. S. v. Richard D., 410 U. S. 614, 617 (1973), that is “real and immediate” and not con*411jectural or hypothetical. Golden v. Zwickler, 394 U. S. 103, 108-109 (1969); Maryland Casualty Co. v. Pacific Coal & Oil Co., 312 U. S. 270, 273 (1941); Public Workers v. Mitchell, 330 U. S. 75, 89-91 (1947). Furthermore, and of greatest relevance here:

“The fundamental aspect of standing is that it focuses on the party seeking to get his complaint before a federal court and not on the issues he wishes to have adjudicated. The 'gist of the question of standing’ is whether the party seeking relief has ‘alleged such a personal stake in the outcome of the controversy as to assure that concrete adverseness which sharpens the presentation of issues upon which the court so largely depends for illumination of difficult constitutional questions.’ Baker v. Carr, 369 U. S. 186, 204 (1962). In other words, when standing is placed in issue in a case, the question is whether the person whose standing is challenged is a proper party to request an adjudication of a particular issue and not whether the issue itself is justi-ciable.” Flast v. Cohen, supra, at 99-100 (footnote omitted).

All of this the Court concedes. It is conceded as well that had the named plaintiff in this case not brought a class action, the case would now be dismissed as moot because the plaintiff, appellant here, has now satisfied the Iowa residency requirement and, what is more, has secured a divorce in another State. Appellant could not have begun this suit either for herself or for a class if at the time of filing she had been an Iowa resident for a year or had secured a divorce in another jurisdiction. There must be a named plaintiff initiating the action who has an existing controversy with the defendant, whether the plaintiff is suing on his own behalf or on behalf of a class as well. However unquestioned it may *412be that a class of persons in the community has a “real” dispute of substance with the defendant, an attorney may not initiate a class action without having a client with a personal stake in the controversy who is a member of the class, and who is willing to be the named plaintiff in the case. The Court recently made this very clear when it said that “if none of the named plaintiffs purporting to represent a class establishes the requisite of a case or controversy with the defendants, none may seek relief on behalf of himself or any other member of the class.” O’Shea v. Littleton, supra, at 494 (footnote omitted).

The Court nevertheless holds that once a case is certified as a class action, the named plaintiff may lose that status which had qualified him to bring the suit and still be acceptable as a party to prosecute the suit to conclusion on behalf of the class. I am unable to . agree. The appellant now satisfies the Iowa residence requirement and has secured a divorce. She retains no real interest whatsoever in this controversy, certainly not. an interest that would have entitled her to be a plaintiff in the first place, either alone or as representing a class. In reality, there is no longer a named plaintiff in the case, no member of the class before the Court. The unresolved issue, the attorney, and a class of unnamed litigants remain. None of the anonymous members of the class is present to direct counsel and ensure that class interests are being properly served. For all practical purposes, this case has become one-sided and has lost the adversary quality necessary to satisfy the constitutional “case or controversy” requirement. A real issue unquestionably remains, but the necessary adverse party to press it has disappeared.

The Court thus dilutes the jurisdictional command of Art. Ill to a mere prudential guideline. The only specific, identifiable individual with an evident continuing *413interest in presenting an attack upon- the residency-requirement is appellant’s counsel. The Court in reality holds that an attorney’s competence in presenting his case, evaluated post hoc through a review of his performance as revealed by the record, fulfills the “case or controversy” mandate. The legal fiction employed to cloak this reality is the reification of an abstract entity, “the class,” constituted of faceless, unnamed individuals who are deemed to have a live case or controversy with appellees.1

*414No prior decision supports the Court’s broad rationale. In cases in which the inadequacy of the named representative’s claim has become apparent prior to class certifi-fication, the Court has been emphatic in rejecting the argument that the class action could still be pursued. O’Shea v. Littleton, supra, at 494-495; Bailey v. Patterson, 369 U. S. 31, 32-33 (1962). Cf. Richardson v. Ramirez, 418 U. S. 24 (1974); Hall v. Beals, 396 U. S. 45, 48-49 (1969).

It is true that Dunn v. Blumstein, 405 U. S. 330, 333 n. 2 (1972), looks in the other direction. There, by the time the Court rendered its decision, the class representative in an action challenging a durational residency requirement for voting had satisfied the requirement and was eligible to vote in the next election. The Court indicated that the case was not moot, saying that the issue was “capable of repetition, yet evading review.” But the question was not contested between the parties and was noted only in passing. Its ramifications for the question of mootness in a class action setting were not explored. Although I joined the opinion in that case, I do not deem it dispositive of the jurisdictional issue here, especially in light of Indiana Employment Division v. Burney, 409 U. S. 540 (1973). There the class representative’s claim had been fully settled, and the Court remanded the case to the District Court for consideration of mootness, a course which the majority, relying on Dunn, rejects here. As I see it, the question of whether a class action survives after the representative’s claim has been mooted remains unsettled by prior decisions. Indeed, what authority there is provides more support for a conclusion that when the personal stake of the named plaintiff terminates, the class action fails.

*415Although the Court cites Dunn v. Blumstein, supra, as controlling authority, the principal basis for its approach is a conception of the class action that substantially dissipates the case-or-controversy requirement as well as the necessity for adequate representation under Fed. Rule Civ. Proc. 23(a)(4). In the Court’s view, the litigation before us is saved from mootness only by the fact that class certification occurred prior to appellant’s change in circumstance. In justification, the Court points to two significant consequences of certification. First, once certified, the class action may not be settled or dismissed without the district court’s approval. Second, if the action results in a judgment on the merits, the decision will bind all members found at the time of certification to be members of the class. These are significant aspects of class-action procedure, but it is not evident and not explained how and why these procedural consequences of certification modify the normal mootness considerations which would otherwise attach. Certification is no substitute for a live' plaintiff with a personal interest in the case sufficient to make it an adversary proceeding. Moreover, certification is not irreversible or inalterable; it “may be conditional, and may be altered or amended before the decision on the merits.” Rule 23(c)(1).2 Furthermore, under Rule 23(d) the court may make various types of orders in conducting the litigation, including an order that notice be given “of the opportunity of members to signify whether they consider the representation fair and adequate, to intervene and present claims or defenses, or otherwise to come into the action” and “requiring that the pleadings be amended to eliminate therefrom allegations as to representation *416of absent persons . 3 Class litigation is most often characterized by its complexity and concomitant flexibility of a court in managing it, and emphasis upon one point in the process flies in the face of that reality.

The new certification procedure of Rule 23 (c)(1), as amended in 1966, was not intended to modify the strictures of Fed. Rule Civ. Proc. 82 that “[t]hese rules shall not be construed to extend ... the jurisdiction of the United States district courts . . ..” Cf. Snyder v. Harris, 394 U. S. 332, 337-338 (1969). The intention behind the certification amendment, which had no counterpart in the earlier version of the rule, was merely “to give clear definition to the action . . . ,” Advisory Committee Note, 28 U. S. C. App., p. 7767; 3B J. Moore, Federal Practice ¶ 23.50, pp. 23-1101 to 23-1102 (1974), not as the Court would now have it, to avoid jurisdictional problems of mootness.4

It is claimed that the certified class supplies the necessary adverse parties for a continuing case or controversy *417with appellees. This is not true; but even if it were, the Court is left with the problem of determining whether the class action is still a good one and whether under Rule 23 (a) (4) appellant is a fair and adequate representative of the class. That appellant can no longer in any realistic sense be considered a member of the class makes these determinations imperative. The Court disposes of the problem to its own satisfaction by saying that it is unlikely that segments of the class appellant represents would have conflicting interests with those she has sought to advance and that because the interests of the class have been competently urged at each level of the proceeding the test of Rule 23 (a) (4) is met. The Court cites no authority for this retrospective decision as to the adequacy of representation which seems to focus on the competence of counsel rather than a party plaintiff who is a representative member of the class.5 At the very least, the case should be remanded to the District Court where these considerations could be explored and the desirability of issuing orders under Rule 23 (d) to protect the class might be considered.

The Court’s refusal to remand for consideration of mootness and adequacy of representation can be explained only by its apparent notion that there may be categories of issues which will permit lower courts to pass upon them but which by their very nature will become moot before this Court can address them. Thus it is said that "no single challenger will remain subject to [the residency requirement] for the period necessary to see such a lawsuit to its conclusion.” Ante, at 400. Hence, *418the Court perceives the need for a general rule which will eliminate the problem. Article III, however, is an “awkward” limitation. It prevents all federal courts from addressing some important questions; there is nothing surprising in the fact that it may permit only the lower federal courts to address other questions. Article III is not a rule always consistent with judicial economy. Its overriding purpose is to define the boundaries separating the branches and to keep this Court from assuming a legislative perspective and function. See Flast v. Cohen, 392 U. S. 83, 96 (1968). The ultimate basis of the Court’s decision must be a conclusion that the issue presented is an important and recurring one which should be finally resolved here. But this notion cannot override constitutional limitations.

Because I find that the case before the Court has become moot, I must respectfully dissent.

dissenting.

The Court today departs sharply from the course we have followed in analyzing durational residency requirements since Shapiro v. Thompson, 394 U. S. 618 (1969). Because I think the principles set out in that case and its progeny compel reversal here, I respectfully dissent.

As we have made clear in Shapiro and subsequent cases, any classification that penalizes exercise of the constitutional right to travel is invalid unless it is justified by a compelling governmental interest. As recently as last Term we held that the right to travel requires that States provide the same vital governmental benefits and privileges to recent immigrants that they do to longtime residents. Memorial Hospital v. Maricopa County, 415 U. S. 250, 261 (1974). Although we recognized that not all durational residency requirements are penalties *419upon the exercise of the right to travel interstate,1 we held that free medical aid, like voting, see Dunn v. Blumstein, 405 U. S. 330 (1972), and welfare assistance, see Shapiro v. Thompson, supra, was of such fundamental importance that the State could not constitutionally condition its receipt upon long-term residence. After examining Arizona’s justifications for restricting the availability of free medical services, we concluded that the State had failed to show that in pursuing legitimate objectives it had chosen means that did not impinge unnecessarily upon constitutionally protected interests.

The Court’s failure to address the instant case in these terms suggests a new distaste for the mode of analysis we have applied to this corner of equal protection law. In its stead, the Court has employed what appears to be an ad hoc balancing test, under which the State’s putative interest in ensuring that its divorce petitioners establish some roots in Iowa is said to justify the one-year residency requirement. I am concerned not only about the disposition of this case, but also about the implications of the majority’s analysis for other divorce statutes and for durational residency requirement cases in general.

I

The Court omits altogether what should be the first inquiry: whether the right to obtain a divorce is of sufficient importance that its denial to recent immigrants constitutes a penalty on interstate travel. In my view, it clearly meets that standard. The previous decisions of this Court make it plain that the right of marital association is one of the most basic rights conferred on the individual by the State. The interests associated *420with marriage and divorce have repeatedly been accorded particular deference, and the right to marry has been termed “one of the vital personal rights essential to the orderly pursuit of happiness by free men.” Loving v. Virginia, 388 U. S. 1, 12 (1967). In Boddie v. Connecticut, 401 U. S. 371 (1971), we recognized that the right to seek dissolution of the marital relationship was closely related to the right to marry, as both involve the voluntary adjustment of the same fundamental human relationship. Id., at 383. Without further laboring the point, I think it is clear beyond cavil that the right to seek dissolution of the marital relationship is of such fundamental importance that denial of this right to the class of recent interstate travelers penalizes interstate travel within the meaning of Shapiro, Dunn, and Maricopa County.

II

Having determined that the interest in obtaining a divorce is of substantial social importance, I would scrutinize Iowa’s durational residency requirement to determine whether it constitutes a reasonable means of furthering important interests asserted by the State. The Court, however, has not only declined to apply the “compelling interest” test to this case, it has conjured up possible justifications for the State’s restriction in a manner much more akin to the lenient standard we have in the past applied in analyzing equal protection challenges to business regulations. See McGowan v. Maryland, 366 U. S. 420, 425-428 (1961); Kotch v. Board of River Port Pilot Comm’rs, 330 U. S. 552, 557 (1947); but see Johnson v. Robison, 415 U. S. 361, 376 (1974). I continue to be of the view that the “rational basis” test has no place in equal protection analysis when important individual interests with constitutional implications are at stake, see San Antonio School District v. Rodriguez, 411 *421U. S. 1, 109 (1973) (Marshall, J., dissenting); Dandridge v. Williams, 397 U. S. 471, 520-522 (1970) (Marshall, J., dissenting). But whatever the ultimate resting point of the current readjustments in equal protection analysis, the Court has clearly directed that the proper standard to apply to cases in which state statutes have penalized the exercise of the right to interstate travel is the “compelling interest” test. Shapiro v. Thompson, 394 U. S., at 634, 638; Oregon v. Mitchell, 400 U. S. 112, 238 (1970) (opinion of Brennan, White, and Marshall, JJ.); Dunn v. Blumstein, 405 U. S., at 342-343; Memorial Hospital v. Maricopa County, 415 U. S., at 262-263.

The Court proposes three defenses for the Iowa statute: first, the residency requirement merely delays receipt of the benefit in question — it does not deprive the applicant of the benefit altogether; second, since significant social consequences may follow from the conferral of a divorce, the State may legitimately regulate the divorce process; and third, the State has interests both in protecting itself from use as a “divorce mill” and in protecting its judgments from possible collateral attack in other States. In my view, the first two defenses provide no significant support for the statute in question here. Only the third has any real force.

A

With the first justification, the Court seeks to distinguish the Shapiro, Dunn, and Maricopa County cases. Yet the distinction the Court draws seems to me specious. Iowa’s residency requirement, the Court says, merely forestalls access to the courts; applicants seeking welfare payments, medical aid, and the right to vote, on the other hand, suffer unrecoverable losses throughout the waiting period. This analysis, however, ignores the severity of the deprivation suffered by the divorce petitioner who is forced to wait a year for relief. See Stanley v. Illinois, *422405 U. S. 645, 647 (1972). The injury accompanying that delay is not directly measurable in money terms like the loss of welfare benefits, but it cannot reasonably be argued that when the year has elapsed, the petitioner is made whole. The year’s wait prevents remarriage and locks both partners into what may be an intolerable, destructive relationship. Even applying the Court’s argument on its own terms, I fail to see how the Maricopa County case can be distinguished. A potential patient may well need treatment for a single ailment. Under Arizona statutes he would have had to wait a year before he could be treated. Yet the majority’s analysis would suggest that Mr. Evaro’s claim for nonemergency medical aid is not cognizable because he would “eventually qualify for the same sort of [service],” ante, at 406. The Court cannot mean that Mrs. Sosna has not suffered any injury by being foreclosed from seeking a divorce in Iowa for a year. It must instead mean that it does not regard that deprivation as being very severe.2

B

1 find the majority’s second argument no more persuasive. The Court forgoes reliance on the usual justifications for durational residency requirements — budgetary considerations and administrative convenience, see Shapiro, 394 U. S., at 627-638; Maricopa County, 415 U. S., at 262-269. Indeed, it would be hard to make a persuasive argument that either of these interests is significantly *423implicated in this case. In their place, the majority invokes a more amorphous justification — the magnitude of the interests affected and resolved by a divorce proceeding. Certainly the stakes in a divorce are weighty both for the individuals directly involved in the adjudication and for others immediately affected by it. The critical importance of the divorce process, however, weakens the argument for a long residency requirement rather than strengthens it. The impact of the divorce decree only underscores the necessity that the State’s regulation be evenhanded.3

It is not enough to recite the State’s traditionally exclusive responsibility for regulating family law matters; some tangible interference with the State’s regulatory scheme must be shown. Yet in this ease, I fail to see how any legitimate objective of Iowa’s divorce regulations would be frustrated by granting equal access to new state residents.4 To draw on an analogy, the States have great interests in the local voting process and wide latitude in regulating that process. Yet one regulation that the States may not impose is an unduly long residency requirement. Dunn v. Blumstein, 405 U. S. 330 (1972). To remark, as the Court does, that because of the consequences riding on a divorce decree “Iowa may insist that one seeking to initiate such a proceeding have the modicum of attachment to the State required here” *424is not to make an argument, but merely to state the result.

C

The Court’s third justification seems to me the only one that warrants close consideration. Iowa has a legitimate interest in protecting itself against invasion by those seeking quick divorces in a forum with relatively lax divorce laws, and it may have some interest in avoiding collateral attacks on its decree in other States.5 These interests, however, would adequately be protected by a simple requirement of domicile — physical presence plus intent to remain — which would remove the rigid one-year barrier while permitting the State to restrict the availability of its divorce process to citizens who are genuinely its own.6

*425The majority notes that in Williams v. North Carolina, 325 U. S. 226 (1945), the Court held that for ex parte divorces one State’s finding of domicile could, under limited circumstances, be challenged in the courts of another. From this, the majority concludes that since Iowa’s findings of domicile' might be subject to collateral attack elsewhere, it should be permitted to cushion its findings with a one-year residency requirement.

For several reasons, the year’s waiting period seems to me neither necessary nor much of a cushion. First, the Williams opinion was not aimed at States seeking to avoid becoming divorce mills. Quite the opposite, it was rather plainly directed at States that had cultivated a “quickie divorce” reputation by playing fast and loose with findings of domicile. See id., at 236-237; id., at 241 (Murphy, J., concurring). If Iowa wishes to avoid becoming a haven for divorce seekers, it is inconceivable that its good-faith determinations of domicile would not meet the rather lenient full faith and credit standards set out in Williams.

A second problem with the majority’s argument on this score is that Williams applies only to ex parte divorces. This Court has held that if both spouses were before the divorcing court, a foreign Staté cannot recognize a collateral challenge that would not be permissible in the divorcing State. Sherrer v. Sherrer, 334 U. S. 343 (1948); Coe v. Coe, 334 U. S. 378 (1948); Johnson v. Muelberger, 340 U. S. 581 (1951); Cook v. Cook, 342 U. S. 126 (1951). Therefore, the Iowa statute sweeps too broadly even as a defense to possible collateral attacks, since it imposes a one-year requirement whenever the respondent does not reside in the State, regardless of whether the proceeding is ex parte 7

*426Third, even a one-year period does not provide complete protection against collateral attack. It merely makes it somewhat less likely that a second State will be able to find “cogent evidence” that Iowa’s determination of domicile was incorrect. But if the Iowa court has erroneously determined the question of domicile, the year’s residence will do nothing to preclude collateral attack under Williams.

Finally, in one sense the year’s residency requirement may technically increase rather than reduce the exposure of Iowa’s decrees to collateral attack. Iowa appears to be among the States that have interpreted their divorce residency requirements as being of jurisdictional import.8 Since a State’s divorce decree is subject to collateral challenge in a foreign forum for any jurisdictional flaw that would void it in the State’s own courts, New York ex rel. Halvey v. Halvey, 330 U. S. 610 (1947), the residency requirement exposes Iowa divorce proceedings to attack both for failure to prove domicile and for failure to prove one year’s residence. If nothing else, this casts doubt on the majority’s speculation that Iowa’s residency requirement may have been intended as a statutory shield for its divorce decrees. In sum, concerns about the need *427for a long residency requirement to defray collateral attacks on state judgments seem more fanciful than real. If, as the majority assumes, Iowa is interested in assuring itself that its divorce petitioners are legitimately Iowa citizens, requiring petitioners to provide convincing evidence of bona fide domicile should be more than adequate to the task.9

Ill

I conclude that the course Iowa has chosen in , restricting access to its divorce courts unduly interferes with the right to “migrate, resettle, find a new job, and start a new life.” Shapiro v. Thompson, 394 U. S., at 629. I would reverse the judgment of the District Court and remand for entry of an order granting relief if the court finds that there is a continuing controversy in this case. See Steffel v. Thompson, 415 U. S. 452 (1974); Johnson v. New York State Education Dept., 409 U. S. 75, 79 n. 7 (1972) (Marshall, J., concurring).