CERTIORARI TO THE UNITED STATES COURT OF APPEALS FOR THE THIRD CIRCUIT.

[738] Kalvin M. Grove argued the cause for petitioner. With him on the briefs were Lawrence M. Cohen, Jeffrey S. Goldman, and Robert A. Penney.

Howard A. Specter argued the cause and filed a brief for respondents.^[1]

[739] MR. JUSTICE REHNQUIST delivered the opinion of the Court.

Respondents filed a complaint in the United States District Court for the Western District of Pennsylvania in which they asserted that petitioner's employee insurance benefits and maternity leave regulations discriminated against women in violation of Title VII of the Civil Rights Act of 1964, 78 Stat. 253, as amended by the Equal Employment Opportunity Act of 1972, 42 U. S. C. § 2000e et seq. (1970 ed. and Supp. IV). The District Court ruled in favor of respondents on the issue of petitioner's liability under that Act, and petitioner appealed to the Court of Appeals for the Third Circuit. That court held that it had jurisdiction of petitioner's appeal under 28 U. S. C. § 1291, and proceeded to affirm on the merits the judgment of the District Court. We [740] granted certiorari, 421 U. S. 987 (1975), and heard argument on the merits. Though neither party has questioned the jurisdiction of the Court of Appeals to entertain the appeal, we are obligated to do so on our own motion if a question thereto exists. Mansfield, Coldwater & Lake Michigan R. Co. v. Swan, 111 U. S. 379 (1884). Because we conclude that the District Court's order was not appealable to the Court of Appeals, we vacate the judgment of the Court of Appeals with instructions to dismiss petitioner's appeal from the order of the District Court.

Respondents' complaint, after alleging jurisdiction and facts deemed pertinent to their claim, prayed for a judgment against petitioner embodying the following relief:

"(a) requiring that defendant establish non-discriminatory hiring, payment, opportunity, and promotional plans and programs;

"(b) enjoining the continuance by defendant of the illegal acts and practices alleged herein;

"(c) requiring that defendant pay over to plaintiffs and to the members of the class the damages sustained by plaintiffs and the members of the class by reason of defendant's illegal acts and practices, including adjusted backpay, with interest, and an additional equal amount as liquidated damages, and exemplary damages;

"(d) requiring that defendant pay to plaintiffs and to the members of the class the costs of this suit and a reasonable attorneys' fee, with interest; and

"(e) such other and further relief as the Court deems appropriate." App. 19.

After extensive discovery, respondents moved for partial summary judgment only as to the issue of liability. Fed. Rule Civ. Proc. 56 (c). The District Court on January 9, 1974, finding no issues of material fact in dispute, [741] entered an order to the effect that petitioner's pregnancy-related policies violated Title VII of the Civil Rights Act of 1964. It also ruled that Liberty Mutual's hiring and promotion policies violated Title VII.^[2] Petitioner thereafter filed a motion for reconsideration which was denied by the District Court. Its order of February 20, 1974, denying the motion for reconsideration, contains the following concluding language:

"In its Order the court stated it would enjoin the continuance of practices which the court found to be in violation of Title VII. The Plaintiffs were invited to submit the form of the injunction order and the Defendant has filed Notice of Appeal and asked for stay of any injunctive order. Under these circumstances the court will withhold the issuance of the injunctive order and amend the Order previously issued under the provisions of Fed. R. Civ. P. 54 (b), as follows:

"And now this 20th day of February, 1974, it is directed that final judgment be entered in favor of Plaintiffs that Defendant's policy of requiring female employees to return to work within three months of delivery of a child or be terminated is in violation of the provisions of Title VII of the Civil Rights Act of 1964; that Defendant's policy of denying disability income protection plan benefits to female employees for disabilities related to pregnancies or childbirth are [sic] in violation of Title VII of the Civil Rights Act of 1964 and that it is expressly directed that Judgment be entered for the [742] Plaintiffs upon these claims of Plaintiffs' Complaint; there being no just reason for delay." 372 F. Supp. 1146, 1164.

It is obvious from the District Court's order that respondents, although having received a favorable ruling on the issue of petitioner's liability to them, received none of the relief which they expressly prayed for in the portion of their complaint set forth above. They requested an injunction, but did not get one; they requested damages, but were not awarded any; they requested attorneys' fees, but received none.

Counsel for respondents when questioned during oral argument in this Court suggested that at least the District Court's order of February 20 amounted to a declaratory judgment on the issue of liability pursuant to the provisions of 28 U. S. C. § 2201. Had respondents sought only a declaratory judgment, and no other form of relief, we would of course have a different case. But even if we accept respondents' contention that the District Court's order was a declaratory judgment on the issue of liability, it nonetheless left unresolved respondents' requests for an injunction, for compensatory and exemplary damages, and for attorneys' fees. It finally disposed of none of respondents' prayers for relief.

The District Court and the Court of Appeals apparently took the view that because the District Court made the recital required by Fed. Rule Civ. Proc. 54 (b) that final judgment be entered on the issue of liability, and that there was no just reason for delay, the orders thereby became appealable as a final decision pursuant to 28 U. S. C. § 1291. We cannot agree with this application of the Rule and statute in question.

Rule 54 (b)^[3] "does not apply to a single claim [743] action . . . . It is limited expressly to multiple claims actions in which `one or more but less than all' of the multiple claims have been finally decided and are found otherwise to be ready for appeal." Sears, Roebuck & Co. v. Mackey, 351 U. S. 427, 435 (1956).^[4] Here, however, respondents set forth but a single claim: that petitioner's employee insurance benefits and maternity leave regulations discriminated against its women employees in violation of Title VII of the Civil Rights Act of 1964. They prayed for several different types of relief in the event that they sustained the allegations of their complaint, see Fed. Rule Civ. Proc. 8 (a) (3), but their complaint advanced a single legal theory which was applied to only one set of facts.^[5] Thus, despite the fact that the District Court undoubtedly made the findings required [744] under the Rule, had it been applicable, those findings do not in a case such as this make the order appealable pursuant to 28 U. S. C. § 1291. See Mackey, supra, at 437-438.

We turn to consider whether the District Court's order might have been appealed by petitioner to the Court of Appeals under any other theory. The order, viewed apart from its discussion of Rule 54 (b), constitutes a grant of partial summary judgment limited to the issue of petitioner's liability. Such judgments are by their terms interlocutory, see Fed. Rule Civ. Proc. 56 (c), and where assessment of damages or awarding of other relief remains to be resolved have never been considered to be "final" within the meaning of 28 U. S. C. § 1291. See, e. g., Borges v. Art Steel Co., 243 F. 2d 350 (CA2 1957); Leonidakis v. International Telecoin Corp., 208 F. 2d 934 (CA2 1953); Tye v. Hertz Drivurself Stations, 173 F. 2d 317 (CA3 1949); Russell v. Barnes Foundation, 136 F. 2d 654 (CA3 1943). Thus the only possible authorization for an appeal from the District Court's order would be pursuant to the provisions of 28 U. S. C. § 1292.

If the District Court had granted injunctive relief but had not ruled on respondents' other requests for relief, this interlocutory order would have been appealable under § 1292 (a) (1).^[6] But, as noted above, the court did not issue an injunction. It might be argued that the order of the District Court, insofar as it failed to include the injunctive relief requested by respondents, is an interlocutory [745] order refusing an injunction within the meaning of § 1292 (a) (1). But even if this would have allowed respondents to then obtain review in the Court of Appeals, there was no denial of any injunction sought by petitioner and it could not avail itself of that grant of jurisdiction.

Nor was this order appealable pursuant to 28 U. S. C. § 1292 (b).^[7] Although the District Court's findings made with a view to satisfying Rule 54 (b) might be viewed as substantial compliance with the certification requirement of that section, there is no showing in this record that petitioner made application to the Court of Appeals within the 10 days therein specified. And that court's holding that its jurisdiction was pursuant to § 1291 makes it clear that it thought itself obliged to consider on the merits petitioner's appeal. There can be no assurance that had the other requirements of § 1292 (b) been complied with, the Court of Appeals would have exercised its discretion to entertain the interlocutory appeal.

Were we to sustain the procedure followed here, we would condone a practice whereby a district court in virtually any case before it might render an interlocutory decision on the question of liability of the defendant, [746] and the defendant would thereupon be permitted to appeal to the court of appeals without satisfying any of the requirements that Congress carefully set forth. We believe that Congress, in enacting present §§ 1291 and 1292 of Title 28, has been well aware of the dangers of an overly rigid insistence upon a "final decision" for appeal in every case, and has in those sections made ample provision for appeal of orders which are not "final" so as to alleviate any possible hardship. We would twist the fabric of the statute more than it will bear if we were to agree that the District Court's order of February 20, 1974, was appealable to the Court of Appeals.

The judgment of the Court of Appeals is therefore vacated, and the case is remanded with instructions to dismiss the petitioner's appeal.

It is so ordered.

MR. JUSTICE BLACKMUN took no part in the consideration or decision of this case.

[1] Briefs of amici curiae urging reversal were filed by Gordon Dean Booth, Jr., for Alaska Airlines, Inc., et al.; by Edward Silver, Larry M. Lavinsky, Sara S. Portnoy, and Kenneth L. Kimble for the American Life Insurance Assn. et al.; by William Martin and Paul C. Blume for the American Mutual Insurance Alliance et al.; by Thompson Powers for the American Telephone & Telegraph Co.; by Simon H. Rifkind, Frazer F. Hilder, and Edmond J. Dilworth, Jr., for General Motors Corp.; by Richard D. Godown for the National Association of Manufacturers; by Lloyd Sutter for Owens-Illinois, Inc., et al.; and by John G. Wayman and Scott F. Zimmermanfor Westinghouse Electric Corp.

Briefs of amici curiae urging affirmance were filed by Solicitor General Bork, Assistant Attorney General Pottinger, Brian K. Landsberg, Walter W. Barnett, Abner W. Sibal, Joseph T. Eddins, and Beatrice Rosenberg for the United States et al.; by Francis X. Bellotti, Attorney General, and Barbara J. Rouse and Terry Jean Seligmann, Assistant Attorneys General, for the Commonwealth of Massachusetts et al.; by William J. Brown, Attorney General, and Andrew J. Ruzicho and Earl M. Manz, Assistant Attorneys General, for the State of Ohio; by Henry Spitz and Paul Hartman for the New York State Division of Human Rights; by Ruth Bader Ginsburg, Melvin L. Wulf, and David Rubin for the American Civil Liberties Union et al.; by J. Albert Woll, Laurence Gold, Stephen I. Schlossberg, and John Fillion for the American Federation of Labor and Congress of Industrial Organizations et al.; by Diane Serafin Blank and Nancy E. Stanley for Blank Goodman Kelly Rone & Stanley; by Wendy W. Williams, Rhonda Copelon, Sylvia Roberts, Marilyn Hall Patel, Judith Lonnquist, Gladys Kessler, and Peter Hart Weiner for the Center for Constitutional Rights et al.; and by Mary K. O'Melveny, Jonathan W. Lubell, H. Howard Ostrin, and Charles V. Koons for the Communication Workers of America, AFL-CIO.

[2] The portion of the District Court's order concerning petitioner's hiring and promotion policies was separately appealed to a different panel of the Court of Appeals. The judgment rendered by the Third Circuit upon that appeal is not before us in this case. See Wetzel v. Liberty Mutual Ins. Co., 508 F. 2d 239, cert. denied, 421 U. S. 1011 (1975).

[3] "Judgment upon multiple claims or involving multiple parties. "When more than one claim for relief is presented in an action, whether as a claim, counterclaim, cross-claim, or third-party claim, or when multiple parties are involved, the court may direct the entry of a final judgment as to one or more but fewer than all of the claims or parties only upon an express determination that there is no just reason for delay and upon an express direction for the entry of judgment. In the absence of such determination and direction, any order or other form of decision, however designated, which adjudicates fewer than all the claims or the rights and liabilities of fewer than all the parties shall not terminate the action as to any of the claims or parties, and the order or other form of decision is subject to revision at any time before the entry of judgment adjudicating all the claims and the rights and liabilities of all the parties."

[4] Following Mackey, the Rule was amended to insure that orders finally disposing of some but not all of the parties could be appealed pursuant to its provisions. That provision is not implicated in this case, however, to which Mackey's exposition of the Rule remains fully accurate.

[5] We need not here attempt any definitive resolution of the meaning of what constitutes a claim for relief within the meaning of the Rules. See 6 J. Moore, Federal Practice ¶¶ 54.24, 54.33 (2d ed. 1975). It is sufficient to recognize that a complaint asserting only one legal right, even if seeking multiple remedies for the alleged violation of that right, states a single claim for relief.

[6] 

"The courts of appeals shall have jurisdiction of appeals from:

"(1) Interlocutory orders of the district courts of the United States, the United States District Court for the District of the Canal Zone, the District Court of Guam, and the District Court of the Virgin Islands, or of the judges thereof, granting, continuing, modifying, refusing or dissolving injunctions, or refusing to dissolve or modify injunctions, except where a direct review may be had in the Supreme Court."

[7] "When a district judge, in making in a civil action an order not otherwise appealable under this section, shall be of the opinion that such order involves a controlling question of law as to which there is substantial ground for difference of opinion and that an immediate appeal from the order may materially advance the ultimate termination of the litigation, he shall so state in writing in such order. The Court of Appeals may thereupon, in its discretion, permit an appeal to be taken from such order, if application is made to it within ten days after the entry of the order: Provided, however, That application for an appeal hereunder shall not stay proceedings in the district court unless the district judge or the Court of Appeals or a judge thereof shall so order."

SEARS, ROEBUCK & CO. v. MACKEY et al.

No. 34.

Argued February 28, 1956.

Decided June 11, 1956.

*428Walter J. Rockier argued the cause and filed a brief for petitioner.

Edward I. Rothschild argued the cause for respondents. With him on a brief for Mackey, respondent, was John Paul Stevens.

Mr. Justice Burton

delivered the opinion of the Court.

This action, presenting multiple claims for relief, was brought by Mackey and another in the United States District Court for the Northern District of Illinois, Eastern Division, in 1953. The court expressly directed that judgment be entered for the defendant, Sears, Roebuck & Co., on two, but less than all, of the claims presented. It also expressly determined that there was no just reason for delay in making the entry. After Mackey’s notice of appeal from that judgment to the Court of Appeals for the Seventh Circuit, Sears, Roebuck & Co. moved to dismiss the appeal for lack of appellate jurisdiction. The Court of Appeals upheld its jurisdiction and denied the *429motion, relying upon 28 U. S. C. § 1291 and Rule 54 (b) of the Federal Rules of Civil Procedure, as amended in 1946. 218 F. 2d 295. Because of the importance of the issue in determining appellate jurisdiction and because of a conflict of judicial views on the subject,1 we granted certiorari. 348 U. S. 970. For the reasons hereafter stated, we sustain the Court of Appeals and its appellate jurisdiction.

Although we are here concerned with the present ap-pealability of the judgment of the District Court and not with its merits, we must examine the claims stated in the complaint so as to consider adequately the issue of appealability.

The complaint contains six counts. We disregard the fifth because it has been abandoned and the sixth because it duplicates others. The claims stated in Counts I and II are material and have been dismissed without leave to amend. The claim contained in Count III and that in amended Count IV are at issue on the answers filed by Sears, Roebuck & Co. The appeal before us is from a *430judgment striking out Counts I and II without disturbing Counts III and IV, and the question presented is whether such a judgment is presently appealable when the District Court, pursuant to amended Rule 54 (b), has made “an express determination that there is no just reason for delay” and has given “an express direction for the entry of judgment.”

In Count I, Mackey, a citizen of Illinois, and Time Saver Tools, Inc., an Illinois corporation owned by Mackey, are the original plaintiffs and the respondents here. Sears, Roebuck & Co., a New York corporation doing business in Illinois, is the original defendant and the petitioner here. Mackey charges Sears with conduct violating the Sherman Antitrust Act in a manner prejudicial to three of Mackey’s commercial ventures causing him $190,000 damages, for which he seeks $570,000 as treble damages. His first charge is unlawful destruction by Sears, since 1949, of the market for nursery lamps manufactured by General Metalcraft Company, a corporation wholly owned by Mackey. Mackey claims that this caused him a loss of $150,000. His second charge is unlawful interference by Sears, in 1952, with Mackey’s contract to sell, on commission, certain tools and other products of the Vascoloy-Ramet Corporation, causing Mackey to lose $15,000. His third charge is unlawful destruction by Sears, in 1952, of the market for a new type of carbide-tipped lathe bit and for other articles manufactured by Time Saver Tools, Inc., resulting in a loss to Mackey of $25,000. Mackey combines such charges with allegations that Sears has used its great size to monopolize commerce and restrain competition in these fields. He asks for damages and equitable relief.

In Count II, Mackey claims federal jurisdiction by virtue of diversity of citizenship. He incorporates the allegations of Count I as to the Metalcraft transactions and asks for $250,000 damages for Sears’ wilful destruc*431tion of the business of Metalcraft, plus $50,000 for Mackey’s loss on obligations guaranteed by him.

In Count III, Mackey seeks $75,000 in a common-law proceeding against Sears for unlawfully inducing a breach of his Yascoloy commission contract.

In Count IV, Time Saver seeks $200,000 in a common-law proceeding against Sears for unlawfully destroying Time Saver’s business by unfair competition and patent infringement.

The jurisdiction of the Court of Appeals to entertain Mackey’s appeal from the District Court’s judgment depends upon 28 U. S. C. § 1291, which provides that “The courts of appeals shall have jurisdiction of appeals from all final decisions of the district courts of the United States . . . .” (Emphasis supplied.)

If Mackey’s complaint had contained only Count I, there is no doubt that a judgment striking out that count and thus dismissing, in its entirety, the claim there stated would be both a final and an appealable decision within the meaning of § 1291. Similarly, if his complaint had contained Counts I, II, III and IV, there is no doubt that a judgment striking out all four would be a final and appealable decision under § 1291. The controversy before us arises solely because, in this multiple claims action, the District Court has dismissed the claims stated in Counts I and II, but has left unadjudicated those stated in Counts III and IV.2

Before the adoption of the Federal Rules of Civil Procedure in 1939, such a situation was generally regarded as leaving the appellate court without jurisdiction- of an attempted appeal. It was thought that, although the judgment was a final decision on the respective claims in Counts I and II, it obviously was not a final decision of *432the whole case, and there was no authority for treating anything less than the whole case as a judicial unit for purposes of appeal.3 This construction of the judicial unit was developed from the common law which had dealt with litigation generally less complicated than much of that of today.4

With the Federal Rules of Civil Procedure, there came an increased opportunity for the liberal joinder of claims in multiple claims actions. This, in turn, demonstrated a need for relaxing the restrictions upon what should be treated as a judicial unit for purposes of appellate jurisdiction. Sound judicial administration did not require relaxation of the standard of finality in the disposition of the individual adjudicated claims for the purpose of their appealability. It did, however, demonstrate that, at least in multiple claims actions, some final decisions, on less than all of the claims, should be appealable without waiting for a final decision on all of the claims. Largely to *433meet this need, in 1939, Rule 54 (b) was promulgated in its original form through joint action of Congress and this Court.5 It read as follows:

“(b) Judgment at Various Stages. When more than one claim for relief is presented in an action, the court at any stage, upon a determination of the issues material to a particular claim and all counterclaims arising out of the transaction or occurrence which is the subject matter of the claim, may enter a judgment disposing of such claim. The judgment shall terminate the action with respect to the claim so disposed of and the action shall proceed as to the remaining claims. In case a separate judgment is so entered, the court by order may stay its enforcement until the entering of a subsequent judgment or judgments and may prescribe such conditions as are necessary to secure the benefit thereof to the party in whose favor the judgment is entered.”

It gave limited relief. The courts interpreted it as not relaxing the requirement of a “final decision” on each individual claim as the basis for an appeal, but as author*434izing a limited relaxation of the former general practice that, in multiple claims actions, all the claims had to be finally decided before an appeal could be entertained from a final decision upon any of them.6 Thus, original Rule 54 (b) modified the single judicial unit theory but left unimpaired the statutory concept of finality prescribed by § 1291. However, it was soon found to be inherently difficult to determine by any automatic standard of unity which of several multiple claims were sufficiently separable from others to qualify for this relaxation of the unitary principle in favor of their appealability. The result was that the jurisdictional time for taking an appeal from a final decision on less than all of the claims in a multiple claims action in some instances expired earlier than was foreseen by the losing party. It thus became prudent to take immediate appeals in all cases of doubtful appealability and the volume of appellate proceedings was undesirably increased.

Largely to overcome this difficulty, Rule 54 (b) was amended, in 1946, to take effect in 1948.7 Since then it has read as follows:

“(b) Judgment Upon Multiple Claims. When more than one claim for relief is presented in an action, whether as a claim, counterclaim, cross-claim, or third-party claim, the court may direct the entry of a final judgment upon one or more but less than all of the claims only upon an express determination *435that there is no just reason for delay and upon an express direction for the entry of judgment. In the absence of such determination and direction, any order or other form of decision, however designated, which adjudicates less than all the claims shall not terminate the action as to any of the claims, and the order or other form of decision is subject to revision at any time before the entry of judgment adjudicating all the claims.” (Emphasis supplied.)

In this form, it does not relax the finality required of each decision, as an individual claim, to render it appeal-able, but it does provide a practical means of permitting an appeal to be taken from one or more final decisions on individual claims, in multiple claims actions, without waiting for final decisions to be rendered on all the claims in the case. The amended rule does not apply to a single claim action nor to a multiple claims action in which all of the claims have been finally decided. It is limited expressly to multiple claims actions in which “one or more but less than all” of the multiple claims have been finally decided and are found otherwise to be ready for appeal.

To meet the demonstrated need for flexibility, the District Court is used as a “dispatcher.” It is permitted to determine, in the first instance, the appropriate time when each “final decision” upon “one or more but less than all” of the claims in a multiple claims action is ready for appeal. This arrangement already has lent welcome certainty to the appellate procedure. Its “negative effect” has met with uniform approval. The effect so referred to is the rule’s specific requirement that for “one or more but less than all” multiple claims to become appealable, the District Court must make both “an express determination that there is no just reason for delay” and “an express direction for the entry of judgment.” A party adversely affected by a final decision thus knows that *436his time for appeal will not run against him until this certification has been made.8

In the instant case, the District Court made this certification, but Sears, Roebuck & Co. nevertheless moved to dismiss the appeal for lack of appellate jurisdiction under § 1291. The grounds for such a motion ordinarily might be (1) that the judgment of the District Court was not a decision upon a “claim for relief,” (2) that the decision was not a “final decision” in the sense of an ultimate disposition of an individual claim entered in the course of a multiple claims action, or (3) that the District Court abused its discretion in certifying the order.

In the case before us, there is no doubt that each of the claims dismissed is a “claim for relief” within the meaning of Rule 54 (b), or that their dismissal constitutes a “final decision” on individual claims. Also, it cannot well be argued that the claims stated in Counts I and II are so inherently inseparable from, or closely related to, those stated in Counts III and IV that the District Court has abused its discretion in certifying that there exists no just reason for delay. They certainly can be decided independently of each other.

Petitioner contends that amended Rule 54 (b) attempts to make an unauthorized extension of § 1291. We disagree. It could readily be argued here that the claims stated in Counts I and II are sufficiently independent of those stated in Counts III and IV to satisfy the requirements of Rule 54 (b) even in its original form. If that were so, the decision dismissing them would also be appealable under the amended rule. It is nowhere contended today that a decision that would have been appealable under the original rule is not also appealable under the amended rule, provided the District Court makes the required certification.

*437While it thus might be possible to hold that in this case the Court of Appeals had jurisdiction under original Rule 54 (b), there at least would be room for argument on the issue of whether the decided claims were separate and independent from those still pending in the District Court.9 Thus the instant case affords an excellent illustration of the value of the amended rule which was designed to overcome that difficulty. Assuming that the requirements of the original rule are not met in this case, we nevertheless are enabled to recognize the present appellate jurisdiction of the Court of Appeals under the amended rule. The District Court cannot, in the exercise of its discretion, treat as “final” that which is not “final” within the meaning of § 1291. But the District Court may, by the exercise of its discretion in the interest of sound judicial administration, release for appeal final decisions upon one or more, but less than all, claims in multiple claims actions. The timing of such a release is, with good reason, vested by the rule primarily in the discretion of the District Court as the one most likely to be familiar with the case and with any justifiable reasons for delay. With equally good reason, any abuse of that discretion remains reviewable by the Court of Appeals.

Rule 54 (b), in its original form, thus may be said to have modified the single judicial unit practice which had been developed by court decisions. The validity of that rule is no longer questioned. In fact, it was applied by this Court in Reeves v. Beardall, 316 U. S. 283, without its validity being questioned.

*438Rule 54 (b), in its amended form, is a comparable exercise of the rulemaking authority of this Court. It does not supersede any statute controlling appellate jurisdiction. It scrupulously recognizes the statutory requirement of a “final decision” under § 1291 as a basic requirement for an appeal to the Court of Appeals. It merely administers that requirement in a practical manner in multiple claims actions and does so by rule instead of by judicial decision. By its negative effect, it operates to restrict in a valid manner the number of appeals in multiple claims actions.

We reach a like conclusion as to the validity of the amended rule where the District Court acts affirmatively and thus assists in properly timing the release of final decisions in multiple claims actions. The amended rule adapts the single judicial unit theory so that it better meets the current needs of judicial administration. Just as Rule 54 (b), in its original form, resulted in the release of some decisions on claims in multiple claims actions before they otherwise would have been released,10 so amended Rule 54 (b) now makes possible the release of more of such decisions subject to judicial supervision. The amended rule preserves the historic federal policy against piecemeal appeals in many cases more effectively than did the original rule.11

Accordingly, the appellate jurisdiction of the Court of Appeals is sustained,12 and its judgment denying the motion to dismiss the appeal for lack of appellate jurisdiction is

Affirmed.

1

For decisions directly or impliedly sustaining the validity of amended Rule 54 (b), as applied in the instant case, see Rieser v. Baltimore & O. R. Co., 224 F. 2d 198 (C. A. 2d Cir.), and cases cited at 203, n. 7; United Artists Corp. v. Masterpiece Productions, Inc., 221 F. 2d 213 (C. A. 2d Cir.); Clarksville v. United States, 198 F. 2d 238 (C. A. 4th Cir.); Boston Medical Supply Co. v. Lea & Febiger, 195 F. 2d 853 (C. A. 1st Cir.); Bendix Aviation Corp. v. Glass, 195 F. 2d 267 (C. A. 3d Cir.); Lopinsky v. Hertz Drive-Ur-Self Systems, Inc., 194 F. 2d 422 (C. A. 2d Cir.), concr. op. of Judge Clark, at 424-430; Pabellon v. Grace Line, Inc., 191 F. 2d 169 (C. A. 2d Cir.). See 6 Moore’s Federal Practice (2d ed. 1953) 220-230, and Note, 62 Yale L. J. 263.

Contra: See Rieser v. Baltimore & O. R. Co., supra, concr. op. of Judge Frank, at 205-208; Bendix Aviation Corp. v. Glass, supra, concr. op. of Judge Hastie, at 277-282; Pabellon v. Grace Line, Inc., supra, concr. op. of Judge Frank, at 176-181; Flegenheimer v. General Mills, Inc., 191 F. 2d 237 (C. A. 2d Cir.). See also, Gold Seal Co. v. Weeks, 93 U. S. App. D. C. 249, 209 F. 2d 802.

2

Sears also contends that the Court of Appeals misconstrued amended Rule 54 (b). See Flegenheimer v. General Mills, Inc., supra. The meaning of that rule is considered hereafter.

3

At common law, a writ of error did not lie to review a judgment that failed to adjudicate every cause of action asserted in the controversy. See Holcombe v. McKusick, 20 How. 552; United States v. Girault, 11 How. 22; Metcalfe’s Case, 11 Co. Rep. 38a, 77 Eng. Rep. 1193. The rule generally followed in the federal courts was that, in a case involving a single plaintiff and a single defendant, a judgment was not appealable if it disposed of some, but less than all, of the claims presented. See Collins v. Miller, 252 U. S. 364; Sheppy v. Stevens, 200 F. 946 (C. A. 2d Cir.). In eases involving multiple parties where the alleged liability was joint, a judgment was not appealable unless it terminated the action as to all the defendants. See Hohorst v. Hamburg-American Packet Co., 148 U. S. 262. But if, in a multiple party case, a judgment finally disposed of a claim that was recognized to be separate and distinct from the others, that judgment, under some circumstances, was appealable. See Republic of China v. American Express Co., 190 F. 2d 334 (C. A. 2d Cir.).

4

The appellate jurisdiction of the United States Courts of Appeals, with exceptions not directly pertinent here, is still largely restricted to the review of cases appealed under 28 U. S. C. § 1291. But see Forgay v. Conrad, 6 How. 201; Cohen v. Beneficial Industrial Loan Corp., 337 U. S. 541; 28 U. S. C. §§ 1292, 1651.

5

The Supreme Court’s authority to promulgate the Federal Rules of Civil Procedure is derived from the Enabling Act, now appearing as 28 U. S. C. § 2072. It authorizes this Court to promulgate rules governing “the forms of process, writs, pleadings, and motions, and the practice and procedure of the district courts of the United States ... in civil actions.” It provides that such rules “shall not abridge, enlarge or modify any substantive right . . . .” And, by reason of Article I, § 8, of the Constitution, it has been held repeatedly that only Congress may define the jurisdiction of the lower federal courts. See Sibbach v. Wilson & Co., 312 U. S. 1; Baltimore Contractors, Inc. v. Bodinger, 348 U. S. 176; and Fed. Rules Civ. Proc., 82. “Such rules shall not take effect until they have been reported to Congress by the Chief Justice at or after the beginning of a regular session thereof but not later than the first day of May, and until the expiration of ninety days after they have been thus reported.” 28 U. S. C. § 2072.

6

See Pabellon v. Grace Line, Inc., supra, at 174.

7

“. . . situations arose where district courts made a piecemeal disposition of an action and entered what the parties thought amounted to a judgment, although a trial remained to be had on other claims similar or identical with those disposed of. In the interim the parties did not know their ultimate rights, and accordingly took an appeal, thus putting the finality of the partial judgment in question.” Report of Advisory Committee on Proposed Amendments to Rules of Civil Procedure 70-71 (June 1946).

8

For favorable comment on this aspect of the rule, see Dickinson v. Petroleum Corp., 338 U. S. 507, 511-512.

9

In the instant case, the claim dismissed by striking out Count I is based on the Sherman Act, while Counts III and IV do not rely on, or even refer to, that Act. They are largely predicated on common-law rights. The basis of liability in Count I is independent of that on which the claims in Counts III and IV depend. But the claim in Count I does rest in part on some of the facts that are involved in Counts III and IV. The claim stated in Count II is clearly independent of those in Counts III and IV.

10

See Collins v. Metro-Goldwyn Pictures Corp., 106 F. 2d 83 (C. A. 2d Cir.), cited in Reeves v. Beardall, 316 U. S. 283.

11

See Cobbledick v. United States, 309 U. S. 323.

12

Mackey also argues that the Court of Appeals has jurisdiction under 28 U. S. C. § 1292 (1). In view of our disposition of this case, we do not reach that contention.

*439Mr. Justice Frankfurter,

whom Mr. Justice Harlan joins,

concurring in No. 34 and dissenting in No. 76.

The result in these two litigations of course has significance for the parties. That is, however, of relative insignificance compared to the directions which judges in the district courts and courts of appeals will draw from the Court’s opinions. For me, what is said has not a little kinship with the pronouncements of the Delphic oracle.

The opinion in Cold Metal Process Co. v. United Engineering & Foundry Co., post, p. 445, declares that 28 U. S. C. § 1291 remains unimpaired, but surely that section does not remain what it was before these opinions were written. Rule 54 (b) is apparently the transforming cause. The Court could have said that Rule 54 (b), promulgated under congressional authority and having the force of statute, has qualified 28 U. S. C. § 1291. It does not say so. The Court could have said that it rejects the reasoning of the decisions in which this Court for over a century has interpreted § 1291 as expressing a hostility toward piecemeal appeals. It does not say so. The Court could have said that Rule 54 (b)’s requirement of a certificate from a district judge means that the district judges alone determine the content of finality. The Court does not say that either.

The Court does indicate that what has been the core of the doctrine of finality as applied to multiple claims litigation — that only that part of a litigation which is separate from, and independent of, the remainder of the litigation can be appealed before the completion of the entire litigation — is no longer to be applied as a standard, or at least as an exclusive standard, for deciding what is final for purposes of § 1291. The Court does not, however, indicate what standards the district courts and the courts of appeals are now to apply in determining when a decision is final. It leaves this problem in the first *440instance to the district courts, subject to review by the courts of appeals for an abuse of discretion. In other instances where a district court’s ruling can be upset only for an abuse of its discretion, the scope of review is necessarily narrow. Here, in regard to the present problem, what is to come under review is a newly modified requirement of finality. But the requirement continues to be based upon a statute, viz., 28 U. S. C. § 1291, and that statute defines and constricts the jurisdiction of the courts of appeals. Therefore the issue of compliance with this congressional command would, I should suppose, cast upon the courts of appeals a duty of independent judgment broader than is implied by the usual flavor of the phrase “abuse of discretion.”

For me, the propositions emerging from analysis of the relationship of Rule 54 (b) to 28 U. S. C. § 1291 are clear.

1. 28 U. S. C. § 1291 is left intact by Rule 54 (b). It could not be otherwise with due regard for the congressional policy embodied in that section and in view of what the Advisory Committee on the Rules said in its Note to amended Rule 54 (b):

“The historic rule in the federal courts has always prohibited piecemeal disposal of litigation and permitted appeals only from final judgments except in those special instances covered by statute. . . . Rule 54 (b) was originally adopted in view of the wide scope and possible content of the newly created ‘civil action’ in order to avoid the possible injustice of a delay in judgment of a distinctly separate claim to await adjudication of the entire case. It was not designed to overturn the settled federal rule stated above ....

“. . . After extended consideration, it [the Committee] concluded that a retention of the older fed*441eral rule was desirable, and that this rule needed only the exercise of a discretionary power to afford a remedy in the infrequent harsh case to provide a simple, definite, workable rule. This is afforded by amended Rule 54 (b). It re-establishes an ancient policy with clarity and precision. . . .” Report of Advisory Committee on Proposed Amendments to Rules of Civil Procedure 70-72.

2. 28 U. S. C. § 1291 is not a technical rule in a game. It expresses not only a deeply rooted but a wisely sanctioned principle against piecemeal appeals governing litigation in the federal courts. See Cobbledick v. United States, 309 U. S. 323; Radio Station WOW v. Johnson, 326 U. S. 120, 123-127. The great importance of this characteristic feature of the federal judicial system — its importance in administering justice — is made luminously manifest by considering the evils where, as in New York, piecemeal reviews are allowed.

3. While the principle against piecemeal appeals has been compendiously and therefore, at times, loosely phrased as implying that the whole of a litigation, no matter what its nature, must be completed before any appeal is allowed, see Collins v. Miller, 252 U. S. 364, 370, the underlying rationale of the principle has been respected when not susceptible of this mechanical way of putting it. What have been called exceptions are not exceptions at all in the sense of inroads on the principle. They have not qualified the core, that is, that there should be no premature, intermediate appeal.

Thus the Court has permitted appeal before completion of the whole litigation when failure to do so would preclude any effective review or would result in irreparable injury. See Forgay v. Conrad, 6 How. 201; Cohen v. Beneficial Loan Corp., 337 U. S. 541, 545-547; Swift & Co. v. Compania Caribe, 339 U. S. 684, 688-689. A *442second situation in which the Court has found that an appeal before termination of the entire litigation did not conflict with the congressional policy against piecemeal appeals is that in which a party to the completed portion of the litigation has no interest in the rest of the proceedings and to make him await their outcome would merely cause unfairness. See Williams v. Morgan, 111 U. S. 684, 699; United States v. River Rouge Co., 269 U. S. 411, 413-414.

4. The expansion by the Federal Rules of the allowable content of a proceeding and the range of a litigation inevitably enlarged the occasions for severing one aspect or portion of a litigation from what remains under the traditional test of a “final decision.” On the basis of prior cases, we held that it was not a departure from the policy against piecemeal appeals to permit an appeal with respect to that part of a multiple claims litigation based on a set of facts separate and independent from the facts on which the remainder of the litigation was based. Reeves v. Beardall, 316 U. S. 283. The Note of the Advisory Committee, quoted supra, demonstrates that the amended Rule 54 (b) was designed in accordance with the historic policy against premature appeal and with the decisions of this Court allowing appeal from a “judgment of a distinctly separate claim.” What the Rule did introduce, however, was a discretionary power in the district judge to control appealability by preventing a party from even attempting to appeal a severable part of a litigation unless the district judge has expressly certified that there is no just reason for delay and has expressly directed entry of judgment on that phase of the litigation. This provision was directed to the kind of difficulty encountered in Dickinson v. Petroleum Conversion Corp., 338 U. S. 507, in ascertaining whether the district judge is in fact finished with a separable part of the litigation.

*443The Court casually disregards this long history of § 1291 and the bearing of Rule 54 (b) to it by rejecting the separate-and-independent test as the basis for determining the finality of a part of a multiple claims litigation. The Court says that its decision “does not impair the statutory concept of finality embraced in § 1291.” The Court may not do so in words, for it pays lip-service to § 1291. But that section’s function as a brake against piecemeal appeals in future multiple claims litigation is greatly impaired. Encouragement is abundantly given to parties to seek such appeals.

The principles which this Court has heretofore enunciated over a long course of decisions under § 1291 furnish ready guides for deciding the appealability of the certified parts of the litigation in the two cases now before the Court. Count II in Sears, Roebuck & Co. v. Mackey, ante, p. 427, is appealable since the transactions and occurrences involved in it do not involve any of those embraced in Counts III and IV. Count I involves at least two transactions which are also the subject matter of Counts III and IV, but is appealable under § 1292 (1) as an interlocutory order denying an injunction. In Cold Metal Process Co. v. United Engineering & Foundry Co., post, p. 445, the counterclaim, even if not compulsory, is based in substantial part on the transactions involved in the main litigation and hence not appealable.

5. Of course, as the Court’s opinion appears to recognize, that crucial principle of the doctrine of finality that the court of appeals has no jurisdiction unless there is a “final decision” cannot be left to the district court. It is one thing for a district court to determine whether it is or is not through with a portion of a litigation. It is quite another thing for it to determine whether the requirements of § 1291 are satisfied so as to give jurisdiction to the court of appeals. A district court can no more confer *444jurisdiction on a court of appeals outside the limits of 28 U. S. C. § 1291 than a state supreme court can confer jurisdiction on this Court beyond the bounds of 28 U. S. C. § 1257. In a particular litigation the opinion of the district judge may properly be deemed a valuable guide. But flexibility would be a strange name for authority in the district court to command the court of appeals to exercise jurisdiction.

6. In summary, then, the Court rightly states, even if it does not hold, that § 1291 is unimpaired by Rule 54 (b). Section 1291 is what a long course of decision has construed it to be. The unifying principle of decisions for over a century is observance of hostility in the federal judicial system to piecemeal appellate review (with a few strictly defined exceptions not here relevant, see 28 U. S. C. § 1292) of one litigation, no matter how many phases or parts there may be to a single judicial proceeding, so long as no part has become separated from, and independent of, the others. This rooted principle against piecemeal appeals of an organic whole — the core of § 1291 — is not left unimpaired when its enforcement is committed without guidance to the individualized notions about finality of some two hundred and fifty district judges, themselves accountable to the discordant views of eleven essentially independent courts of appeals. Allowing such leeway to the district courts and courts of appeals is not flexibility but anarchy.

CERTIORARI TO THE UNITED STATES COURT OF APPEALS FOR THE THIRD CIRCUIT.^[1]

[542] Charles Hershenstein and Philip B. Kurland argued the cause for petitioners in No. 442 and respondents in [543] No. 512. With them on the brief were Edward J. O'Mara, Samuel Dreskin and David F. Cohen.

John M. Harlan argued the cause for the Beneficial Industrial Loan Corp., respondent in No. 442 and petitioner in No. 512. With him on the brief were Charles Danzig and Walter Pond.

Briefs of amici curiae in support of petitioners in No. 442 and respondents in No. 512 were filed by Julius Levy for Weinberger; and by Lewis M. Dabney, Jr.

MR. JUSTICE JACKSON delivered the opinion of the Court.

The ultimate question here is whether a federal court, having jurisdiction of a stockholder's derivative action only because the parties are of diverse citizenship, must apply a statute of the forum state which makes the plaintiff, if unsuccessful, liable for the reasonable expenses, including attorney's fees, of the defense and entitles the corporation to require security for their payment.

Petitioners' decedent, as plaintiff, brought in the United States District Court for New Jersey an action in the right of the Beneficial Industrial Loan Corporation, a Delaware corporation doing business in New Jersey. The defendants were the corporation and certain of its managers and directors. The complaint alleged generally that since 1929 the individual defendants engaged in a continuing and successful conspiracy to enrich themselves at the expense of the corporation. Specific charges of mismanagement and fraud extended over a period of eighteen years and the assets allegedly wasted or diverted thereby were said to exceed $100,000,000. The stockholder had demanded that the corporation institute proceedings for its recovery but, by their control of the corporation, the individual defendants prevented it from doing so. This stockholder, therefore, sought to assert [544] the right of the corporation. One of 16,000 stockholders, he owned 100 of its more than two million shares, so that his holdings, together with 150 shares held by the intervenor, approximated 0. 0125% of the outstanding stock and had a market value that had never exceeded $9,000.

The action was brought in 1943, and various proceedings had been taken therein when, in 1945, New Jersey enacted the statute which is here involved.^[2] Its general effect is to make a plaintiff having so small an interest liable for the reasonable expenses and attorney's fees of [545] the defense if he fails to make good his complaint and to entitle the corporation to indemnity before the case can be prosecuted. These conditions are made applicable to pending actions. The corporate defendant therefore moved to require security, pointed to its by-laws by which it might be required to indemnify the individual defendants, and averred that a bond of $125,000 would be appropriate.

The District Court was of the opinion that the state enactment is not applicable to such an action when pending in a federal court, 7 F.R.D. 352. The Court of Appeals was of a contrary opinion and reversed, 170 F.2d 44, and we granted certiorari. 336 U.S. 917.

APPEALABILITY.

At the threshold we are met with the question whether the District Court's order refusing to apply the statute was an appealable one. Title 28 U.S.C. § 1291 provides, as did its predecessors, for appeal only "from all final decisions of the district courts," except when direct appeal to this Court is provided. Section 1292 allows appeals also from certain interlocutory orders, decrees and judgments, not material to this case except as they indicate the purpose to allow appeals from orders other than final judgments when they have a final and irreparable effect on the rights of the parties. It is obvious that, if Congress had allowed appeals only from those final judgments which terminate an action, this order would not be appealable.

[546] The effect of the statute is to disallow appeal from any decision which is tentative, informal or incomplete. Appeal gives the upper court a power of review, not one of intervention. So long as the matter remains open, unfinished or inconclusive, there may be no intrusion by appeal. But the District Court's action upon this application was concluded and closed and its decision final in that sense before the appeal was taken.

Nor does the statute permit appeals, even from fully consummated decisions, where they are but steps towards final judgment in which they will merge. The purpose is to combine in one review all stages of the proceeding that effectively may be reviewed and corrected if and when final judgment results. But this order of the District Court did not make any step toward final disposition of the merits of the case and will not be merged in final judgment. When that time comes, it will be too late effectively to review the present order, and the rights conferred by the statute, if it is applicable, will have been lost, probably irreparably. We conclude that the matters embraced in the decision appealed from are not of such an interlocutory nature as to affect, or to be affected by, decision of the merits of this case.

This decision appears to fall in that small class which finally determine claims of right separable from, and collateral to, rights asserted in the action, too important to be denied review and too independent of the cause itself to require that appellate consideration be deferred until the whole case is adjudicated. The Court has long given this provision of the statute this practical rather than a technical construction. Bank of Columbia v. Sweeny, 1 Pet. 567, 569; United States v. River Rouge Co., 269 U.S. 411, 414; Cobbledick v. United States, 309 U.S. 323, 328.

We hold this order appealable because it is a final disposition of a claimed right which is not an ingredient [547] of the cause of action and does not require consideration with it. But we do not mean that every order fixing security is subject to appeal. Here it is the right to security that presents a serious and unsettled question. If the right were admitted or clear and the order involved only an exercise of discretion as to the amount of security, a matter the statute makes subject to reconsideration from time to time, appealability would present a different question.

Since this order may be reviewed on appeal, the petition in No. 512, whereby the corporation asserts the right to compel security by mandamus, is dismissed.

CONSTITUTIONALITY.

Petitioners deny the validity of the statute under the Federal Constitution and the New Jersey Constitution. The latter question is ultimately for the state courts, and since they have made no contrary determination, we shall presume in the circumstances of this case that the statute conforms with the state constitution.

Federal constitutional questions we must consider, because a federal court would not give effect, in either a diversity or nondiversity case, to a state statute that violates the Constitution of the United States.

The background of stockholder litigation with which this statute deals requires no more than general notice. As business enterprise increasingly sought the advantages of incorporation, management became vested with almost uncontrolled discretion in handling other people's money. The vast aggregate of funds committed to corporate control came to be drawn to a considerable extent from numerous and scattered holders of small interests. The director was not subject to an effective accountability. That created strong temptation for managers to profit personally at expense of their trust. The business code became all too tolerant of such practices. Corporate laws [548] were lax and were not self-enforcing, and stockholders, in face of gravest abuses, were singularly impotent in obtaining redress of abuses of trust.

Equity came to the relief of the stockholder, who had no standing to bring civil action at law against faithless directors and managers. Equity, however, allowed him to step into the corporation's shoes and to seek in its right the restitution he could not demand in his own. It required him first to demand that the corporation vindicate its own rights, but when, as was usual, those who perpetrated the wrongs also were able to obstruct any remedy, equity would hear and adjudge the corporation's cause through its stockholder with the corporation as a defendant, albeit a rather nominal one. This remedy, born of stockholder helplessness, was long the chief regulator of corporate management and has afforded no small incentive to avoid at least grosser forms of betrayal of stockholders' interests. It is argued, and not without reason, that without it there would be little practical check on such abuses.

Unfortunately, the remedy itself provided opportunity for abuse, which was not neglected. Suits sometimes were brought not to redress real wrongs, but to realize upon their nuisance value. They were bought off by secret settlements in which any wrongs to the general body of share owners were compounded by the suing stockholder, who was mollified by payments from corporate assets. These litigations were aptly characterized in professional slang as "strike suits." And it was said that these suits were more commonly brought by small and irresponsible than by large stockholders, because the former put less to risk and a small interest was more often within the capacity and readiness of management to compromise than a large one.

We need not determine the measure of these abuses or the evils they produced on the one hand or prevented [549] and redressed on the other. The Legislature of New Jersey, like that of other states,^[3] considered them sufficient to warrant some remedial measures.

The very nature of the stockholder's derivative action makes it one in the regulation of which the legislature of a state has wide powers. Whatever theory one may hold as to the nature of the corporate entity, it remains a wholly artificial creation whose internal relations between management and stockholders are dependent upon state law and may be subject to most complete and penetrating regulation, either by public authority or by some form of stockholder action. Directors and managers, if not technically trustees, occupy positions of a fiduciary nature, and nothing in the Federal Constitution prohibits a state from imposing on them the strictest measure of responsibility, liability and accountability, either as a condition of assuming office or as a consequence of holding it.

Likewise, a stockholder who brings suit on a cause of action derived from the corporation assumes a position, not technically as a trustee perhaps, but one of a fiduciary character. He sues, not for himself alone, but as representative of a class comprising all who are similarly situated. The interests of all in the redress of the wrongs are taken into his hands, dependent upon his diligence, wisdom and integrity. And while the stockholders have chosen the corporate director or manager, they have no such election as to a plaintiff who steps forward to represent them. He is a self-chosen representative and a volunteer champion. The Federal Constitution does not oblige the state to place its litigating and adjudicating processes at the disposal of such a [550] representative, at least without imposing standards of responsibility, liability and accountability which it considers will protect the interests he elects himself to represent. It is not without significance that this Court has found it necessary long ago in the Equity Rules^[4] and now in the Federal Rules of Civil Procedure^[5] to impose procedural regulations of the class action not applicable to any other. We conclude that the state has plenary power over this type of litigation.

In considering specific objections to the way in which the state has exercised its power in this particular statute, it should be unnecessary to say that we are concerned only with objections which go to constitutionality. The wisdom and the policy of this and similar statutes are involved in controversies amply debated in legal literature^[6] but not for us to judge, and hence not for us to remark upon. The Federal Constitution does not invalidate state legislation because it fails to embody the [551] highest wisdom or provide the best conceivable remedies. Nor can legislation be set aside by courts because of the fact, if it be such, that it has been sponsored and promoted by those who advantage from it.^[7] In dealing with such difficult and controversial subjects, only experience will verify or disclose weaknesses and defects of any policy and teach lessons which may be applied by amendment. Within the area of constitutionality, the states should not be restrained from devising experiments, even those we might think dubious, in the effort to preserve the maximum good which equity sought in creating the derivative stockholder's action and at the same time to eliminate as much as possible its defects and evils.

It is said that this statute transgresses the Due Process Clause by being "arbitrary, capricious and unreasonable"; the Equal Protection Clause by singling out small stockholders to burden most heavily; that it violates the Contract Clause; and that its application to pending litigation renders it unconstitutionally retroactive.

The contention that this statute violates the Contract Clause of the Constitution is one in which we see not the slightest merit. Plaintiff's suit is entertained by equity largely because he had no contract rights on which to base an action at law, and hence none which is impaired by this legislation.

In considering whether the statute offends the Due Process Clause we can judge it only by its own terms, for it has had no interpretation or application as yet. It imposes liability and requires security for "the reasonable expenses, including counsel fees, which may be incurred" (emphasis supplied) by the corporation and by other parties defendant. The amount of security is subject to increase if the progress of the litigation reveals [552] that it is inadequate, or to decrease if it is proved to be excessive. A state may set the terms on which it will permit litigations in its courts. No type of litigation is more susceptible of regulation than that of a fiduciary nature. And it cannot seriously be said that a state makes such unreasonable use of its power as to violate the Constitution when it provides liability and security for payment of reasonable expenses if a litigation of this character is adjudged to be unsustainable. It is urged that such a requirement will foreclose resort by most stockholders to the only available judicial remedy for the protection of their rights. Of course, to require security for the payment of any kind of costs, or the necessity for bearing any kind of expense of litigation, has a deterring effect. But we deal with power, not wisdom; and we think, notwithstanding this tendency, it is within the power of a state to close its courts to this type of litigation if the condition of reasonable security is not met.

The contention that the statute denies equal protection of the laws is based upon the fact that it enables a stockholder who owns 5% of a corporation's outstanding shares, or $50,000 in market value, to proceed without either security or liability and imposes both upon those who elect to proceed with a smaller interest. We do not think the state is forbidden to use the amount of one's financial interest, which measures his individual injury from the misconduct to be redressed, as some measure of the good faith and responsibility of one who seeks at his own election to act as custodian of the interests of all stockholders, and as an indication that he volunteers for the large burdens of the litigation from a real sense of grievance and is not putting forward a claim to capitalize personally on its harassment value. These may not be the best ways of precluding "strike lawsuits," but we are unable to say that a classification for these purposes, [553] based upon the percentage or market value of the stock alleged to be injured by the wrongs, is an unconstitutional one. Where any classification is based on a percentage or an amount, it is necessarily somewhat arbitrary. It is difficult to say of many lines drawn by legislation that they give those just above and those just below the line a perfectly equal protection. A taxpayer with $10,000.01 of income does not think it is equality to tax him at a different rate than one who has $9,999.99, or to require returns from one just above and not from one just below a certain figure. It is difficult to say that a stockholder who has 49.99% of a company's stock should be unable to elect any representative to its Board of Directors while one who owns 50.01% may name the entire Board. If there is power, as we think there is, to draw a line based on considerations of proportion of amount, it is a rare case, of which this is not one, that a constitutional objection may be made to the particular point which the legislature has chosen.

The contention also is made that the provision which applies this statute to actions pending upon its enactment, in which no final judgment has been entered, renders it void under the Due Process Clause for retroactivity. While by its terms the statute applies to pending cases, it does not provide the manner of application; nor do the New Jersey courts appear to have settled what its effect is to be. Its terms do not appear to require an interpretation that it creates new liability against the plaintiff for expenses incurred by the defense previous to its enactment. The statute would admit of a construction that plaintiff's liability begins only from the time when the Act was passed or perhaps when the corporation's application for security is granted and that security for expenses and counsel fees which "may be incurred" does not include those which have been incurred [554] before one or the other of these periods. We would not, for the purpose of considering constitutionality, construe the statute in absence of a state decision as imposing liability for events before its enactment. On this basis its alleged retroactivity amounts only to a stay of further proceedings unless and until security is furnished for expense incurred in the future, and does not extend either to destruction of an existing cause of action or to creation of a new liability for past events.

The mere fact that a statute applies to a civil action retrospectively does not render it unconstitutional. Blount v. Windley, 95 U.S. 173, 180; Western Union Telegraph Co. v. L. & N.R. Co., 258 U.S. 13; Chase Securities Corp. v. Donaldson, 325 U.S. 304. Looking upon the statute as we have indicated, its retroactive effect, if any, is certainly less drastic and prejudicial than that held not to be unconstitutional in these decisions. We do not find in the bare statute any such retroactive effect as renders it unconstitutional under the Due Process Clause, and of course we express no opinion as to the effect of an application other than we have indicated.

It is also contended that this statute may not be applied in this case because the cause of action derives from a Delaware corporation and hence Delaware law governs it. But it is the plaintiff who has brought the case in New Jersey. The trial will very likely involve questions of conflict of laws as to which the law of New Jersey will apply, Klaxon Co. v. Stentor Co., 313 U.S. 487; Griffin v. McCoach, 313 U.S. 498, and perhaps questions of full faith and credit. These are not before us now. A plaintiff cannot avail himself of the New Jersey forum and at the same time escape the terms on which it is made available, if the law is applicable to a federal court sitting in that State, which we later consider.

We conclude, therefore, that, so far as the Federal Constitution is concerned, New Jersey's security statute is [555] a valid law of that State and the question remains as to whether it must be applied by federal courts in that State to suits brought therein on diversity grounds.

APPLICABILITY IN FEDERAL COURT.

The Rules of Decision Act, in effect since the First Congress of the United States and now found at 28 U.S.C. § 1652, provides: "The laws of the several states, except where the Constitution or treaties of the United States or Acts of Congress otherwise require or provide, shall be regarded as rules of decision in civil actions in the courts of the United States, in cases where they apply." This Court in Erie R. Co. v. Tompkins 304 U.S. 64, held that judicial decisions are laws of the states within its meaning. But Erie R. Co. v. Tompkins and its progeny have wrought a more far-reaching change in the relation of state and federal courts and the application of state law in the latter whereby in diversity cases the federal court administers the state system of law in all except details related to its own conduct of business. Guaranty Trust Co. v. York, 326 U.S. 99. The only substantial argument that this New Jersey statute is not applicable here is that its provisions are mere rules of procedure rather than rules of substantive law.

Even if we were to agree that the New Jersey statute is procedural, it would not determine that it is not applicable. Rules which lawyers call procedural do not always exhaust their effect by regulating procedure. But this statute is not merely a regulation of procedure. With it or without it the main action takes the same course. However, it creates a new liability where none existed before, for it makes a stockholder who institutes a derivative action liable for the expense to which he puts the corporation and other defendants, if he does not make good his claims. Such liability is not usual and it goes beyond payment of what we know as "costs." If all [556] the Act did was to create this liability, it would clearly be substantive. But this new liability would be without meaning and value in many cases if it resulted in nothing but a judgment for expenses at or after the end of the case. Therefore, a procedure is prescribed by which the liability is insured by entitling the corporate defendant to a bond of indemnity before the outlay is incurred. We do not think a statute which so conditions the stockholder's action can be disregarded by the federal court as a mere procedural device.

It is urged, however, that Federal Rule of Civil Procedure No. 23 deals with plaintiff's right to maintain such an action in federal court and that therefore the subject is recognized as procedural and the federal rule alone prevails. Rule 23 requires the stockholder's complaint to be verified by oath and to show that the plaintiff was a stockholder at the time of the transaction of which he complains or that his share thereafter devolved upon him by operation of law. In other words, the federal court will not permit itself to be used to litigate a purchased grievance or become a party to speculation in wrongs done to corporations. It also requires a showing that an action is not a collusive one to confer jurisdiction and to set forth the facts showing that the plaintiff has endeavored to obtain his remedy through the corporation itself. It further provides that the class action shall not be dismissed or compromised without approval of the court, with notice to the members of the class. These provisions neither create nor exempt from liabilities, but require complete disclosure to the court and notice to the parties in interest. None conflict with the statute in question and all may be observed by a federal court, even if not applicable in state court.

We see no reason why the policy stated in Guaranty Trust Co. v. York, 326 U.S. 99, should not apply.

[557] We hold that the New Jersey statute applies in federal courts and that the District Court erred in declining to fix the amount of indemnity reasonably to be exacted as a condition of further prosecution of the suit.

The judgment of the Court of Appeals is

Affirmed.

MR. JUSTICE DOUGLAS, with whom MR. JUSTICE FRANKFURTER concurs, dissenting in part.

The cause of action on which this suit is brought is a derivative one. Though it belongs to the corporation, the stockholders are entitled under state law to enforce it. The measure of the cause of action is the claim which the corporation has against the alleged wrongdoers. This New Jersey statute does not add one iota to nor subtract one iota from that cause of action. It merely prescribes the method by which stockholders may enforce it. Each state has numerous regulations governing the institution of suits in its courts. They may favor the litigation or they may affect it adversely. But they do not fall under the principle of Erie R. Co. v. Tompkins, 304 U.S. 64, unless they define, qualify or delimit the cause of action or otherwise relate to it.

This New Jersey statute, like statutes governing security for costs, regulates only the procedure for instituting a particular cause of action and hence need not be applied in this diversity suit in the federal court. Rule 23 of the Federal Rules of Civil Procedure defines that procedure for the federal courts.

MR. JUSTICE RUTLEDGE, dissenting.

I am in accord with the dissenting opinion of MR. JUSTICE DOUGLAS in this case. I also agree with the dissenting views of MR. JUSTICE JACKSON in No. 465, Woods v. [558] Interstate Realty Co., ante, p. 538, decided today. And I have noted my dissent in No. 522, Ragan v. Merchants Transfer & Warehouse Co., ante, p. 530, also decided today.

Without undertaking to discuss each case in detail, I think the three decisions taken together demonstrate the extreme extent to which the Court is going in submitting the control of diversity litigation in the federal courts to the states rather than to Congress, where it properly belongs. This is done in the guise of applying the rule of Erie R. Co. v. Tompkins, 304 U.S. 64. But in my opinion it was never the purpose of that decision to put such matters as those involved here outside the power of Congress to regulate and to confer that authority exclusively upon the states.

What is being applied is a gloss on the Erie rule, not the rule itself. That case held that federal courts in diversity cases must apply state law, decisional as well as statutory, in determining matters of substantive law, in particular and apart from procedural limitations upon its assertion — whether a cause of action exists. I accept that view generally and insofar as it involves a wise rule of administration for the federal courts, though I have grave doubt that it has any solid constitutional foundation.

But the Erie case made no ruling that in so deciding diversity cases a federal court is "merely another court of the state in which it sits." and hence that in every situation in which the doors of state courts are closed to a suitor, so must be also those of the federal courts. Not only is this not true when the state bar is raised by a purely procedural obstacle. There is sound historical reason for believing that one of the purposes of the diversity clause was to afford a federal court remedy when, for at least some reasons of state policy, none would be available in the state courts. It is the gloss which has [559] been put upon the Erie ruling by later decisions, e.g., Guaranty Trust Co. v. York, 326 U.S. 99, which in my opinion is being applied to extend the Erie ruling far beyond its original purpose or intent and, in my judgment, with consequences and implications seriously impairing Congress' power, within its proper sphere of action, to control this type of litigation in the federal courts.

The accepted dichotomy is the familiar "procedural-substantive" one. This of course is a subject of endless discussion, which hardly needs to be repeated here. Suffice it to say that actually in many situations procedure and substance are so interwoven that rational separation becomes well-nigh impossible. But, even so, this fact cannot dispense with the necessity of making a distinction. For, as the matter stands, it is Congress which has the power to govern the procedure of the federal courts in diversity cases, and the states which have that power over matters clearly substantive in nature. Judges therefore cannot escape making the division. And they must make it where the two constituent elements are Siamese twins as well as where they are not twins or even blood brothers. The real question is not whether the separation shall be made, but how it shall be made: whether mechanically by reference to whether the state courts' doors are open or closed, or by a consideration of the policies which close them and their relation to accommodating the policy of the Erie rule with Congress' power to govern the incidents of litigation in diversity suits.

It is in these close cases, this borderland area, that I think we are going too far. It is one thing to decide that Pennsylvania does or does not create a cause of action in tort for injuries inflicted by specified conduct and to have that determination govern the outcome of [560] a diversity suit in Pennsylvania or New York.^[8] It is another, in my view, to require a bond for costs or for payment of the opposing party's expenses and attorney's fees in the event the claimant is unsuccessful. Whether or not the latter is conceived as creating a new substantive right, it is too close to controlling the incidents of the litigation rather than its outcome to be identified with the former. It is a matter which in my opinion lies within Congress' control for diversity cases, not one for state control or to be governed by the fact that the state shuts the doors of its courts unless the state requirements concerning such incidents of litigation are complied with.

In my view Rule 23 of the Federal Rules of Civil Procedure, derived from the former Equity Rules and now having the sanction of Congress, is valid and governs in the Cohen case. If, however, the State of New Jersey has the power to govern federal diversity suits within its [561] borders as to all matters having a substantive tinge or aspect, then it may be questioned whether, in the event of conflict with some local policy, a federal court sitting in that state could give effect to the Rule's requirement that the complaint aver "that the plaintiff was a shareholder at the time of the transaction of which he complains or that his share thereafter devolved on him by operation of law. . . ." For in any strict and abstract sense that provision would seem to be as much a "substantive" one as the New Jersey requirements for bond, etc. And, if so, then it would seem highly doubtful, on any automatic or mechanical application of the substantive-procedural dichotomy, that either Congress or this Court could create such a limitation on diversity litigation, since as a substantive matter this would be for the states to control. See 3 Moore, Federal Practice (2d ed., 1948) 3493-3506.

For myself I have no doubt of the validity of Rule 23 or of the power of Congress to enact such a rule, even though it has a substantive aspect. Notwithstanding that aspect, the rule is too closely related to procedural and other matters affecting litigation in the federal courts for me to conceive of its invalidity. So also in the present cases I think the state regulations, though each may be regarded as having a substantive aspect, are too closely related to the modes and methods of conducting litigation in the federal courts to be capable of displacing Congress' power of regulation in those respects or the federal courts' power to hear and determine the respective controversies.

Accordingly I would reverse the judgments in the Cohen and Ragan cases and affirm that in the Woods case.

[1] Together with No. 512, Beneficial Industrial Loan Corp. v. Smith, United States District Judge, et al., also on certiorari to the same Court.

[2]Chapter 131, New Jersey Laws of 1945, provides in pertinent part as follows:

"1. In any action instituted or maintained in the right of any domestic or foreign corporation by the holder or holders of shares, or of voting trust certificates representing shares, of such corporation having a total par value or stated capital value of less than five per centum (5%) of the aggregate par value or stated capital value of all the outstanding shares of such corporation's stock of every class . . . unless the shares or voting trust certificates held by such holder or holders have a market value in excess of fifty thousand dollars ($50,000.00), the corporation in whose right such action is brought shall be entitled, at any stage of the proceeding before final judgment, to require the complainant or complainants to give security for the reasonable expenses, including counsel fees, which may be incurred by it in connection with such action and by the other parties defendant in connection therewith for which it may become subject pursuant to law, its certificate of incorporation, its by-laws or under equitable principles, to which the corporation shall have recourse in such amount as the court having jurisdiction shall determine upon the termination of such action. The amount of such security may thereafter, from time to time, be increased or decreased in the discretion of the court having jurisdiction of such action upon showing that the security provided has or may become inadequate or is excessive.

"2. In any action, suit or proceeding brought or maintained in the right of a domestic or foreign corporation by the holder or holders of shares, or of voting trust certificates representing shares, of such corporation, it must be made to appear that the complainant was a shareholder or the holder of a voting trust certificate at the time of the transaction of which he complains or that his share or voting trust certificate thereafter devolved upon him by operation of law.

"3. This act shall take effect immediately and shall apply to all such actions, suits or proceedings now pending in which no final judgment has been entered, and to all future actions, suits and proceedings."

[3] See New York General Corporation Law. § 61-b; 12 Pa. Stat. Ann. § 1322; Laws of Maryland, 1945, c. 989; Wisconsin Stat. § 180.13 (1945).

[4] Old Equity Rule 94, 104 U.S. ix; Equity Rule 27, 226 U.S. 649, 656.

[5] Rule 23 (b).

[6] See Hornstein, Problems of Procedure in Stockholder's Derivative Suits, 42 Col. L.R. 574; Hornstein, Directors' Expenses in Stockholders' Suits, 43 id. 301; Koessler, The Stockholder's Suit; A Comparative View, 46 id. 238; Hornstein, New Aspects of Stockholders' Derivative Suits, 47 id. 1; Carson, Current Phases of Derivative Actions Against Directors, 40 Mich. L.R. 1125; P.E. Jackson, Reorganization of the Corporate Concept And the Effect of Section 61-b of the New York General Corporation Law, A5 Am. Bankr. Rev. 323; Carson, Further Phases of Derivative Actions Against Directors, 29 Cornell L.Q. 431; House, Stockholders' Suits And the Coudert-Mitchell Laws, 20 N.Y.U.L.Q. Rev. 377; Hornstein, The Death Knell of Stockholders' Derivative Suits in New York, 32 California L.R. 123; Zlinkoff. The American Investor And the Constitutionality of Section 61-b of the New York General Corporation Law, 54 Yale L.J. 352. See Douglas, Directors Who Do Not Direct, 47 Harv. L.R. 1305.

[7] Daniel v. Family Insurance Co., 336 U.S. 220.

[8] It may be noted that the disposition of the local law problem apparently presented in Erie was not consistent, either here or on remand, with the current view that a federal district court is required to treat a diversity case exactly as would a state court of the state in which the district court is sitting: The Erie case arose out of an alleged Pennsylvania tort, and this Court stated that the court of appeals had erred when it "declined to decide the issue of state law," 304 U.S. at 80 — i.e., "the Pennsylvania law." Ibid. But the Erie case was initiated by Tompkins, "a citizen of Pennsylvania . . . in the federal court for southern New York, which had jurisdiction because the company is a corporation of that State." 304 U.S. at 69 (emphasis added). Accordingly, as Erie is now construed, the issue on remand should have been what law a New York state court would have applied to the Pennsylvania tort. But the sole issue determined on remand was the applicable Pennsylvania law, without mention of the probable attitude of the New York courts. Tompkins v. Erie R. Co., 98 F.2d 49. It was not until after Justice Brandeis had retired that this Court held that federal district courts were required to follow local conflict of laws doctrine in the resolution of diversity cases. Klaxon Co. v. Stentor Co., 313 U.S. 487.