WINTER, SECRETARY OF THE NAVY, et al. v. NATURAL RESOURCES DEFENSE COUNCIL, INC., et al.

No. 07-1239.

Argued October 8, 2008

Decided November 12, 2008

*10Roberts, C. J., delivered the opinion of the Court, in which Scalia, Kennedy, Thomas, and Alito, JJ., joined. Breyer, J., filed an opinion *11concurring in part and dissenting in part, in which Stevens, J., joined as to Part I, post, p. 34. Ginsburg, J., filed a dissenting opinion, in which Souter, J., joined, post, p. 43.

Solicitor General Garre argued the cause for petitioners. With him on the briefs were Assistant Attorney General Tenpas, Deputy Solicitor General Kneedler, Deputy Assistant Attorney General Nelson, Anthony A. Yang, Andrew C. Mergen, Michael R. Eitel, Luther L. Hajek, Allen M. Brdbender, Daniel J. DelVOrto, Thomas N. Ledvina, J. Page Turney, Jane C. Luxton, Mary Beth Ward, Joel La Bissonniere, and Edward A. Boling.

Richard B. Kendall argued the cause for respondents. With him on the brief for respondent environmentalists were Robert N. Klieger, Gregory A. Fayer, Joel R. Reynolds, Cara A. Horowitz, and Stephen Zak Smith.

Edmund G. Brown, Jr., Attorney General of California, filed a brief for respondent California Coastal Commission. With him on the brief were Jarme Jordan Patterson, Supervising Deputy Attorney General, J. Matthew Rodriquez, Chief Assistant Attorney General, Manuel M. Medeiros, Solicitor General, and Gordon Burns, Deputy Solicitor General.*

*

Briefs of amici curiae urging reversal were filed for the Navy League of the United States — Honolulu Council et al. by Robert H. Thomas, Mark M. Murakami, Theodore G. Meeker, and Michael A. Lilly; for the Pacific Legal Foundation by M. Reed Hopper and Steven Geoffrey Gieseler; and for the Washington Legal Foundation et al. by Andrew G. McBride, Daniel J. Popeo, and Paul D. Kamenar.

Briefs of amici curiae urging affirmance were filed for the Ecological Society of America by Peter E. Perkowski; for California Assembly Member Julia Brownley et al. by Stephen B. Kinnaird, Peter H. Weiner, Paul W. Cane, Jr., and Sean D. Unger; and for Michael C. Small et al. by Mr. Small, pro se, Jonathan D. Varat, and Edward P. Lazarus.

Briefs of amici curiae were filed for the California Forestry Association et al. by Thomas R. Dundquist, J. Michael Klise, Michele Dias, William R. Murray, Douglas T. Nelson, Duane J. Desiderio, and Thomas J. Ward; and for Defenders of Wildlife et al. by Eric R. Glitzenstein.

*12Chief Justice Roberts

delivered the opinion of the Court.

“To be prepared for war is one of the most effectual means of preserving peace.” 1 Messages and Papers of the Presidents 57 (J. Richardson comp. 1897). So said George Washington in his first Annual Address to Congress, 218 years ago. One of the most important ways the Navy prepares for war is through integrated training exercises at sea. These exercises include training in the use of modern sonar to detect and track enemy submarines, something the Navy has done for the past 40 years. The plaintiffs, respondents here, complained that the Navy’s sonar-training program harmed marine mammals, and that the Navy should have prepared an environmental impact statement before commencing its latest round of training exercises. The Court of Appeals upheld a preliminary injunction imposing restrictions on the Navy’s sonar training, even though that court acknowledged that “the record contains no evidence that marine mammals have been harmed” by the Navy’s exercises. 518 F. 3d 658, 696 (CA9 2008).

The Court of Appeals was wrong, and its decision is reversed.

I

The Navy deploys its forces in “strike groups,” which are groups of surface ships, submarines, and aircraft centered around either an aircraft carrier or an amphibious assault ship. App. to Pet. for Cert. 316a-317a (Pet. App.). Seamless coordination among strike-group assets is critical. Before deploying a strike group, the Navy requires extensive integrated training in analysis and prioritization of threats, execution of military missions, and maintenance of force protection. App. 110-111.

Antisubmarine warfare is currently the Pacific Fleet’s top war-fighting priority. Pet. App. 270a-271a. Modern diesel-electric submarines pose a significant threat to Navy vessels because they can operate almost silently, making them ex*13tremely difficult to detect and track. Potential adversaries of the United States possess at least 300 of these submarines. App. 571.

The most effective technology for identifying submerged diesel-electric submarines within their torpedo range is active sonar, which involves emitting pulses of sound underwater and then receiving the acoustic waves that echo off the target. Pet. App. 266a-267a, 274a. Active sonar is a particularly useful tool because it provides both the bearing and the distance of target submarines; it is also sensitive enough to allow the Navy to track enemy submarines that are quieter than the surrounding marine environment.1 This case concerns the Navy’s use of “mid-frequency active” (MFA) sonar, which transmits sound waves at frequencies between 1 kHz and 10 kHz.

Not surprisingly, MFA sonar is a complex technology, and sonar operators must undergo extensive training to become proficient in its use. Sonar reception can be affected by countless different factors, including the time of day, water density, salinity, currents, weather conditions, and the contours of the sea floor. Id., at 278a-279a. When working as part of a strike group, sonar operators must be able to coordinate with other Navy ships and planes while avoiding interference. The Navy conducts regular training exercises under realistic conditions to ensure that sonar operators are thoroughly skilled in its use in a variety of situations.

The waters off the coast of southern California (SOCAL) are an ideal location for conducting integrated training exercises, as this is the only area on the west coast that is relatively close to land, air, and sea bases, as well as amphibious *14landing areas. App. 141-142. At issue in this case are the Composite Training Unit Exercises and the Joint Tactical Force Exercises, in which individual naval units (ships, submarines, and aircraft) train together as members of a strike group. A strike group cannot be certified for deployment until it has successfully completed the integrated training exercises, including a demonstration of its ability to operate under simulated hostile conditions. Id., at 564-565. In light of the threat posed by enemy submarines, all strike groups must demonstrate proficiency in antisubmarine warfare. Accordingly, the SOCAL exercises include extensive training in detecting, tracking, and neutralizing enemy submarines. The use of MFA sonar during these exercises is “mission-critical,” given that MFA sonar is the only proven method of identifying submerged diesel-electric submarines operating on battery power. Id., at 568-571.

Sharing the waters in the SOCAL operating area are at least 37 species of marine mammals, including dolphins, whales, and sea lions. The parties strongly dispute the extent to which the Navy’s training activities will harm those animals or disrupt their behavioral patterns. The Navy emphasizes that it has used MFA sonar during training exercises in SOCAL for 40 years, without a single documented sonar-related injury to any marine mammal. The Navy asserts that, at most, MFA sonar may cause temporary hearing loss or brief disruptions of marine mammals’ behavioral patterns.

The plaintiffs are the Natural Resources Defense Council, Inc., Jean-Michael Cousteau (an environmental enthusiast and filmmaker), and several other groups devoted to the protection of marine mammals and ocean habitats. They contend that MFA sonar can cause much more serious injuries to marine mammals than the Navy acknowledges, including permanent hearing loss, decompression sickness, and major behavioral disruptions. According to the plaintiffs, several mass strandings of marine mammals (outside of SOCAL) *15have been “associated” with the use of active sonar. They argue that certain species of marine mammals — such as beaked whales — are uniquely susceptible to injury from active sonar; these injuries would not necessarily be detected by the Navy, given that beaked whales are “very deep divers” that spend little time at the surface.

II

The procedural history of this case is rather complicated. The Marine Mammal Protection Act of 1972 (MMPA), 86 Stat. 1027, generally prohibits any individual from “taking” a marine mammal, defined as harassing, hunting, capturing, or killing it. 16 U. S. C. §§ 1362(13), 1372(a). The Secretary of Defense may “exempt any action or category of actions” from the MMPA if such actions are “necessary for national defense.” § 1371(f)(1). In January 2007, the Deputy Secretary of Defense — acting for the Secretary — granted the Navy a 2-year exemption from the MMPA for the training exercises at issue in this case. Pet. App. 219a-220a. The exemption was conditioned on the Navy adopting several mitigation procedures, including: (1) training lookouts and officers to watch for marine mammals; (2) requiring at least five lookouts with binoculars on each vessel to watch for anomalies on the water surface (including marine mammals); (3) requiring aircraft and sonar operators to report detected marine mammals in the vicinity of the training exercises; (4) requiring reduction of active sonar transmission levels by 6 dB if a marine mammal is detected within 1,000 yards of the bow of the vessel, or by 10 dB if detected within 500 yards; (5) requiring complete shutdown of active sonar transmission if a marine mammal is detected within 200 yards of the vessel; (6) requiring active sonar to be operated at the “lowest practicable level”; and (7) adopting coordination and reporting procedures. Id., at 222a-230a.

The National Environmental Policy Act of 1969 (NEPA), 83 Stat. 852, requires federal agencies “to the fullest extent *16possible” to prepare an environmental impact statement (EIS) for “every . . . major Federal actio[n] significantly affecting the quality of the human environment.” 42 U. S. C. §4332(2)(C) (2000 ed.). An agency is not required to prepare a full EIS if it determines — based on a shorter environmental assessment (EA) — that the proposed action will not have a significant impact on the environment. 40 CFR §§ 1508.9(a), 1508.13 (2007).

In February 2007, the Navy issued an EA concluding that the 14 SOCAL training exercises scheduled through January 2009 would not have a significant impact on the environment. App. 226-227. The EA divided potential injury to marine mammals into two categories: Level A harassment, defined as the potential destruction or loss of biological tissue (1 e., physical injury), and Level B harassment, defined as temporary injury or disruption of behavioral patterns such as migration, feeding, surfacing, and breeding. Id., at 160-161.

The Navy’s computer models predicted that the SOCAL training exercises would cause only eight Level A harassments of common dolphins each year, and that even these injuries could be avoided through the Navy’s voluntary mitigation measures, given that dolphins travel in large pods easily located by Navy lookouts. Id., at 176-177, 183. The EA also predicted 274 Level B harassments of beaked whales per year, none of which would result in permanent injury. Id., at 185-186. Beaked whales spend little time at the surface, so the precise effect of active sonar on these mammals is unclear. Erring on the side of caution, the Navy classified all projected harassments of beaked whales as Level A. Id., at 186, 223. In light of its conclusion that the SOCAL training exercises would not have a significant impact on the environment, the Navy determined that it was unnecessary to prepare a full EIS. See 40 CFR § 1508.13.

Shortly after the Navy released its EA, the plaintiffs sued the Navy, seeking declaratory and injunctive relief on the grounds that the Navy’s SOCAL training exercises violated *17NEPA, the Endangered Species Act of 1973 (ESA), and the Coastal Zone Management Act of 1972 (CZMA).2 The District Court granted plaintiffs’ motion for a preliminary injunction and prohibited the Navy from using MFA sonar during its remaining training exercises. The court held that plaintiffs had “demonstrated a probability of success” on their claims under NEPA and the CZMA. Pet. App. 207a, 215a. The court also determined that equitable relief was appropriate because, under Ninth Circuit precedent, plaintiffs had established at least a “ ‘possibility’ ” of irreparable harm to the environment. Id., at 217a. Based on scientific studies, declarations from experts, and other evidence in the record, the District Court concluded that there was in fact a “near certainty” of irreparable injury to the environment, and that this injury outweighed any possible harm to the Navy. Id., at 217a-218a.

The Navy filed an emergency appeal, and the Ninth Circuit stayed the injunction pending appeal. 502 F. 3d 859, 865 (2007). After hearing oral argument, the Court of Appeals agreed with the District Court that preliminary injunctive relief was appropriate. The appellate court concluded, however, that a blanket injunction prohibiting the Navy from using MFA sonar in SOCAL was overbroad, and remanded the case to the District Court “to narrow its injunction so as to provide mitigation conditions under which the Navy may conduct its training exercises.” 508 F. 3d 885, 887 (2007).

On remand, the District Court entered a new preliminary injunction allowing the Navy to use MFA sonar only as long as it implemented the following mitigation measures (in addition to the measures the Navy had adopted pursuant to its MMPA exemption): (1) imposing a 12 nautical mile “exclusion *18zone” from the coastline; (2) using lookouts to conduct additional monitoring for marine mammals; (3) restricting the use of “helicopter-dipping” sonar; (4) limiting the use of MFA sonar in geographic “choke points”; (5) shutting down MFA sonar when a marine mammal is spotted within 2,200 yards of a vessel; and (6) powering down MFA sonar by 6 dB during significant surface ducting conditions, in which sound travels further than it otherwise would due to temperature differences in adjacent layers of water. 530 F. Supp. 2d 1110, 1118-1121 (CD Cal. 2008). The Navy filed a notice of appeal, challenging only the last two restrictions.

The Navy then sought relief from the Executive Branch. The President, pursuant to 16 U. S. C. § 1456(c)(1)(B), granted the Navy an exemption from the CZMA. Section 1456(c)(1)(B) permits such exemptions if the activity in question is “in the paramount interest of the United States.” The President determined that continuation of the exercises as limited by the Navy was “essential to national security.” Pet. App. 232a. He concluded that compliance with the District Court’s injunction would “undermine the Navy’s ability to conduct realistic training exercises that are necessary to ensure the combat effectiveness of... strike groups.” Ibid.

Simultaneously, the Council on Environmental Quality (CEQ) authorized the Navy to implement “alternative arrangements” to NEPA compliance in light of “emergency circumstances.” See 40 CFR § 1506.11.3 The CEQ determined that alternative arrangements were appropriate because the District Court’s injunction “create[s] a significant and unreasonable risk that Strike Groups will not be *19able to train and be certified as fully mission capable.” Pet. App. 238a. Under the alternative arrangements, the Navy would be permitted to conduct its training exercises under the mitigation procedures adopted in conjunction with the exemption from the MMPA. The CEQ also imposed additional notice, research, and reporting requirements.

In light of these actions, the Navy then moved to vacate the District Court’s injunction with respect to the 2,200-yard shutdown zone and the restrictions on training in surface ducting conditions. The District Court refused to do so, 527 F. Supp. 2d 1216 (2008), and the Court of Appeals affirmed. The Ninth Circuit held that there was a serious question regarding whether the CEQ’s interpretation of the “emergency circumstances” regulation was lawful. Specifically, the court questioned whether there was a true “emergency” in this case, given that the Navy has been on notice of its obligation to comply with NEPA from the moment it first planned the SOCAL training exercises. 518 F. 3d, at 681. The Court of Appeals concluded that the preliminary injunction was entirely predictable in light of the parties’ litigation history. Ibid. The court also held that plaintiffs had established a likelihood of success on their claim that the Navy was required to prepare a full EIS for the SOCAL training exercises. Id., at 693. The Ninth Circuit agreed with the District Court’s holding that the Navy’s EA — which resulted in a finding of no significant environmental impact — was “cursory, unsupported by cited evidence, or unconvincing.” Ibid.4

The Court of Appeals further determined that plaintiffs had carried their burden of establishing a “possibility” of irreparable injury. Even under the Navy’s own figures, the court concluded, the training exercises would cause 564 physical injuries to marine mammals, as well as 170,000 disturb*20anees of marine mammals’ behavior. Id., at 696. Lastly, the Court of Appeals held that the balance of hardships and consideration of the public interest weighed in favor of the plaintiffs. The court emphasized that the negative impact on the Navy’s training exercises was “speculative,” since the Navy has never before operated under the procedures required by the District Court. Id., at 698-699. In particular, the court determined that: (1) The 2,200-yard shutdown zone imposed by the District Court was unlikely to affect the Navy’s operations, because the Navy often shuts down its MFA sonar systems during the course of training exercises; and (2) the power-down requirement during significant surface ducting conditions was not unreasonable because such conditions are rare, and the Navy has previously certified strike groups that had not trained under such conditions. Id., at 699-702. The Ninth Circuit concluded that the District Court’s preliminary injunction struck a proper balance between the competing interests at stake.

We granted certiorari, 554 U. S. 916 (2008), and now reverse and vacate the injunction.

Ill

A

A plaintiff seeking a preliminary injunction must establish that he is likely to succeed on the merits, that he is likely to suffer irreparable harm in the absence of preliminary relief, that the balance of equities tips in his favor, and that an injunction is in the public interest. See Munaf v. Geren, 553 U. S. 674, 689-690 (2008); Amoco Production Co. v. Gambell, 480 U. S. 531, 542 (1987); Weinberger v. Romero-Barcelo, 456 U. S. 305, 311-312 (1982).

The District Court and the Ninth Circuit concluded that plaintiffs have shown a likelihood of success on the merits of their NEPA claim. The Navy strongly disputes this determination, arguing that plaintiffs’ likelihood of success is low because the CEQ reasonably concluded that “emergency *21circumstances” justified alternative arrangements to NEPA compliance. 40 CFR § 1506.11. Plaintiffs’briefs before this Court barely discuss the ground relied upon by the lower courts — that the plain meaning of “emergency circumstances” does not encompass a court order that was “entirely predictable” in light of the parties’ litigation history. 518 F. 3d, at 681. Instead, plaintiffs contend that the CEQ’s actions violated the separation of powers by readjudicating a factual issue already decided by an Article III court. Moreover, they assert that the CEQ’s interpretations of NEPA are not entitled to deference because the CEQ has not been given statutory authority to conduct adjudications.

The District Court and the Ninth Circuit also held that when a plaintiff demonstrates a strong likelihood of prevailing on the merits, a preliminary injunction may be entered based only on a. “possibility” of irreparable harm. Id., at 696-697; 530 F. Supp. 2d, at 1118 (quoting Faith Center Church Evangelistic Ministries v. Glover, 480 F. 3d 891, 906 (CA9 2007); Earth Island Inst. v. United States Forest Serv., 442 F. 3d 1147,1159 (CA9 2006)). The lower courts held that plaintiffs had met this standard because the scientific studies, declarations, and other evidence in the record established to “a near certainty” that the Navy’s training exercises would cause irreparable harm to the environment. 530 F. Supp. 2d, at 1118.

The Navy challenges these holdings, arguing that plaintiffs must demonstrate a likelihood of irreparable injury— not just a possibility — in order to obtain preliminary relief. On the facts of this case, the Navy contends that plaintiffs’ alleged injuries are too speculative to give rise to irreparable injury, given that ever since the Navy’s training program began 40 years ago, there has been no documented case of sonar-related injury to marine mammals in SOCAL. And even if MFA sonar does cause a limited number of injuries to individual marine mammals, the Navy asserts that plaintiffs have failed to offer evidence of species-level harm that *22would adversely affect their scientific, recreational, and ecological interests. For their part, plaintiffs assert that they would prevail under any formulation of the irreparable injury standard, because the District Court found that they had established a “near certainty” of irreparable harm.

We agree with the Navy that the Ninth Circuit’s “possibility” standard is too lenient. Our frequently reiterated standard requires plaintiffs seeking preliminary relief to demonstrate that irreparable injury is likely in the absence of an injunction. Los Angeles v. Lyons, 461 U. S. 95, 103 (1983); Granny Goose Foods, Inc. v. Teamsters, 415 U. S. 423, 441 (1974); O’Shea v. Littleton, 414 U. S. 488, 502 (1974); see also 11A C. Wright, A. Miller, & M. Kane, Federal Practice and Procedure §2948.1, p. 139 (2d ed. 1995) (hereinafter Wright & Miller) (applicant must demonstrate that in the absence of a preliminary injunction, “the applicant is likely to suffer irreparable harm before a decision on the merits can be rendered”); id., at 154-155 (“[A] preliminary injunction will not be issued simply to prevent the possibility of some remote future injury”). Issuing a preliminary injunction based only on a possibility of irreparable harm is inconsistent with our characterization of injunctive relief as an extraordinary remedy that may only be awarded upon a clear showing that the plaintiff is entitled to such relief. Mazurek v. Armstrong, 520 U. S. 968, 972 (1997) (per curiam).

It is not clear that articulating the incorrect standard affected the Ninth Circuit’s analysis of irreparable harm. Although the court referred to the “possibility” standard, and cited Circuit precedent along the same lines, it affirmed the District Court’s conclusion that plaintiffs had established a “ ‘near certainty’ ” of irreparable harm. 518 F. 3d, at 696-697. At the same time, however, the nature of the District Court’s conclusion is itself unclear. The District Court originally found irreparable harm from sonar-training exercises generally. But by the time of the District Court’s final decision, the Navy challenged only two of six restrictions *23imposed by the court. See supra, at 17-19. The District Court did not reconsider the likelihood of irreparable harm in light of the four restrictions not challenged by the Navy. This failure is significant in light of the District Court’s own statement that the 12 nautical mile exclusion zone from the coastline — one of the unchallenged mitigation restrictions— “would bar the use of MFA sonar in a significant portion of important marine mammal habitat.” 530 F. Supp. 2d, at 1119.

We also find it pertinent that this is not a case in which the defendant is conducting a new type of activity with completely unknown effects on the environment. When the Government conducts an activity, “NEPA itself does not mandate particular results.” Robertson v. Methow Valley Citizens Council, 490 U. S. 332, 350 (1989). Instead, NEPA imposes only procedural requirements to “ensur[e] that the agency, in reaching its decision, will have available, and will carefully consider, detailed information concerning significant environmental impacts.” Id., at 349. Part of the harm NEPA attempts to prevent in requiring an EIS is that, without one, there may be little if any information about prospective environmental harms and potential mitigating measures. Here, in contrast, the plaintiffs are seeking to enjoin — or substantially restrict — training exercises that have been taking place in SOCAL for the last 40 years. And the latest series of exercises were not approved until after the defendant took a “hard look at environmental consequences,” id., at 350 (quoting Kleppe v. Sierra Club, 427 U. S. 390, 410, n. 21 (1976); internal quotation marks omitted), as evidenced by the issuance of a detailed, 293-page EA.

As explained in the next section, even if plaintiffs have shown irreparable injury from the Navy’s training exercises, any such injury is outweighed by the public interest and the Navy’s interest in effective, realistic training of its sailors. A proper consideration of these factors alone requires denial of the requested injunctive relief. For the same reason, we *24do not address the lower courts’ holding that plaintiffs have also established a likelihood of success on the merits.

B

A preliminary injunction is an extraordinary remedy never awarded as of right. Munaf, 553 U. S., at 689-690. In each case, courts “must balance the competing claims of injury and must consider the effect on each party of the granting or withholding of the requested relief.” Amoco Production Co., 480 U. S., at 542. “In exercising their sound discretion, courts of equity should pay particular regard for the public consequences in employing the extraordinary remedy of injunction.” Romero-Barcelo, 456 U. S., at 312; see also Railroad Comm’n of Tex. v. Pullman Co., 312 U. S. 496, 500 (1941). In this case, the District Court and the Ninth Circuit significantly understated the burden the preliminary injunction would impose on the Navy’s ability to conduct realistic training exercises, and the injunction’s consequent adverse impact on the public interest in national defense.

This case involves “complex, subtle, and professional decisions as to the composition, training, equipping, and control of a military force,” which are “essentially professional military judgments.” Gilligan v. Morgan, 413 U. S. 1, 10 (1973). We “give great deference to the professional judgment of military authorities concerning the relative importance of a particular military interest.” Goldman v. Weinberger, 475 U. S. 503, 507 (1986). As the Court emphasized just last Term, “neither the Members of this Court nor most federal judges begin the day with briefings that may describe new and serious threats to our Nation and its people.” Boumediene v. Bush, 553 U. S. 723, 797 (2008).

Here, the record contains declarations from some of the Navy’s most senior officers, all of whom underscored the threat posed by enemy submarines and the need for extensive sonar training to counter this threat. Admiral Gary *25Roughead — the Chief of Naval Operations — stated that during training exercises:

“It is important to stress the ship crews in all dimensions of warfare simultaneously. If one of these training elements were impacted — for example, if effective sonar training were not possible — the training value of the other elements would also be degraded . . . .” Pet. App. 342a.

Captain Martin May — the Third Fleet’s Assistant Chief of Staff for Training and Readiness — emphasized that the use of MFA sonar is “mission-critical.” App. 570-571. He described the ability to operate MFA sonar as a “highly perishable skill” that must be repeatedly practiced under realistic conditions. Id., at 577. During training exercises, MFA sonar operators learn how to avoid sound-reducing “clutter” from ocean floor topography and environmental conditions; they also learn how to avoid interference and how to coordinate their efforts with other sonar operators in the strike group. Id., at 574. Several Navy officers emphasized that realistic training cannot be accomplished under the two challenged restrictions imposed by the District Court — the 2,200-yard shutdown zone and the requirement that the Navy power down its sonar systems during significant surface ducting conditions. See, e. g., Pet. App. 333a (powering down in presence of surface ducting “unreasonably prevent[s] realistic training”); id., at 356a (shutdown zone would “result in a significant, adverse impact to realistic training”). We accept these officers’ assertions that the use of MFA sonar under realistic conditions during training exercises is of the utmost importance to the Navy and the Nation.

These interests must be weighed against the possible harm to the ecological, scientific, and recreational interests that are legitimately before this Court. Plaintiffs have submitted declarations asserting that they take whale watching trips, observe marine mammals underwater, conduct scien*26tifie research on marine mammals, and photograph these animals in their natural habitats. Plaintiffs contend that the Navy’s use of MFA sonar will injure marine mammals or alter their behavioral patterns, impairing plaintiffs’ ability to study and observe the animals.

While we do not question the seriousness of these interests, we conclude that the balance of equities and consideration of the overall public interest in this case tip strongly in favor of the Navy. For the plaintiffs, the most serious possible injury would be harm to an unknown number of the marine mammals that they study and observe. In contrast, forcing the Navy to deploy an inadequately trained antisubmarine force jeopardizes the safety of the fleet. Active sonar is the only reliable technology for detecting and tracking enemy diesel-electric submarines, and the President— the Commander in Chief — has determined that training with active sonar is “essential to national security.” Id., at 232a.

The public interest in conducting training exercises with active sonar under realistic conditions plainly outweighs the interests advanced by the plaintiffs. Of course, military interests do not always trump other considerations, and we have not held that they do. In this case, however, the proper determination of where the public interest lies does not strike us as a close question.

C

Despite the importance of assessing the balance of equities and the public interest in determining whether to grant a preliminary injunction, the District Court addressed these considerations in only a cursory fashion. The court’s entire discussion of these factors consisted of one (albeit lengthy) sentence: “The Court is also satisfied that the balance of hardships tips in favor of granting an injunction, as the harm to the environment, Plaintiffs, and public interest outweighs the harm that Defendants would incur if prevented from using MFA sonar, absent the use of effective mitigation *27measures, during a subset of their regular activities in one part of one state for a limited period.” Id., at 217a-218a. As the prior Ninth Circuit panel in this case put it, in staying the District Court’s original preliminary injunction, “[t]he district court did not give serious consideration to the public interest factor.” 502 F. 3d, at 863. The District Court’s order on remand did nothing to cure this defect, but simply repeated nearly verbatim the same sentence from its previous order. Compare 530 F. Supp. 2d, at 1118, with Pet. App. 217a-218a. The subsequent Ninth Circuit panel framed its opinion as reviewing the District Court’s exercise of discretion, 518 F. 3d, at 697-699, but that discretion was barely exercised here.

The Court of Appeals held that the balance of equities and the public interest favored the plaintiffs, largely based on its view that the preliminary injunction would not in fact impose a significant burden on. the Navy’s ability to conduct its training exercises and certify its strike groups. Id., at 698-699. The court deemed the Navy’s concerns about the preliminary injunction “speculative” because the Navy had not operated under similar procedures before. Ibid. But this is almost always the case when a plaintiff seeks injunctive relief to alter a defendant’s conduct. The lower courts failed properly to defer to .senior Navy officers’ specific, predictive judgments about how the preliminary injunction would reduce the effectiveness of the Navy’s SOCAL training exercises. See Wright & Miller §2948.2, at 167-168 (“The policy against the imposition of judicial restraints prior to an adjudication of the merits becomes more significant when there is reason to believe that the decree will be burdensome”).

The preliminary injunction requires the Navy to shut down its MFA sonar if a marine mammal is detected within 2,200 yards of a sonar-emitting vessel. The Ninth Circuit stated that the 2,200-yard shutdown zone would not be overly burdensome because sightings of marine mammals *28during training exercises are relatively rare. But regardless of the frequency of marine mammal sightings, the injunction will greatly increase the size of the shutdown zone. Pursuant to its exemption from the MMPA, the Navy agreed to reduce the power of its MFA sonar at 1,000 yards and 500 yards, and to completely turn off the system at 200 yards. Pet. App. 222a-230a. The District Court’s injunction does not include a graduated power-down, instead requiring a total shutdown of MFA sonar if a marine mammal is detected within 2,200 yards of a sonar-emitting vessel. There is an exponential relationship between radius length and surface area (Area = % r2). Increasing the radius of the shutdown zone from 200 to 2,200 yards would accordingly expand the surface area of the shutdown zone by a factor of over 100 (from 125,664 square yards to 15,205,308 square yards).

The lower courts did not give sufficient weight to the views of several top Navy officers, who emphasized that because training scenarios can take several days to develop, each additional shutdown can result in the loss of several days’ worth of training. Id., at 344a. Limiting the number of sonar shutdowns is particularly important during the Joint Tactical Force Exercises, which usually last for less than two weeks. Ibid. Rear Admiral John Bird explained that the 2,200-yard shutdown zone would cause operational commanders to “lose awareness of the tactical situation through the constant stopping and starting of MFA [sonar].” Id., at 332a; see also id., at 356a (“It may take days to get to the pivotal attack in antisubmarine warfare, but only minutes to confound the results upon which certification is based”). Even if there is a low likelihood of a marine mammal sighting, the preliminary injunction would clearly increase the number of disruptive sonar shutdowns the Navy is forced to perform during its SOCAL training exercises.

The Court of Appeals also concluded that the 2,200-yard shutdown zone would not be overly burdensome because the Navy had shut down MFA sonar 27 times during its eight *29prior training exercises in SOCAL; in several of these cases, the Navy turned off its sonar when marine mammals were spotted well beyond the Navy’s self-imposed 200-yard shutdown zone. 518 F. 3d, at 700, n. 65. Vice Admiral Samuel Locklear III — the Commander of the Navy’s Third Fleet— stated that any shutdowns beyond the 200-yard zone were voluntary avoidance measures that likely took place at tactically insignificant times; the Ninth Circuit discounted this explanation as not supported by the record. Ibid. In reaching this conclusion, the Court of Appeals ignored key portions of Vice Admiral Locklear’s declaration, in which he stated unequivocally that commanding officers “would not shut down sonar until legally required to do so if in contact with a submarine.” Pet. App. 354a-355a. Similarly, if a commanding officer is in contact with a target submarine, “the CO will be expected to continue to use active sonar unless another ship or helicopter can gain contact or if regulatory reasons dictate otherwise.” Id., at 355a. The record supports the Navy’s contention that its shutdowns of MFA sonar during prior training exercises only occurred during tactically insignificant times; those voluntary shutdowns do not justify the District Court’s imposition of a mandatory 2,200-yard shutdown zone.

Lastly, the Ninth Circuit stated that a 2,200-yard shutdown zone was feasible because the Navy had previously adopted a 2,000-meter zone for low-frequency active (LFA) sonar. The Court of Appeals failed to give sufficient weight to the fact that LFA sonar is used for long-range detection of enemy submarines, and thus its use and shutdown involve tactical considerations quite different from those associated with MFA sonar. See App. 508 (noting that equating MFA sonar with LFA sonar “is completely misleading and is like comparing 20 degrees Fahrenheit to 20 degrees Celsius”).

The Court of Appeals also concluded that the Navy’s training exercises would not be significantly affected by the requirement that it power down MFA sonar by 6 dB during *30significant surface ducting conditions. Again, we think the Ninth Circuit understated the burden this requirement would impose on the Navy’s ability to conduct realistic training exercises.

Surface ducting is a phenomenon in which relatively little sound energy penetrates beyond a narrow layer near the surface of the water. When surface ducting occurs, active sonar becomes more useful near the surface but less useful at greater depths. Pet. App. 299a-300a. Diesel-electric submariners are trained to take advantage of these distortions to avoid being detected by sonar. Id., at 333a.

The Ninth Circuit determined that the power-down requirement during surface ducting conditions was unlikely to affect certification of the Navy’s strike groups because surface ducting occurs relatively rarely, and the Navy has previously certified strike groups that did not train under such conditions. 518 F. 3d, at 701-702. This reasoning is backwards. Given that surface ducting is both rare and unpredictable, it is especially important for the Navy to be able to train under these conditions when they occur. Rear Admiral Bird explained that the 6 dB power-down requirement makes the training less valuable because it “exposes [sonar operators] to unrealistically lower levels of mutual interference caused by multiple sonar systems operating together by the ships within the Strike Group.” Pet. App. 281a (footnote and some capitalization omitted). Although a 6 dB reduction may not seem terribly significant, decibels are measured on a logarithmic scale, so a 6 dB decrease in power equates to a 75% reduction. Id., at 284a-285a.

The District Court acknowledged that “ ‘the imposition of these mitigation measures will require the Navy to alter and adapt the way it conducts antisubmarine warfare training — a substantial challenge. Nevertheless, evidence presented to the Court reflects that the Navy has employed mitigation measures in the past, without sacrificing training *31objectives.’ ” 527 F. Supp. 2d, at 1238. Apparently no good deed goes unpunished. The fact that the Navy has taken measures in the past to address concerns about marine mammals — or, for that matter, has elected not to challenge four additional restrictions imposed by the District Court in this case, see supra, at 17-19 — hardly means that other, more intrusive restrictions pose no threat to preparedness for war.

The Court of Appeals concluded its opinion by stating that “the Navy may return to the district court to request relief on an emergency basis” if the preliminary injunction “actually result[s] in an inability to train and certify sufficient naval forces to provide for the national defense.” 518 F. 3d, at 703. This is cold comfort to the Navy. The Navy contends that the injunction will hinder efforts to train sonar operators under realistic conditions, ultimately leaving strike groups more vulnerable to enemy submarines. Unlike the Ninth Circuit, we do not think the Navy is required to wait until the injunction “actually result[s] in an inability to train ... sufficient naval forces to provide for the national defense” before seeking its dissolution. By then it may be too late.

IV

As noted above, we do not address the underlying merits of plaintiffs’ claims. While we have authority to proceed to such a decision at this point, see Munaf, 553 U. S., at 691-692, doing so is not necessary here. In addition, reaching the merits is complicated by the fact that the lower courts addressed only one of several issues raised, and plaintiffs have largely chosen not to defend the decision below on that ground.5

*32At the same time, what we have said makes clear that it would be an abuse of discretion to enter a permanent injunction, after final decision on the merits, along the same lines as the preliminary injunction. An injunction is a matter of equitable discretion; it does not follow from success on the merits as a matter of course. Romero-Barcelo, 456 U. S., at 313 (“[A] federal judge sitting as chancellor is not mechanically obligated to grant an injunction for every violation of law”).

The factors examined above — the balance of equities and consideration of the public interest — are pertinent in assessing the propriety of any injunctive relief, preliminary or permanent. See Amoco Production Co., 480 U. S., at 546, n. 12 (“The standard for a preliminary injunction is essentially the same as for a permanent injunction with the exception that the plaintiff must show a likelihood of success on the merits rather than actual success”). Given that the ultimate legal claim is that the Navy must prepare an EIS, not that it must cease sonar training, there is no basis for enjoining such *33training in a manner credibly alleged to pose a serious threat to national security. This is particularly true in light of the fact that the training has been going on for 40 years with no documented episode of harm to a marine mammal. A court concluding that the Navy is required to prepare an EIS has many remedial tools at its disposal, including declaratory relief or an injunction tailored to the preparation of an EIS rather than the Navy’s training in the interim. See, e. g., Steffel v. Thompson, 415 U. S. 452, 466 (1974) (“Congress plainly intended declaratory relief to act as an alternative to the strong medicine of the injunction”). In the meantime, we see no basis for jeopardizing national security, as the present injunction does. Plaintiffs confirmed at oral argument that the preliminary injunction was “the whole ball game,” Tr. of Oral Arg. 33, and our analysis of the propriety of preliminary relief is applicable to any permanent injunction as well.

* Hi

President Theodore Roosevelt explained that “the only way in which a navy can ever be made efficient is by practice at sea, under all the conditions which would have to be met if war existed.” President’s Annual Message, 42 Cong. Rec. 81 (1907). We do not discount the importance of plaintiffs’ ecological, scientific, and recreational interests in marine mammals. Those interests, however, are plainly outweighed by the Navy’s need to conduct realistic training exercises to ensure that it is able to neutralize the threat posed by enemy submarines. The District Court abused its discretion by imposing a 2,200-yard shutdown zone and by requiring the Navy to power down its MPA sonar during significant surface ducting conditions. The judgment of the Court of Appeals is reversed, and the preliminary injunction is vacated to the extent it has been challenged by the Navy.

It is so ordered.

1

In contrast, passive sonar “listens” for sound waves but does not introduce sound into the water. Passive sonar is not effective for tracking diesel-electric submarines because those vessels can operate almost silently. Passive sonar also has a more limited range than active sonar, and cannot identify the exact location of an enemy submarine. Pet. App. 266a-271a.

2

The CZMA states that federal agencies taking actions “that affec[t] any land or water use or natural resource of the coastal zone” shall carry out these activities “in a manner which is consistent to the maximum extent practicable with the enforceable policies of approved State management programs.” 16 U. S. C. § 1456(c)(1)(A).

3

That provision states in full: ‘Where emergency circumstances make it necessary to take an action with significant environmental impact without observing the provisions of these regulations, the Federal agency taking the action should consult with the Council about alternative arrangements. Agencies and the Council will limit such arrangements to actions necessary to control the immediate impacts of the emergency. Other actions remain subject to NEPA review.”

4

The Ninth Circuit’s discussion of the plaintiffs’ likelihood of success was limited to their NEPA claims. The court did not discuss claims under the CZMA or ESA.

5

The bulk of Justice Ginsburg’s dissent is devoted to the merits. For the reasons stated, we find the injunctive relief granted in this case an abuse of discretion, even if plaintiffs are correct on the underlying merits. As to the injunction, the dissent barely mentions the Navy’s in*32terests. Post, at 53. We find that those interests, and the documented risks to national security, clearly outweigh the harm on the other side of the balance.

We agree with much of Justice Breyer's analysis, post, at 36-41 (opinion concurring in part and dissenting in part), but disagree with his conclusion that the modified conditions imposed by the stay order should remain in force until the Navy completes its EIS, post, at 42-43. The Court is reviewing the District Court’s imposition of the preliminary injunction; once we conclude, as Justice Breyer does, post, at 41, that the preliminary injunction should be vacated, the stay order is no longer pertinent. A stay is a useful tool for managing the impact of injunctive relief pending further appeal, but once the Court resolves the merits of the appeal, the stay ceases to be relevant. See 518 F. 3d 704, 706 (CA9 2008) (“[T]he partial stay... shall remain in effect until final disposition by the Supreme Court”). Unexamined conditions imposed by the stay order are certainly no basis for what would be in effect the entry of a new preliminary injunction by this Court.

*34Justice Breyer,

with whom Justice Stevens joins as to Part I, concurring in part and dissenting in part.

As of December 2006, the United States Navy planned to engage in a series of 14 antisubmarine warfare training exercises off the southern California coast. The Natural Resources Defense Council, Inc., and others (NRDC) brought this case in Federal District Court claiming that the National Environmental Policy Act of 1969 (NEPA) requires the Navy to prepare an environmental impact statement (EIS) (assessing the impact of the exercises on marine mammals) prior to its engaging in the exercise's. As the case reaches us, the District Court has found that the NRDC will likely prevail on its demand for an EIS; the Navy has agreed to prepare an EIS; the District Court has forbidden the Navy to proceed with the exercises unless it adopts six mitigating measures; and the Navy has agreed to adopt all but two of those measures.

The controversy between the parties now concerns the two measures that the Navy is unwilling to adopt. The first concerns the “shutdown zone,” a circle with a ship at the center within which the Navy must try to spot marine mammals and shut down its sonar if one is found. The controverted condition would enlarge the radius of that circle from about one-tenth of a mile (200 yards) to one and one-quarter miles (2,200 yards). The second, concerns special ocean conditions called “surface ducting conditions.” The controverted condition would require the Navy, when it encounters any such condition, to diminish the sonar’s power by 75%. The Court of Appeals affirmed the District Court order that contained these two conditions. 518 F. 3d 658, 703 (CA9 2008).

I

We must now decide whether the District Court was legally correct in forbidding the training exercises unless the Navy implemented the two controverted conditions. In *35doing so, I assume, like the Court, that the NRDC will prevail on its demand for an EIS. (Indeed, the Navy is in the process of preparing one.) And, I would ask whether, in imposing these conditions, the District Court properly “balance[d the] harms.” See, e.g., Amoco Production Co. v. Gambell, 480 U. S. 531, 545 (1987).

Respondents’ (the plaintiffs) argument favoring the District Court injunction is a strong one. As Justice Ginsburg well points out, see post, at 47-48 (dissenting opinion), the very point of NEPA’s insistence upon the writing of an EIS is to force an agency “carefully” to “consider ... detailed information concerning significant environmental impacts,” while “giv[ing] the public the assurance that the agency ‘has indeed considered- environmental concerns in its decision-making process.’” Robertson v. Methow Valley Citizens Council, 490 U. S. 332, 349 (1989). NEPA seeks to assure that when Government officials consider taking action that may affect the environment, they do so fully aware of the relevant environmental considerations. An EIS does not force them to make any particular decision, but it does lead them to take environmental considerations into account when they decide whether, or how, to act. Id., at 354. Thus, when a decision to which EIS obligations attach is made without the informed environmental consideration that NEPA requires, much of the harm that NEPA seeks to prevent has already taken place. In this case, for example, the absence of an injunction means that the Navy will proceed with its exercises in the absence of the fuller consideration of environmental effects that an EIS is intended to bring. The absence of an injunction thereby threatens to cause the very environmental harm that a full preaction EIS might have led the Navy to avoid (say, by adopting the two additional mitigation measures that the NRDC proposes). Consequently, if the exercises are to continue, conditions designed to mitigate interim environmental harm may well be appropriate.

*36On the other hand, several features of this case lead me to conclude that the record, as now before us, lacks adequate support for an injunction imposing the two controverted requirements. First, the evidence of need for the two special conditions is weak or uncertain. The record does show that the exercises as the Navy originally proposed them could harm marine mammals. The District Court found (based on the Navy’s study of the matter) that the exercises might cause 466 instances of Level A harm and 170,000 instances of Level B harm. App. to Pet. for Cert. 196a-197a. (The environmental assessment actually predicted 564 instances of Level A harm. See App. 223-224.) The study defines Level A injury as “any act that injures or has the significant potential to injure a marine mammal or marine mammal stock in the wild” through “destruction or loss of biological tissue,” whether “slight to severe.” Id., at 160. It defines Level B harm as “ ‘any act that disturbs or is likely to disturb a marine mammal... by causing disruption of natural behavioral patterns including, but not limited to, migration, surfacing, nursing, breeding, feeding, or sheltering to a point where such behaviors are abandoned or significantly altered’ ” and describes it as a “short term” and “temporary” “disturbance.” Id., at 161, 175.

The raw numbers seem large. But the parties argue about the extent to which they mean likely harm. The Navy says the classifications and estimates err on the side of caution. (When in doubt about the amount of harm to a mammal, the study assumed the harm would qualify as Level A harassment. Id., at 200.) The Navy also points out that, by definition, mammals recover from Level B injuries, often very quickly. It notes that, despite 40 years of naval exercises off the southern California coast, no injured marine mammal has ever been found. App. to Pet. for Cert. 274a-275a. (It adds that dolphins often swim alongside the ships. Id., at 290a, 346a.) At the same time, plaintiffs point to instances where whales have been found stranded. They add *37that scientific studies have found a connection between those beachings and the Navy’s use of sonar, see, e. g., App. 600-602, and the Navy has acknowledged one stranding where “U. S. Navy mid-frequency sonar has been identified as the most plausible contributory source to the stranding event,” id., at 168.

Given the uncertainty the figures create in respect to the harm caused by the Navy’s original training plans, it would seem important to have before us at least some estimate of the harm likely avoided by the Navy’s decision not to contest here four of the six mitigating conditions that the District Court ordered. Without such evidence, it is difficult to assess the relevant harm — that is, the environmental harm likely caused by the Navy’s exercises with the four uncontested mitigation-measures (but without the two contested mitigation measures) in place.

Second, the Navy has filed multiple affidavits from Navy officials explaining in detail the seriousness of the harm that the delay associated with completion of this EIS (approximately one year) would create in respect to the Navy’s ability to maintain an adequate national defense. See generally App. to Pet. for Cert. 260a-357a. Taken by themselves, those affidavits make a strong case for the proposition that insistence upon the two additional mitigating conditions would seriously interfere with necessary defense training.

The affidavits explain the importance of training in antisubmarine warfare, id., at 263a; the need to use active sonar to detect enemy submarines, id., at 266a-267a, App. 566; the complexity of a training exercise involving sonar, App. to Pet. for Cert. 343a; the need for realistic conditions when training exercises take place, id., at 299a-300a, App. 566; the “cascading” negative “effect” that delay in one important aspect of a set of coordinated training exercises has upon the Navy’s ability “to provide combat ready forces,” App. to Pet. for Cert. 343a; the cost and disruption that would accompany the adoption of the two additional mitigating conditions that *38the NRDC seeks, ibid.; the Navy’s resulting inability adequately to train personnel, id., at 278a; the effectiveness of the mammal-protecting measures that the Navy has taken in the past, id., at 285a-298a; and the reasonable likelihood that the mitigating conditions to which it has agreed will prove adequate, id., at 296a.

Third, and particularly important in my view, the District Court did not explain why it rejected the Navy’s affidavit-supported contentions. In its first opinion enjoining the use of sonar, the District Court simply stated:

“The Court is ... satisfied that the balance of hardships tips in favor of granting an injunction, as the harm to the environment, Plaintiffs, and public interest outweighs the harm that Defendants would incur if prevented from using [mid-frequency active (MFA)] sonar, absent the use of effective mitigation measures, during a subset of their regular activities in one part of one state for a limited period.” Id., at 217a-218a.

Following remand from the Court of Appeals, the District Court simply repeated, word for word, this same statement. It said:

“The Court is ... satisfied that the balance of hardships tips in favor of granting an injunction, as the harm to the environment, Plaintiffs, and public interest outweighs the harm that Defendants would incur (or the public interest would suffer) if Defendants were prevented from using MFA sonar, absent the use of effective mitigation measures, during a subset of their regular activities in one part of one state for a limited period.” 530 F. Supp. 2d 1110, 1118 (CD Cal. 2008).

With respect to the imposition of the 2,200-yard shutdown zone, the District Court noted evidence of the harm that MFA sonar poses to marine mammals, and then concluded that “[t]he Court therefore is persuaded that while the 2200 yard shutdown requirement may protect marine mammals *39from the harshest of sonar-related consequences, it represents a minimal imposition [on] the Navy’s training exercises.” Id., at 1119. The District Court did not there explain the basis for that conclusion. With respect to the imposition of the surface ducting condition, the District Court said nothing about the Navy’s interests at all. Id., at 1120-1121.

While a district court is often free simply to state its conclusion in summary fashion, in this instance neither that conclusion, nor anything else I have found in the District Court’s opinion, answers the Navy’s documented claims that the two extra conditions the District Court imposed will, in effect, seriously interfere with its ability to carry out necessary training exercises.

The first condition requires the Navy to reduce the power of its sonar equipment by 75% when the ship encounters a condition called “surface ducting” that occurs when the presence of layers of water of different temperature make it unusually difficult for sonar operators to determine whether a diesel submarine is hiding below. Rear Admiral John Bird, an expert in submarine warfare, made clear that the 75% power-reduction requirement was equivalent to forbidding any related training. App. to Pet. for Cert. 297a. But he says in paragraph 52 of his declaration: “Training in surface ducting conditions is critical to effective training because sonar operators need to learn how sonar transmissions are altered due to surface ducting and how submarines may take advantage of them.” Id., at 299a-300a. The District Court, as far as I can tell, did not even acknowledge in its opinion the Navy’s asserted interest in being able to train under these conditions. 530 F. Supp. 2d, at 1120-1121.

The second condition requires the Navy to expand the sonar “shutdown” area surrounding a ship (i e., turn off the sonar if a mammal is spotted in the area) from a circle with a radius of about one-tenth of a mile to a circle with a radius of about one mile and a quarter. Both sides agree that this *40requirement will lead to more shutdowns. Admiral Gary Roughead, Chief of Naval Operations, states in paragraph 12 of his declaration that this expanded zone requirement “will result in increased interruptions to training exercises, . . . vastly increasing] the risk of negating training effectiveness, preventing strike group certification, and disrupting carefully orchestrated deployment plans to meet world-wide operational commitments.” App. to Pet. for Cert. 344a. Again, I can find nothing in the District Court’s opinion that specifically explains why this is not so. 530 F. Supp. 2d, at 1119-1120.

Fourth, the Court of Appeals sought, through its own thorough examination of the record, to supply the missing explanations. But those explanations are not sufficient. In respect to the surface ducting conditions, the Court of Appeals rejected the Navy’s contentions on the ground that those conditions are “rar[e],” and the Navy has certified trainings that did not involve any encounter with those conditions. 518 F. 3d, at 701-702. I am not certain, however, why the rarity of the condition supports the District Court’s conclusion. Rarity argues as strongly for training when the condition is encountered as it argues for the contrary.

In respect to the expansion of the “shutdown” area, the Court of Appeals noted that (1) the Navy in earlier exercises had shut down its sonar when marine mammals were sited within about one-half a mile, (2) the Navy has used a larger shutdown area when engaged in exercises with lower frequency sonar equipment, and (3) foreign navies have used larger shutdown areas. Id., at 699-701, and nn. 63, 67. But the Navy’s affidavits state that (1) earlier shutdowns when marine mammals were spotted at farther distances “likely occurred during tactically insignificant times,” App. to Pet. for Cert. 356a, (2) ships with low frequency sonar (unlike the sonar here at issue) have equipment that makes it easier to monitor the larger area, particularly by significantly reducing the number of monitoring personnel necessarily involved, and (3) foreign navy experience is not relevant given the *41potentially different military demands upon those navies, App. 508-509.

Finally, the Court of Appeals, mirroring a similar District Court suggestion in the language I have quoted, says that “the exercises in southern California are only a subset of the Navy’s training activities involving active sonar.” 518 F. 3d, at 702. It adds that the Navy’s study “shows the Navy is still able to conduct its exercises in alternative locations, in reduced number, or through simulation.” Ibid., n. 69. The Court of Appeals, however, also concluded that the study “provides reasonably detailed justifications for why the Southern California Operating Area is uniquely suited to these exercises, and demonstrates that the Navy would suffer a certain hardship if the considered alternatives were employed instead.” Ibid.

Fifth, when the Court of Appeals first heard this case following the District Court’s imposition of a broad, absolute injunction, it held that any injunction must be crafted so that the Navy could continue its training exercises. Noting that the Navy had, in the past, been able to use mitigation measures to “reduce the harmful effects of its active sonar,” it “vacate[d] the stay and remand[ed] this matter to the district court to narrow its injunction so as to provide mitigation conditions under which the Navy may conduct its training exercises.” 508 F. 3d 885, 887 (CA9 2007) (emphasis added). For the reasons just stated, neither the District Court nor the Court of Appeals has explained why we should reject the Navy’s assertions that it cannot effectively conduct its training exercises under the mitigation conditions imposed by the District Court.

I would thus vacate the preliminary injunction imposed by the District Court to the extent it has been challenged by the Navy. Neither the District Court nor the Court of Appeals has adequately explained its conclusion that the balance of the equities tips in favor of plaintiffs. Nor do those parts of the record to which the parties have pointed supply the missing explanation.

*42II

Nonetheless, as the Court of Appeals held when it first considered this case, the Navy’s past use of mitigation conditions makes clear that the Navy can effectively train under some mitigation conditions. In the ordinary course, I would remand so the District Court could, pursuant to the Court of Appeals’ direction, set forth mitigation conditions that will protect the marine wildlife while also enabling the Navy to carry out its exercises. But, at this point, the Navy has informed us that this set of exercises will be complete by January, at the latest, and an EIS will likely be complete at that point, as well. Thus, by the time the District Court would have an opportunity to impose new conditions, the case could very well be moot.

In February of this year, the Court of Appeals stayed the injunction imposed by the District Court — but only pending this Court’s resolution of the case. The Court of Appeals concluded that “[i]n light of the short time before the Navy is to commence its next exercise, the importance of the Navy’s mission to provide for the national defense and the representation by the Chief of Naval Operations that the district court’s preliminary injunction in its current form will ‘unacceptably risk’ effective training and strike group certification and thereby interfere with his statutory responsibility ... to ‘organiz[e], train[], and equip[] the Navy,’” interim relief was appropriate, and the court then modified the two mitigation conditions at issue. 518 F. 3d 704, 705 (CA9 2008).

With respect to the 2,200-yard shutdown zone, it required the Navy to suspend its use of the sonar if a marine mammal is detected within 2,200 yards, except when sonar is being used at a “critical point in the exercise,” in which case the amount by which the Navy must power down is proportional to the mammal’s proximity to the sonar. Id., at 705-706 (internal quotation marks omitted). With respect to surface ducting, the Navy is only required to shut down sonar alto*43gether when a marine mammal is detected within 500 meters and the amount by which it is otherwise required to power down is again proportional to the mammal’s proximity to the sonar source. Ibid. The court believed these conditions would permit the Navy to go forward with its imminently planned exercises while at the same time minimizing the harm to marine wildlife.

In my view, the modified conditions imposed by the Court of Appeals in its February stay order reflect the best equitable conditions that can be created in the short time available before the exercises are complete and the EIS is ready. The Navy has been training under these conditions since February, so allowing them to remain in place will, in effect, maintain what has become the status quo. Therefore, I would modify the Court of Appeals’ February 29, 2008, order so that the provisional conditions it contains remain in place until the Navy’s completion of an acceptable EIS.

Justice Ginsburg,

with whom Justice Souter joins, dissenting.

The central question in this action under the National Environmental Policy Act of 1969 (NEPA) was whether the Navy must prepare an environmental impact statement (EIS). The Navy does not challenge its obligation to do so, and it represents that the EIS will be complete in January 2009 — one month after the instant exercises conclude. If the Navy had completed the EIS before taking action, as NEPA instructs, the parties and the public could have benefited from the environmental analysis — and the Navy’s training could have proceeded without interruption. Instead, the Navy acted first, and thus thwarted the very purpose an EIS is intended to serve. To justify its course, the Navy sought dispensation not from Congress, but from an executive council that lacks authority to countermand or revise NEPA’s requirements. I would hold that, in imposing manageable measures to mitigate harm until completion of the *44EIS, the District Court conscientiously balanced the equities and did not abuse its discretion.

I

In December 2006, the Navy announced its intent to prepare an EIS to address the potential environmental effects of its naval readiness activities in the Southern California (SOCAL) Range Complex. See 71 Fed. Reg. 76639 (2006). These readiness activities include expansion and intensification of naval training, as well as research, development, and testing of various systems and weapons. Id., at 76639, 76640. The EIS process is underway, and the Navy represents that it will be complete in January 2009. Brief for Petitioners 11; Tr. of Oral Arg. 11.

In February 2007, seeking to commence training before completion of the EIS, the Navy prepared an environmental assessment (EA) for the 14 exercises it planned to undertake in the interim. See App. to Pet. for Cert. 235a.1 On February 12, the Navy concluded the EA with a finding of no significant impact. App. 225-226. The same day, the Navy commenced its training exercises. Id., at 227 (“The Proposed Action is hereby implemented.”).

On March 22, 2007, the Natural Resources Defense Council, Inc. (NRDC), filed suit in the U. S. District Court for the Central District of California, seeking declaratory and injunctive relief based on the Navy’s alleged violations of NEPA and other environmental statutes. As relevant here, the District Court determined that NRDC was likely to succeed on its NEPA claim and that equitable principles warranted preliminary relief. On August 7, 2007, the court *45enjoined the Navy’s use of mid-frequency active (MFA) sonar during the 11 remaining exercises at issue.

On August 31, the Court of Appeals for the Ninth Circuit stayed the injunction pending disposition of the Navy’s appeal, and the Navy proceeded with two more exercises. In a November 13 order, the Court of Appeals vacated the stay, stating that NRDC had shown “a strong likelihood of success on the merits” and that preliminary injunctive relief was appropriate. 508 F. 3d 885, 886 (2007). The Court of Appeals remanded, however, instructing the District Court to provide mitigation measures under which the Navy could conduct its remaining exercises.

On remand, the District Court received briefing from both parties. In addition, the court “toured the USS Milius at the naval base in San Diego, California, to improve its understanding of the Navy’s sonar training procedures and the feasibility of the parties’ proposed mitigation measures. Counsel for both [parties] were present.” 530 F. Supp. 2d 1110, 1112 (2008). On January 3, 2008, the District Court entered a modified preliminary injunction imposing six mitigation measures. The court revised the modified injunction slightly on January 10 in response to filings by the Navy, and four days later, denied the Navy’s application for a stay pending appeal.

On the following day, January 15, the Council on Environmental Quality (CEQ), an advisory body within the Executive Office of the President, responded to the Navy’s request for “alternative arrangements” for NEPA compliance. App. to Pet. for Cert. 233a. The “arrangements” CEQ set out purported to permit the Navy to continue its training without timely environmental review. Id., at 241a-247a. The Navy accepted the arrangements on the same day. App. 228.

The Navy then filed an emergency motion in the Court of Appeals requesting immediate vacatur of the District Court’s modified injunction. CEQ’s action, the Navy urged, *46eliminated the injunction’s legal foundation. In the alternative, the Navy sought a stay of two aspects of the injunction pending its appeal: the 2,200-yard mandatory shutdown zone and the power-down requirement in significant surface ducting conditions, see ante, at 17-18 (opinion of the Court). While targeting in its stay application only two of the six measures imposed by the District Court, the Navy explicitly reserved the right to challenge on appeal each of the six mitigation measures. Responding to the Navy’s emergency motion, the Court of Appeals remanded the matter to allow the District Court to determine in the first instance the effect of the intervening executive action. Pending its own consideration of the Navy’s motion, the District Court stayed the injunction, and the Navy conducted its sixth exercise.

On February 4, after briefing and oral argument, the District Court denied the Navy’s motion. The Navy appealed, reiterating its position that CEQ’s action eliminated all justification for the injunction. The Navy also argued that vacatur of the entire injunction was required irrespective of CEQ’s action, in part because the “conditions imposed, in particular the 2,200 yard mandatory shutdown zone and the six decibel (75%) power-down in significant surface ducting conditions, severely degrade the Navy’s training.” Brief for Appellants in No. 08-55054 (CA9), p. 15. In the February 29 decision now under review, the Court of Appeals affirmed the District Court’s judgment. 518 F. 3d 658, 703 (2008). The Navy has continued training in the meantime and plans to complete its final exercise in December 2008.

As the procedural history indicates, the courts below determined that an EIS was required for the 14 exercises. The Navy does not challenge that decision in this Court. Instead, the Navy defends its failure to complete an EIS before launching the exercises based upon CEQ’s “alternative arrangements” — arrangements the Navy sought and obtained in order to overcome the lower courts’ rulings. As *47explained below, the Navy’s actions undermined NEPA and took an extraordinary course.

II

NEPA “promotes its sweeping commitment” to environmental integrity “by focusing Government and public attention on the environmental effects of proposed agency action.” Marsh v. Oregon Natural Resources Council, 490 U. S. 360, 371 (1989). “By so focusing agency attention, NEPA ensures that the agency will not act on incomplete information, only to regret its decision after it is too late to correct.” Ibid.

The EIS is NEPA’s core requirement. Department of Transportation v. Public Citizen, 541 U. S. 752, 757 (2004). This Court has characterized the requirement as “action-forcing.” Andrus v. Sierra Club, 442 U. S. 347, 350 (1979) (internal quotation marks omitted). Environmental concerns must be “integrated into the very process of agency decisionmaking” and “interwoven into the fabric of agency planning.” Id., at 350-351. In addition to discussing potential consequences, an EIS must describe potential mitigation measures and alternatives to the proposed course of action. See Robertson v. Methow Valley Citizens Council, 490 U. S. 332, 351-352 (1989) (citing 40 CFR §§ 1508.25(b), 1502.14(f), 1502.16(h), 1505.2(c) (1987)). The EIS requirement “ensures that important effects will not be overlooked or underestimated only to be discovered after resources have been committed or the die otherwise cast.” 490 U. S., at 349.

“Publication of an EIS . . . also serves a larger informational role.” Ibid. It demonstrates that an agency has indeed considered environmental concerns, and “perhaps more significantly, provides a springboard for public comment.” Ibid. At the same time, it affords other affected governmental bodies “notice of the expected consequences and the opportunity to plan and implement corrective measures in a timely manner.” Id., at 350.

*48In light of these objectives, the timing of an EIS is critical. CEQ regulations instruct agencies to “integrate the NEPA process with other planning at the earliest possible time to insure that planning and decisions reflect environmental values.” 40 CFR § 1501.2 (1987). An EIS must be prepared “early enough so that it can serve practically as an important contribution to the decisionmaking process and will not be used to rationalize or justify decisions already made.” Andrus, 442 U. S., at 351-352, n. 3 (quoting 43 Fed. Reg. 55995 (1978) (codified in 40 CFR § 1502.5 (1979))).

The Navy’s publication of its EIS in this case, scheduled to occur after the 14 exercises are completed, defeats NEPA’s informational and participatory purposes. The Navy’s inverted timing, it bears emphasis, is the very reason why the District Court had to confront the question of mitigation measures at all. Had the Navy prepared a legally sufficient EIS before beginning the SOCAL exercises, NEPA would have functioned as its drafters intended: The EIS process and associated public input might have convinced the Navy voluntarily to adopt mitigation measures, but NEPA itself would not have impeded the Navy’s exercises. See Public Citizen, 541 U. S., at 756, 769, n. 2 (noting that NEPA does not mandate particular results, but rather establishes procedural requirements with a “focus on improving agency decisionmaking”).

The Navy had other options. Most importantly, it could have requested assistance from Congress. The Government has sometimes obtained congressional authorization to proceed with planned activities without fulfilling NEPA’s requirements. See, e.g., Floyd D. Spence National Defense Authorization Act for Fiscal Year 2001, Pub. L. 106-398, §317, 114 Stat. 1654A-57 (exempting the military from preparing a programmatic EIS for low-level flight training); 42 U. S. C. § 10141(c) (2000 ed.) (exempting the Environmental Protection Agency from preparing an EIS for the development of criteria for handling spent nuclear fuel and high-*49level radioactive waste); 43 U. S. C. § 1652(d) (exempting construction of the trans-Alaska oil pipeline from further NEPA compliance).

Rather than resorting to Congress, the Navy “sought relief from the Executive Branch.” Ante, at 18 (opinion of the Court). On January 10, 2008, the Navy asked CEQ, adviser to the President, to approve alternative arrangements for NEPA compliance pursuant to 40 CFR §1506.11 (1987). App. to Pet. for Cert. 233a; see ante, at 18, n. 3. The next day, the Navy submitted supplementary material to CEQ, including the Navy’s EA and after-action reports, the District Court’s orders, and two analyses by the National Marine Fisheries Service (NMFS). App. to Pet. for Cert. 237a-238a. Neither the Navy nor CEQ notified NRDC, and CEQ did not request or consider any of the materials underlying the District Court orders it addressed.

Four days later, on January 15, the Chairman of CEQ issued a letter to the Secretary of the Navy. Repeating the Navy’s submissions with little independent analysis, the letter stated that the District Court’s orders posed risks to the Navy’s training exercises. See id., at 238a (“You have explained that the training restrictions set forth in the . . . injunctive orders prevent the Navy from providing Strike Groups with adequate proficiency training and create a substantial risk of precluding certification of the Strike Groups as combat ready.”).

The letter continued:

“Discussions between our staffs, your letter and supporting documents, and the classified declaration and briefings I have received, have clearly determined that the Navy cannot ensure the necessary training to certify strike groups for deployment under the terms of the injunctive orders. Based on the record supporting your request . . . CEQ has concluded that the Navy must be able to conduct the [exercises] ... in a timeframe that does not provide sufficient .time to complete an EIS. *50Therefore, emergency circumstances are present for the nine exercises and alternative arrangements for compliance with NEPA under CEQ regulation 40 C. F. R. § 1506.11 are warranted.” Id., at 240a.

The alternative arrangements CEQ set forth do not vindicate NEPA’s objectives. The arrangements provide for “public participation measures,” which require the Navy to provide notices of the alternative arrangements. Id., at 241a, 242a. The notices must “seek input on the process for reviewing post-exercise assessments” and “include an offer to meet jointly with Navy representatives . . . and CEQ to discuss the alternative arrangements.” Id., at 242a-243a. The alternative arrangements also describe the Navy’s existing research and mitigation efforts. Id., at 243a-247a.

CEQ’s hasty decision on a one-sided record is no substitute for the District Court’s considered judgment based on a two-sided record.2 More fundamentally, even an exemplary CEQ review could not have effected the short circuit the Navy sought. CEQ lacks authority to absolve an agency of its statutory duty to prepare an EIS. NEPA established CEQ to assist and advise the President on environmental policy, 42 U. S. C. § 4342, and a 1977 Executive Order charged CEQ with issuing regulations to federal agencies for implementation of NEPA’s procedural provisions, Exec. Order No. 11991, 3 CFR 123 (1977 Comp.). This Court has recognized that CEQ’s regulations are entitled to “substantial deference,” Robertson, 490 U. S., at 355, and 40 CFR § 1506.11 indicates that CEQ may play an important consultative role in emergency circumstances, but we have never suggested that CEQ could eliminate the statute’s command. If the *51Navy sought to avoid its NEPA obligations, its remedy lay in the Legislative Branch. The Navy’s alternative course— rapid, self-serving resort to an office in the White House — is surely not what Congress had in mind when it instructed agencies to comply with NEPA “to the fullest extent possible.” 42U.S.C. §4332.3

Ill

A

Flexibility is a hallmark of equity jurisdiction. “The essence of equity jurisdiction has been the power of the Chancellor to do equity and to mould each decree to the necessities of the particular case. Flexibility rather than rigidity has distinguished it.” Weinberger v. Romero-Barcelo, 456 U. S. 305, 312 (1982) (quoting Hecht Co. v. Bowles, 321 U. S. 321, 329 (1944)). Consistent with equity’s character, courts do not insist that litigants uniformly show a particular, predetermined quantum of probable success or injury before awarding equitable relief. Instead, courts have evaluated claims for equitable relief on a “sliding scale,” sometimes awarding relief based on a lower likelihood of harm when the likelihood of success is very high. 11A C. Wright, A. Miller, & M. Kane, Federal Practice and Procedure § 2948.3, p. 195 (2d ed. 1995). This Court has never rejected that formulation, and I do not believe it does so today.

Equity’s flexibility is important in the NEPA context. Because an EIS is the tool for uncovering environmental harm, environmental plaintiffs may often rely more heavily on their probability of success than the likelihood of harm. The Court is correct that relief is not warranted “simply to prevent the possibility of some remote future injury.” Ante, *52at 22 (quoting Wright & Miller, supra, §2948.1, at 155). “However, the injury need not have been inflicted when application is made or be certain to occur; a strong threat of irreparable injury before trial is an adequate basis.” Wright & Miller, supra, §2948.1, at 155-156 (footnote omitted). I agree with the District Court that NRDC made the required showing here.

B

The Navy’s own EA predicted substantial and irreparable harm to marine mammals. Sonar is linked to mass standings of marine mammals, hemorrhaging around the brain and ears, acute spongiotic changes in the central nervous system, and lesions in vital organs. E. g., App. 600-602; id., at 360-362, 478-479. As the Ninth Circuit noted, the EA predicts that the Navy’s “use of MFA sonar in the SOCAL exercises will result in 564 instances of physical injury including permanent hearing loss (Level A harassment) and nearly 170,000 behavioral disturbances (Level B harassment), more than 8,000 of which would also involve temporary hearing loss.” 518 F. 3d, at 696; see App. 223-224. Within those totals,

“the EA predicts 436 Level A harassments of Cuvier’s beaked whales. According to [the National Oceanic and Atmospheric Administration (NOAA)], as few as 1,121 . . . may exist in California, Oregon and Washington combined. Likewise, the EA predicts 1,092 Level B harassments of bottlenose dolphins, of which only 5,271 may exist in the California Coastal and Offshore stocks.” 518 F. 3d, at 691-692.

The majority acknowledges the lower courts’ findings, ante, at 19, but also states that the EA predicted “only eight Level A harassments of common dolphins each year” and “274 Level B harassments of beaked whales per year, none of which would result in permanent injury,” ante, at 16. Those numbers do not fully capture the EA’s predictions.

*53The EA classified the harassments of beaked whales as Level A, not Level B. The EA does indeed state that “modeling predicts non-injurious Level B exposures.” App. 185. But, as the majority correctly notes, ante, at 16, the EA also states that “all beaked whale exposures are counted as Level A,” App. 185. The EA counted the predicted exposures as Level A “[b]y Navy policy developed in conjunction with NMFS.” Id., at 200. The record reflects “the known sensitivity of these species to tactical sonar,” id., at 365 (NO A A letter), and as the majority acknowledges, beaked whales are difficult to study, ante, at 16. Further, as the Ninth Circuit noted, “the EA . . . maintained that the methodology used was based on the ‘best available science.’” 518 F. 3d, at 669.4

In my view, this likely harm — 170,000 behavioral disturbances, including 8,000 instances of temporary hearing loss; and 564 Level A harms, including 436 injuries to a beaked whale population numbering only 1,121 — cannot be lightly dismissed, even in the face of an alleged risk to the effectiveness of the Navy’s 14 training exercises. There is no doubt that the training exercises serve critical interests. But those interests do not authorize the Navy to violate a statutory command, especially when recourse to the Legislature remains open. “Of course, military interests do not always trump other considerations, and we have not held that they do.” Ante, at 26.

In light of the likely, substantial harm to the environment, NRDC’s almost inevitable success on the merits of its claim *54that NEPA required the Navy to prepare an EIS, the history of this litigation, and the public interest, I cannot agree that the mitigation measures the District Court imposed signal an abuse of discretion. Cf. Amoco Production Co. v. Gambell, 480 U. S. 531, 545 (1987) (“Environmental injury, by its nature, can seldom be adequately remedied by money damages and is often permanent or at least of long duration, i. e., irreparable. If such injury is sufficiently likely, therefore, the balance of harms will usually favor the issuance of an injunction to protect the environment.”).

For the reasons stated, I would affirm the judgment of the Ninth Circuit.

1

An EA is used “for determining whether to prepare” an EIS. Department of Transportation v. Public Citizen, 541 U. S. 752, 757 (2004) (quoting 40 CFR § 1508.9(a) (2003)); see ante, at 15-16 (opinion of the Court). By definition, an EA alone does not satisfy an agency’s obligation under NEPA if the effects of a proposed action require preparation of a full EIS.

2

The District Court may well have given too spare an explanation for the balance of hardships in issuing its injunction of August 7, 2007. The court cured any error in this regard, however, when it closely examined each mitigation measure in issuing the modified injunction of January 3, 2008. The Court of Appeals, too, conducted a detailed analysis of the record.

3

On the same day that CEQ issued its letter, the President granted the Navy an exemption from the requirements of the Coastal Zone Management Act of 1972 (CZMA) pursuant to 16 U. S. C. § 1456(c)(1)(B) (2006 ed.). That exemption, expressly authorized by the CZMA, does not affect NRDC’s NEPA claim.

4

The majority reasons that the environmental harm deserves less weight because the training exercises “have been taking place in SOCAL for the last 40 years,” such that “this is not a ease in which the defendant is conducting a new type of activity with completely unknown effects on the environment.” Ante, at 23. But the EA explains that the proposed action is not a continuation of the “status quo training.” App. 128. Instead, the EA is based on the Navy’s proposal to employ a “surge” training strategy, ibid., in which the commander “would have the option to conduct two concurrent major range events,” id., at 124.

CERTIORARI TO THE UNITED STATES COURT OF APPEALS FOR THE SECOND CIRCUIT

[4] Henry S. Cohn, Assistant Attorney General of Connecticut, argued the cause for petitioners. With him on the briefs were Clarine Nardi Riddle, Attorney General, Arnold B. Feigin and Carolyn K. Querijero, Assistant Attorneys General, and Andrew M. Calamari.

Joanne S. Faulkner argued the cause for respondent. With her on the brief were Brian Wolfman and Alan B. Morrison.^[1]

JUSTICE WHITE delivered an opinion, Parts I, II, and III of which are the opinion of the Court.^[2]

This case requires us to determine whether a state statute that authorizes prejudgment attachment of real estate without prior notice or hearing, without a showing of extraordinary circumstances, and without a requirement that the person seeking the attachment post a bond, satisfies the Due Process Clause of the Fourteenth Amendment. We hold that, as applied to this case, it does not.

[5] I

On March 15, 1988, petitioner John F. DiGiovanni submitted an application to the Connecticut Superior Court for an attachment in the amount of $75,000 on respondent Brian K. Doehr's home in Meriden, Connecticut. DiGiovanni took this step in conjunction with a civil action for assault and battery that he was seeking to institute against Doehr in the same court. The suit did not involve Doehr's real estate, nor did DiGiovanni have any pre-existing interest either in Doehr's home or any of his other property.

Connecticut law authorizes prejudgment attachment of real estate without affording prior notice or the opportunity for a prior hearing to the individual whose property is subject to the attachment. The State's prejudgment remedy statute provides, in relevant part:

"The court or a judge of the court may allow the prejudgment remedy to be issued by an attorney without hearing as provided in sections 52-278c and 52-278d upon verification by oath of the plaintiff or of some competent affiant, that there is probable cause to sustain the validity of the plaintiff's claims and (1) that the prejudgment remedy requested is for an attachment of real property . . . ." Conn. Gen. Stat. § 52-278e (1991).^[3]

[6] The statute does not require the plaintiff to post a bond to insure the payment of damages that the defendant may suffer should the attachment prove wrongfully issued or the claim prove unsuccessful.

As required, DiGiovanni submitted an affidavit in support of his application. In five one-sentence paragraphs, DiGiovanni stated that the facts set forth in his previously submitted complaint were true; that "I was willfully, wantonly and maliciously assaulted by the defendant, Brian K. Doehr"; that "[s]aid assault and battery broke my left wrist and further caused an ecchymosis to my right eye, as well as other injuries"; and that "I have further expended sums of money [7] for medical care and treatment." App. 24A. The affidavit concluded with the statement, "In my opinion, the foregoing facts are sufficient to show that there is probable cause that judgment will be rendered for the plaintiff." Ibid.

On the strength of these submissions the Superior Court Judge, by an order dated March 17, found "probable cause to sustain the validity of the plaintiff's claim" and ordered the attachment on Doehr's home "to the value of $75,000." The sheriff attached the property four days later, on March 21. Only after this did Doehr receive notice of the attachment. He also had yet to be served with the complaint, which is ordinarily necessary for an action to commence in Connecticut. Young v. Margiotta, 136 Conn. 429, 433, 71 A. 2d 924, 926 (1950). As the statute further required, the attachment notice informed Doehr that he had the right to a hearing: (1) to claim that no probable cause existed to sustain the claim; (2) to request that the attachment be vacated, modified, or dismissed or that a bond be substituted; or (3) to claim that some portion of the property was exempt from execution. Conn. Gen. Stat. § 52-278e(b) (1991).

Rather than pursue these options, Doehr filed suit against DiGiovanni in Federal District Court, claiming that § 52-278e (a)(1) was unconstitutional under the Due Process Clause of the Fourteenth Amendment.^[4] The District Court upheld the statute and granted summary judgment in favor of DiGiovanni. Pinsky v. Duncan, 716 F. Supp. 58 (Conn. 1989). On appeal, a divided panel of the United States Court of Appeals for the Second Circuit reversed. Pinsky v. Duncan, 898 F. 2d 852 (1990).^[5] Judge Pratt, who wrote the opinion [8] for the court, concluded that the Connecticut statute violated due process in permitting ex parte attachment absent a showing of extraordinary circumstances. "The rule to be derived from Sniadach v. Family Finance Corp. of Bay View, 395 U. S. 337 (1969), and its progeny, therefore, is not that postattachment hearings are generally acceptable provided that plaintiff files a factual affidavit and that a judicial officer supervises the process, but that a prior hearing may be postponed where exceptional circumstances justify such a delay, and where sufficient additional safeguards are present." Id., at 855. This conclusion was deemed to be consistent with our decision in Mitchell v. W. T. Grant Co., 416 U. S. 600 (1974), because the absence of a preattachment hearing was approved in that case based on the presence of extraordinary circumstances.

A further reason to invalidate the statute, the court ruled, was the highly factual nature of the issues in this case. In Mitchell, there were "uncomplicated matters that len[t] themselves to documentary proof" and "[t]he nature of the issues at stake minimize[d] the risk that the writ [would] be wrongfully issued by a judge." Id., at 609-610. Similarly, in Mathews v. Eldridge, 424 U. S. 319, 343-344 (1976), where an evidentiary hearing was not required prior to the termination of disability benefits, the determination of disability was "sharply focused and easily documented." Judge Pratt observed that in contrast the present case involved the fact-specific event of a fist fight and the issue of assault. He doubted that the judge could reliably determine probable cause when presented with only the plaintiff's version of the altercation. "Because the risk of a wrongful attachment is considerable under these circumstances, we conclude that dispensing with notice and opportunity for a hearing until after the attachment, without a showing of extraordinary circumstances, violates the requirements of due process." 898 F. 2d, at 856. Judge Pratt went on to conclude that in his view, the statute was also constitutionally infirm for its failure [9] to require the plaintiff to post a bond for the protection of the defendant in the event the attachment was ultimately found to have been improvident.

Judge Mahoney was also of the opinion that the statutory provision for attaching real property in civil actions, without a prior hearing and in the absence of extraordinary circumstances, was unconstitutional. He disagreed with Judge Pratt's opinion that a bond was constitutionally required. Judge Newman dissented from the holding that a hearing prior to attachment was constitutionally required and, like Judge Mahoney, disagreed with Judge Pratt on the necessity for a bond.

The dissent's conclusion accorded with the views of the Connecticut Supreme Court, which had previously upheld § 52-278e(b) in Fermont Division, Dynamics Corp. of America v. Smith, 178 Conn. 393, 423 A. 2d 80 (1979). We granted certiorari to resolve the conflict of authority. 498 U. S. 809 (1990).

II

With this case we return to the question of what process must be afforded by a state statute enabling an individual to enlist the aid of the State to deprive another of his or her property by means of the prejudgment attachment or similar procedure. Our cases reflect the numerous variations this type of remedy can entail. In Sniadach v. Family Finance Corp. of Bay View, 395 U. S. 337 (1969), the Court struck down a Wisconsin statute that permitted a creditor to effect prejudgment garnishment of wages without notice and prior hearing to the wage earner. In Fuentes v. Shevin, 407 U. S. 67 (1972), the Court likewise found a due process violation in state replevin provisions that permitted vendors to have goods seized through an ex parte application to a court clerk and the posting of a bond. Conversely, the Court upheld a Louisiana ex parte procedure allowing a lienholder to have disputed goods sequestered in Mitchell v. W. T. Grant Co., supra. Mitchell, however, carefully noted that Fuentes was [10] decided against "a factual and legal background sufficiently different . . . that it does not require the invalidation of the Louisiana sequestration statute." Id., at 615. Those differences included Louisiana's provision of an immediate postdeprivation hearing along with the option of damages; the requirement that a judge rather than a clerk determine that there is a clear showing of entitlement to the writ; the necessity for a detailed affidavit; and an emphasis on the lienholder's interest in preventing waste or alienation of the encumbered property. Id., at 615-618. In North Georgia Finishing, Inc. v. Di-Chem, Inc., 419 U. S. 601 (1975), the Court again invalidated an ex parte garnishment statute that not only failed to provide for notice and prior hearing but also failed to require a bond, a detailed affidavit setting out the claim, the determination of a neutral magistrate, or a prompt postdeprivation hearing. Id., at 606-608.

These cases "underscore the truism that `"[d]ue process," unlike some legal rules, is not a technical conception with a fixed content unrelated to time, place and circumstances.'" Mathews v. Eldridge, supra, at 334 (quoting Cafeteria & Restaurant Workers v. McElroy, 367 U. S. 886, 895 (1961)). In Mathews, we drew upon our prejudgment remedy decisions to determine what process is due when the government itself seeks to effect a deprivation on its own initiative. 424 U. S., at 334. That analysis resulted in the now familiar threefold inquiry requiring consideration of "the private interest that will be affected by the official action"; "the risk of an erroneous deprivation of such interest through the procedures used, and the probable value, if any, of additional or substitute safeguards"; and lastly "the Government's interest, including the function involved and the fiscal and administrative burdens that the additional or substitute procedural requirement would entail." Id., at 335.

Here the inquiry is similar, but the focus is different. Prejudgment remedy statutes ordinarily apply to disputes between private parties rather than between an individual and [11] the government. Such enactments are designed to enable one of the parties to "make use of state procedures with the overt, significant assistance of state officials," and they undoubtedly involve state action "substantial enough to implicate the Due Process Clause." Tulsa Professional Collection Services, Inc. v. Pope, 485 U. S. 478, 486 (1988). Nonetheless, any burden that increasing procedural safeguards entails primarily affects not the government, but the party seeking control of the other's property. See Fuentes v. Shevin, supra, at 99-101 (WHITE, J., dissenting). For this type of case, therefore, the relevant inquiry requires, as in Mathews, first, consideration of the private interest that will be affected by the prejudgment measure; second, an examination of the risk of erroneous deprivation through the procedures under attack and the probable value of additional or alternative safeguards; and third, in contrast to Mathews, principal attention to the interest of the party seeking the prejudgment remedy, with, nonetheless, due regard for any ancillary interest the government may have in providing the procedure or forgoing the added burden of providing greater protections.

We now consider the Mathews factors in determining the adequacy of the procedures before us, first with regard to the safeguards of notice and a prior hearing, and then in relation to the protection of a bond.

III

We agree with the Court of Appeals that the property interests that attachment affects are significant. For a property owner like Doehr, attachment ordinarily clouds title; impairs the ability to sell or otherwise alienate the property; taints any credit rating; reduces the chance of obtaining a home equity loan or additional mortgage; and can even place an existing mortgage in technical default where there is an insecurity clause. Nor does Connecticut deny that any of these consequences occurs.

[12] Instead, the State correctly points out that these effects do not amount to a complete, physical, or permanent deprivation of real property; their impact is less than the perhaps temporary total deprivation of household goods or wages. See Sniadach, supra, at 340; Mitchell, 416 U. S., at 613. But the Court has never held that only such extreme deprivations trigger due process concern. See Buchanan v. Warley, 245 U. S. 60, 74 (1917). To the contrary, our cases show that even the temporary or partial impairments to property rights that attachments, liens, and similar encumbrances entail are sufficient to merit due process protection. Without doubt, state procedures for creating and enforcing attachments, as with liens, "are subject to the strictures of due process." Peralta v. Heights Medical Center, Inc., 485 U. S. 80, 85 (1988) (citing Mitchell, supra, at 604; Hodge v. Muscatine County, 196 U. S. 276, 281 (1905)).^[6]

We also agree with the Court of Appeals that the risk of erroneous deprivation that the State permits here is substantial. By definition, attachment statutes premise a deprivation of property on one ultimate factual contingency—the award of damages to the plaintiff which the defendant may not be able to satisfy. See Ownbey v. Morgan, 256 U. S. 94, 104-105 (1921); R. Thompson & J. Sebert, Remedies: Damages, Equity and Restitution § 5.01 (1983). For attachments [13] before judgment, Connecticut mandates that this determination be made by means of a procedural inquiry that asks whether "there is probable cause to sustain the validity of the plaintiff's claim." Conn. Gen. Stat. § 52-278e(a) (1991). The statute elsewhere defines the validity of the claim in terms of the likelihood "that judgment will be rendered in the matter in favor of the plaintiff." Conn. Gen. Stat. § 52-278c(a)(2) (1991); Ledgebrook Condominium Assn. v. Lusk Corp., 172 Conn. 577, 584, 376 A. 2d 60, 63-64 (1977). What probable cause means in this context, however, remains obscure. The State initially took the position, as did the dissent below, that the statute requires a plaintiff to show the objective likelihood of the suit's success. Brief for Petitioners 12; Pinsky, 898 F. 2d, at 861-862 (Newman, J., dissenting). Doehr, citing ambiguous state cases, reads the provision as requiring no more than that a plaintiff demonstrate a subjective good-faith belief that the suit will succeed. Brief for Respondent 25-26. Ledgebrook Condominium Assn., supra, at 584, 376 A. 2d, at 63-64; Anderson v. Nedovich, 19 Conn. App. 85, 88, 561 A. 2d 948, 949 (1989). At oral argument, the State shifted its position to argue that the statute requires something akin to the plaintiff stating a claim with sufficient facts to survive a motion to dismiss.

We need not resolve this confusion since the statute presents too great a risk of erroneous deprivation under any of these interpretations. If the statute demands inquiry into the sufficiency of the complaint, or, still less, the plaintiff's good-faith belief that the complaint is sufficient, requirement of a complaint and a factual affidavit would permit a court to make these minimal determinations. But neither inquiry adequately reduces the risk of erroneous deprivation. Permitting a court to authorize attachment merely because the plaintiff believes the defendant is liable, or because the plaintiff can make out a facially valid complaint, would permit the deprivation of the defendant's property when the claim would fail to convince a jury, when it rested on factual allegations [14] that were sufficient to state a cause of action but which the defendant would dispute, or in the case of a mere good-faith standard, even when the complaint failed to state a claim upon which relief could be granted. The potential for unwarranted attachment in these situations is self-evident and too great to satisfy the requirements of due process absent any countervailing consideration.

Even if the provision requires the plaintiff to demonstrate, and the judge to find, probable cause to believe that judgment will be rendered in favor of the plaintiff, the risk of error was substantial in this case. As the record shows, and as the State concedes, only a skeletal affidavit need be, and was, filed. The State urges that the reviewing judge normally reviews the complaint as well, but concedes that the complaint may also be conclusory. It is self-evident that the judge could make no realistic assessment concerning the likelihood of an action's success based upon these one-sided, self-serving, and conclusory submissions. And as the Court of Appeals said, in a case like this involving an alleged assault, even a detailed affidavit would give only the plaintiff's version of the confrontation. Unlike determining the existence of a debt or delinquent payments, the issue does not concern "ordinarily uncomplicated matters that lend themselves to documentary proof." Mitchell, 416 U. S., at 609. The likelihood of error that results illustrates that "fairness can rarely be obtained by secret, one-sided determination of facts decisive of rights . . . . [And n]o better instrument has been devised for arriving at truth than to give a person in jeopardy of serious loss notice of the case against him and opportunity to meet it." Joint Anti-Fascist Refugee Comm. v. McGrath, 341 U. S. 123, 170-172 (1951) (Frankfurter, J., concurring).

What safeguards the State does afford do not adequately reduce this risk. Connecticut points out that the statute also provides an "expeditiou[s]" postattachment adversary hearing, [15] § 52-278e(c);^[7] notice for such a hearing, § 52-278e(b); judicial review of an adverse decision, § 52-2781(a); and a double damages action if the original suit is commenced without probable cause, § 52-568(a)(1). Similar considerations were present in Mitchell, where we upheld Louisiana's sequestration statute despite the lack of predeprivation notice and hearing. But in Mitchell, the plaintiff had a vendor's lien to protect, the risk of error was minimal because the likelihood of recovery involved uncomplicated matters that lent themselves to documentary proof, 416 U. S., at 609-610, and the plaintiff was required to put up a bond. None of these factors diminishing the need for a predeprivation hearing is present in this case. It is true that a later hearing might negate the presence of probable cause, but this would not cure the temporary deprivation that an earlier hearing might have prevented. "The Fourteenth Amendment draws no bright lines around three-day, 10-day or 50-day deprivations of property. Any significant taking of property by the State is within the purview of the Due Process Clause." Fuentes, 407 U. S., at 86.

[16] Finally, we conclude that the interests in favor of an ex parte attachment, particularly the interests of the plaintiff, are too minimal to supply such a consideration here. The plaintiff had no existing interest in Doehr's real estate when he sought the attachment. His only interest in attaching the property was to ensure the availability of assets to satisfy his judgment if he prevailed on the merits of his action. Yet there was no allegation that Doehr was about to transfer or encumber his real estate or take any other action during the pendency of the action that would render his real estate unavailable to satisfy a judgment. Our cases have recognized such a properly supported claim would be an exigent circumstance permitting postponing any notice or hearing until after the attachment is effected. See Mitchell, supra, at 609; Fuentes, supra, at 90-92; Sniadach, 395 U. S., at 339. Absent such allegations, however, the plaintiff's interest in attaching the property does not justify the burdening of Doehr's ownership rights without a hearing to determine the likelihood of recovery.

No interest the government may have affects the analysis. The State's substantive interest in protecting any rights of the plaintiff cannot be any more weighty than those rights themselves. Here the plaintiff's interest is de minimis. Moreover, the State cannot seriously plead additional financial or administrative burdens involving predeprivation hearings when it already claims to provide an immediate postdeprivation hearing. Conn. Gen. Stat. §§ 52-278e(b) and (c) (1991); Fermont, 178 Conn., at 397-398, 423 A. 2d, at 83.

Historical and contemporary practices support our analysis. Prejudgment attachment is a remedy unknown at common law. Instead, "it traces its origin to the Custom of London, under which a creditor might attach money or goods of the defendant either in the plaintiff's own hands or in the custody of a third person, by proceedings in the mayor's court or in the sheriff's court." Ownbey, 256 U. S., at 104. Generally speaking, attachment measures in both England and this [17] country had several limitations that reduced the risk of erroneous deprivation which Connecticut permits. Although attachments ordinarily did not require prior notice or a hearing, they were usually authorized only where the defendant had taken or threatened to take some action that would place the satisfaction of the plaintiff's potential award in jeopardy. See C. Drake, Law of Suits by Attachment, §§ 40-82 (1866) (hereinafter Drake); 1 R. Shinn, Attachment and Garnishment § 86 (1896) (hereinafter Shinn). Attachments, moreover, were generally confined to claims by creditors. Drake §§ 9-10; Shinn § 12. As we and the Court of Appeals have noted, disputes between debtors and creditors more readily lend themselves to accurate ex parte assessments of the merits. Tort actions, like the assault and battery claim at issue here, do not. See Mitchell, supra, at 609-610. Finally, as we will discuss below, attachment statutes historically required that the plaintiff post a bond. Drake §§ 114-183; Shinn § 153.

Connecticut's statute appears even more suspect in light of current practice. A survey of state attachment provisions reveals that nearly every State requires either a preattachment hearing, a showing of some exigent circumstance, or both, before permitting an attachment to take place. See Appendix to this opinion. Twenty-seven States, as well as the District of Columbia, permit attachments only when some extraordinary circumstance is present. In such cases, preattachment hearings are not required but postattachment hearings are provided. Ten States permit attachment without the presence of such factors but require prewrit hearings unless one of those factors is shown. Six States limit attachments to extraordinary circumstance cases, but the writ will not issue prior to a hearing unless there is a showing of some even more compelling condition.^[8] Three States always require a [18] preattachment hearing. Only Washington, Connecticut, and Rhode Island authorize attachments without a prior hearing in situations that do not involve any purportedly heightened threat to the plaintiff's interests. Even those States permit ex parte deprivations only in certain types of cases: Rhode Island does so only when the claim is equitable; Connecticut and Washington do so only when real estate is to be attached, and even Washington requires a bond. Conversely, the States for the most part no longer confine attachments to creditor claims. This development, however, only increases the importance of the other limitations.

We do not mean to imply that any given exigency requirement protects an attachment from constitutional attack. Nor do we suggest that the statutory measures we have surveyed are necessarily free of due process problems or other constitutional infirmities in general. We do believe, however, that the procedures of almost all the States confirm our view that the Connecticut provision before us, by failing to provide a preattachment hearing without at least requiring a showing of some exigent circumstance, clearly falls short of the demands of due process.

IV

A

Although a majority of the Court does not reach the issue, JUSTICES MARSHALL, STEVENS, O'CONNOR, and I deem it appropriate to consider whether due process also requires the plaintiff to post a bond or other security in addition to requiring a hearing or showing of some exigency.^[9]

[19] As noted, the impairments to property rights that attachments effect merit due process protection. Several consequences can be severe, such as the default of a homeowner's mortgage. In the present context, it need only be added that we have repeatedly recognized the utility of a bond in protecting property rights affected by the mistaken award of prejudgment remedies. Di-Chem, 419 U. S., at 610, 611 (Powell, J., concurring in judgment); id., at 619 (BLACKMUN, J., dissenting); Mitchell, 416 U. S., at 606, n. 8.

Without a bond, at the time of attachment, the danger that these property rights may be wrongfully deprived remains unacceptably high even with such safegnards as a hearing or exigency requirement. The need for a bond is especially apparent where extraordinary circumstances justify an attachment with no more than the plaintiff's ex parte assertion of a claim. We have already discussed how due process tolerates, and the States generally permit, the otherwise impermissible chance of erroneously depriving the defendant in such situations in light of the heightened interest of the plaintiff. Until a postattachment hearing, however, a defendant has no protection against damages sustained where no extraordinary circumstance in fact existed or the plaintiff's likelihood of recovery was nil. Such protection is what a bond can supply. Both the Court and its individual Members have repeatedly found the requirement of a bond to play an essential role in reducing what would have been too great a degree of risk in precisely this type of circumstance. Mitchell, [20] supra, at 610, 619; Di-Chem, 419 U. S., at 613 (Powell, J., concurring in judgment); id., at 619 (BLACKMUN, J., dissenting); Fuentes, 407 U. S., at 101 (WHITE, J., dissenting).

But the need for a bond does not end here. A defendant's property rights remain at undue risk even when there has been an adversarial hearing to determine the plaintiff's likelihood of recovery. At best, a court's initial assessment of each party's case cannot produce more than an educated prediction as to who will win. This is especially true when, as here, the nature of the claim makes any accurate prediction elusive. See Mitchell, supra, at 609-610. In consequence, even a full hearing under a proper probable-cause standard would not prevent many defendants from having title to their homes impaired during the pendency of suits that never result in the contingency that ultimately justifies such impairment, namely, an award to the plaintiff. Attachment measures currently on the books reflect this concern. All but a handful of States require a plaintiff's bond despite also affording a hearing either before, or (for the vast majority, only under extraordinary circumstances) soon after, an attachment takes place. See Appendix to this opinion. Bonds have been a similarly common feature of other prejudgment remedy procedures that we have considered, whether or not these procedures also included a hearing. See Ownbey, 256 U. S., at 101-102, n. 1; Fuentes, supra, at 73, n. 6, 75-76, n. 7, 81-82; Mitchell, supra, at 606, and n. 6; DiChem, supra, at 602-603, n. 1, 608.

The State stresses its double damages remedy for suits that are commenced without probable cause. Conn. Gen. Stat. § 52-568(a)(1).^[10] This remedy, however, fails to make [21] up for the lack of a bond. As an initial matter, the meaning of "probable cause" in this provision is no more clear here than it was in the attachment provision itself. Should the term mean the plaintiff's good faith or the facial adequacy of the complaint, the remedy is clearly insufficient. A defendant who was deprived where there was little or no likelihood that the plaintiff would obtain a judgment could nonetheless recover only by proving some type of fraud or malice or by showing that the plaintiff had failed to state a claim. Problems persist even if the plaintiff's ultimate failure permits recovery. At best a defendant must await a decision on the merits of the plaintiff's complaint, even assuming that a § 52-568(a)(1) action may be brought as a counterclaim. Hydro Air of Connecticut, Inc. v. Versa Technologies, Inc., 99 F. R. D. 111, 113 (Conn. 1983). Settlement, under Connecticut law, precludes seeking the damages remedy, a fact that encourages the use of attachments as a tactical device to pressure an opponent to capitulate. Blake v. Levy, 191 Conn. 257, 464 A. 2d 52 (1983). An attorney's advice that there is probable cause to commence an action constitutes a complete defense, even if the advice was unsound or erroneous. Vandersluis v. Weil, 176 Conn. 353, 361, 407 A. 2d 982, 987 (1978). Finally, there is no guarantee that the original plaintiff will have adequate assets to satisfy an award that the defendant may win.

Nor is there any appreciable interest against a bond requirement. Section 52-278e(a)(1) does not require a plaintiff to show exigent circumstances nor any pre-existing interest in the property facing attachment. A party must show more than the mere existence of a claim before subjecting an opponent to prejudgment proceedings that carry a significant risk of erroneous deprivation. See Mitchell, supra, at 604-609; Fuentes, supra, at 90-92; Sniadach, 395 U. S., at 339.

[22] B

Our foregoing discussion compels the four of us to consider whether a bond excuses the need for a hearing or other safeguards altogether. If a bond is needed to augment the protections afforded by preattachment and postattachment hearings, it arguably follows that a bond renders these safeguards unnecessary. That conclusion is unconvincing, however, for it ignores certain harms that bonds could not undo but that hearings would prevent. The law concerning attachments has rarely, if ever, required defendants to suffer an encumbered title until the case is concluded without any prior opportunity to show that the attachment was unwarranted. Our cases have repeatedly emphasized the importance of providing a prompt postdeprivation hearing at the very least. Mitchell, 416 U. S., at 606; Di-Chem, 419 U. S., at 606-607. Every State but one, moreover, expressly requires a preattachment or postattachment hearing to determine the propriety of an attachment.

The necessity for at least a prompt postattachment hearing is self-evident because the right to be compensated at the end of the case, if the plaintiff loses, for all provable injuries caused by the attachment is inadequate to redress the harm inflicted, harm that could have been avoided had an early hearing been held. An individual with an immediate need or opportunity to sell a property can neither do so, nor otherwise satisfy that need or recreate the opportunity. The same applies to a parent in need of a home equity loan for a child's education, an entrepreneur seeking to start a business on the strength of an otherwise strong credit rating, or simply a homeowner who might face the disruption of having a mortgage placed in technical default. The extent of these harms, moreover, grows with the length of the suit. Here, oral argument indicated that civil suits in Connecticut commonly take up to four to seven years for completion. Tr. of Oral Arg. 44. Many state attachment statutes require [23] that the amount of a bond be anywhere from the equivalent to twice the amount the plaintiff seeks. See, e. g., Utah Rule of Civ. Proc. 64C(b). These amounts bear no relation to the harm the defendant might suffer even assuming that money damages can make up for the foregoing disruptions. It should be clear, however, that such an assumption is fundamentally flawed. Reliance on a bond does not sufficiently account for the harms that flow from an erroneous attachment to excuse a State from reducing that risk by means of a timely hearing.

If a bond cannot serve to dispense with a hearing immediately after attachment, neither is it sufficient basis for not providing a preattachment hearing in the absence of exigent circumstances even if in any event a hearing would be provided a few days later. The reasons are the same: a wrongful attachment can inflict injury that will not fully be redressed by recovery on the bond after a prompt postattachment hearing determines that the attachment was invalid.

Once more, history and contemporary practices support our conclusion. Historically, attachments would not issue without a showing of extraordinary circumstances even though a plaintiff bond was almost invariably required in addition. Drake §§ 4, 114; Shinn §§ 86, 153. Likewise, all but eight States currently require the posting of a bond. Out of this 42-State majority, all but one requires a preattachment hearing, a showing of some exigency, or both, and all but one expressly require a postattachment hearing when an attachment has been issued ex parte. See Appendix to this opinion. This testimony underscores the point that neither a hearing nor an extraordinary circumstance limitation eliminates the need for a bond, no more than a bond allows waiver of these other protections. To reconcile the interests of the defendant and the plaintiff accurately, due process generally requires all of the above.

[24] V

Because Connecticut's prejudgment remedy provision, Conn. Gen. Stat. § 52-278e(a)(1), violates the requirements of due process by authorizing prejudgment attachment without prior notice or a hearing, the judgment of the Court of Appeals is affirmed, and the case is remanded to that court for further proceedings consistent with this opinion.

It is so ordered.

APPENDIX TO OPINION OF THE COURT

Prejudgment Attachment Statutes

-----------------------------------------------------------------------------------------------                                    Attachment                  Preattach.       Only in Exigent   Preattach.                  Hrg. Required       Circs.;        Hrg. Even in                 Unless Exigent     No Preattach.    Most Exigent    Bond       Postattach.                    Circs.          Hrg. Required       Circs.      Required   Hrg. Required-----------------------------------------------------------------------------------------------Alabama                                   X                            X            XAlaska               Preattachment hrg. always required.               XArizona              X                                                 X            XArkansas                                  X                            X            XCalifornia           X                                                 X            XColorado                                  X                            X            XConnecticut          X (or unless attachment of real estate)                        XDelaware                                  X                            X            XDC                                        X                            X            XFlorida                                   X                            X            XGeorgia                                   X                            X            XHawaii               Preattachment hrg. always required.               X            XIdaho                X                                                 X            XIllinois                                  X                            X            XIndiana                                   X                            X            XIowa                                      X                            X            XKansas                                    X                            X            XKentucky                                                 X             XLouisiana                                 X                            X            X

 [25] Maine                X                                                              XMaryland                                  X                            X            XMassachusetts        X                                                 X/O1         XMichigan                                  X                                         XMinnesota                                                X             X            XMississippi                               X                            X            XMissouri                                  X                            X            XMontana                                   X                            X            XNebraska                                  X                            X            XNevada               X                                                 X            XNew Hampshire        X                                                              XNew Jersey           X                                                 X/O          XNew Mexico                                X                            X            XNew York                                  X                            X            XNorth Carolina                            X                            X            XNorth Dakota                              X                            X            XOhio                                                     X             X            XOklahoma             X                                                 X            XOregon               Preattachment hrg. always required.               XPennsylvania         Rescinded in light of 530 F. 2d 1123 (CA3 1976).Rhode Island         X (but not if equitable claim)                    X/OSouth Carolina                            X                            X            XSouth Dakota                              X                            X            XTennessee                                 X                            X            X2Texas                                                    X             X            XUtah                                                     X             X            XVermont              X                                                              X1. An "X/O" in the "Bond Required" column indicates that a bond may berequired at the discretion of the court.2. The court may, under certain circumstances, quash the attachment atthe defendant's request without a hearing.

 [26] Virginia                                  X                            X            XWashington                                               X             X3      X                                          (except for real estate on a contract claim)West Virginia                             X                            X            XWisconsin                                 X                            X            XWyoming                                                  X             X            X-----------------------------------------------------------------------------------------------3. A bond is required except in situations in which the plaintiff seeks toattach the real property of a defendant who, after diligent efforts, cannotbe served.

CHIEF JUSTICE REHNQUIST, with whom JUSTICE BLACKMUN joins, concurring in part and concurring in the judgment.

I agree with the Court that the Connecticut attachment statute, "as applied to this case," ante, at 4, fails to satisfy the Due Process Clause of the Fourteenth Amendment. I therefore join Parts I, II, and III of its opinion. Unfortunately, the remainder of the opinion does not confine itself to the facts of this case, but enters upon a lengthy disquisition as to what combination of safeguards are required to satisfy due process in hypothetical cases not before the Court. I therefore do not join Part IV.

As the Court's opinion points out, the Connecticut statute allows attachment not merely for a creditor's claim, but for a tort claim of assault and battery; it affords no opportunity for a predeprivation hearing; it contains no requirement that there be "exigent circumstances," such as an effort on the part of the defendant to conceal assets; no bond is required from the plaintiff; and the property attached is one in which the plaintiff has no pre-existing interest. The Court's opinion [27] is, in my view, ultimately correct when it bases its holding of unconstitutionality of the Connecticut statute as applied here on our cases of Sniadach v. Family Finance Corp. of Bay View, 395 U. S. 337 (1969); Fuentes v. Shevin, 407 U. S. 67 (1972), Mitchell v. W. T. Grant Co., 416 U. S. 600 (1974), and North Georgia Finishing, Inc. v. Di-Chem, Inc., 419 U. S. 601 (1975). But I do not believe that the result follows so inexorably as the Court's opinion suggests. All of the cited cases dealt with personalty—bank deposits or chattels —and each involved the physical seizure of the property itself, so that the defendant was deprived of its use. These cases, which represented something of a revolution in the jurisprudence of procedural due process, placed substantial limits on the methods by which creditors could obtain a lien on the assets of a debtor prior to judgment. But in all of them the debtor was deprived of the use and possession of the property. In the present case, on the other hand, Connecticut's prejudgment attachment on real property statute, which secures an incipient lien for the plaintiff, does not deprive the defendant of the use or possession of the property.

The Court's opinion therefore breaks new ground, and I would point out, more emphatically than the Court does, the limits of today's holding. In Spielman-Fond, Inc. v. Hanson's, Inc., 379 F. Supp. 997, 999 (Ariz. 1973), the District Court held that the filing of a mechanics' lien did not cause the deprivation of a significant property interest of the owner. We summarily affirmed that decision. 417 U. S. 901 (1974). Other courts have read this summary affirmance to mean that the mere imposition of a lien on real property, which does not disturb the owner's use or enjoyment of the property, is not a deprivation of property calling for procedural due process safeguards. I agree with the Court, however, that upon analysis the deprivation here is a significant one, even though the owner remains in undisturbed possession. "For a property owner like Doehr, attachment ordinarily clouds title; impairs the ability to sell or otherwise [28] alienate the property; taints any credit rating; reduces the chance of obtaining a home equity loan or additional mortgage; and can even place an existing mortgage in technical default where there is an insecurity clause." Ante, at 11. Given the elaborate system of title records relating to real property which prevails in all of our States, a lienor need not obtain possession or use of real property belonging to a debtor in order to significantly impair its value to him.

But in Spielman-Fond, Inc., supra, there was, as the Court points out, ante, at 12, n. 4, an alternative basis available to this Court for affirmance of that decision. Arizona recognized a pre-existing lien in favor of unpaid mechanics and materialmen who had contributed labor or supplies which were incorporated in improvements to real property. The existence of such a lien upon the very property ultimately posted or noticed distinguishes those cases from the present one, where the plaintiff had no pre-existing interest in the real property which he sought to attach. Materialman's and mechanic's lien statutes award an interest in real property to workers who have contributed their labor, and to suppliers who have furnished material, for the improvement of the real property. Since neither the labor nor the material can be reclaimed once it has become a part of the realty, this is the only method by which workmen or small businessmen who have contributed to the improvement of the property may be given a remedy against a property owner who has defaulted on his promise to pay for the labor and the materials. To require any sort of a contested court hearing or bond before the notice of lien takes effect would largely defeat the purpose of these statutes.

Petitioners in their brief rely in part on our summary affirmance in Bartlett v. Williams, 464 U. S. 801 (1983). That case involved a lis pendens, in which the question presented to this Court was whether such a procedure could be valid when the only protection afforded to the owner of land affected by the lis pendens was a postsequestration hearing. [29] A notice of lis pendens is a well-established, traditional remedy whereby a plaintiff (usually a judgment creditor) who brings an action to enforce an interest in property to which the defendant has title gives notice of the pendency of such action to third parties; the notice causes the interest which he establishes, if successful, to relate back to the date of the filing of the lis pendens. The filing of such notice will have an effect upon the defendant's ability to alienate the property, or to obtain additional security on the basis of title to the property, but the effect of the lis pendens is simply to give notice to the world of the remedy being sought in the lawsuit itself. The lis pendens itself creates no additional right in the property on the part of the plaintiff, but simply allows third parties to know that a lawsuit is pending in which the plaintiff is seeking to establish such a right. Here, too, the fact that the plaintiff already claims an interest in the property which he seeks to enforce by a lawsuit distinguishes this class of cases from the Connecticut attachment employed in the present case.

Today's holding is a significant development in the law; the only cases dealing with real property cited in the Court's opinion, Peralta v. Heights Medical Center, Inc., 485 U. S. 80, 85 (1988), and Hodge v. Muscatine County, 196 U. S. 276, 281 (1905), arose out of lien foreclosure sales in which the question was whether the owner was entitled to proper notice. The change is dramatically reflected when we compare today's decision with the almost casual statement of Justice Holmes, writing for a unanimous Court in Coffin Brothers & Co. v. Bennett, 277 U. S. 29, 31 (1928):

"[N]othing is more common than to allow parties alleging themselves to be creditors to establish in advance by attachment a lien dependent for its effect upon the result of the suit."

The only protection accorded to the debtor in that case was the right to contest his liability in a postdeprivation proceeding.

[30] It is both unwise and unnecessary, I believe, for the plurality to proceed, as it does in Part IV, from its decision of the case before it to discuss abstract and hypothetical situations not before it. This is especially so where we are dealing with the Due Process Clause which, as the Court recognizes, "`"unlike some legal rules, is not a technical conception with a fixed content unrelated to time, place and circumstances,"'" ante, at 10. And it is even more true in a case involving constitutional limits on the methods by which the States may transfer or create interests in real property; in other areas of the law, dicta may do little damage, but those who insure titles or write title opinions often do not enjoy the luxury of distinguishing detween dicta and holding.

The two elements of due process with which the Court concerns itself in Part IV—the requirements of a bond and of "exigent circumstances" — prove to be upon analysis so vague that the discussion is not only unnecessary, but not particularly useful. Unless one knows what the terms and conditions of a bond are to be, the requirement of a "bond" in the abstract means little. The amount to be secured by the bond and the conditions of the bond are left unaddressed —is there to be liability on the part of a plaintiff if he is ultimately unsuccessful in the underlying lawsuit, or is it instead to be conditioned on some sort of good-faith test? The "exigent circumstances" referred to by the Court are admittedly equally vague; nonresidency appears to be enough in some States, an attempt to conceal assets is required in others, an effort to flee the jurisdiction in still others. We should await concrete cases which present questions involving bonds and exigent circumstances before we attempt to decide when and if the Due Process Clause of the Fourteenth Amendment requires them as prerequisites for a lawful attachment.

JUSTICE SCALIA, concurring in part and concurring in the judgment.

Since the manner of attachment here was not a recognized procedure at common law, cf. Pacific Mut. Life Ins. Co. v. [31] Haslip, 499 U. S. 1, 24 (1991) (SCALIA, J., concurring in judgment), I agree that its validity under the Due Process Clause should be determined by applying the test we set forth in Mathews v. Eldridge, 424 U. S. 319 (1976); and I agree that it fails that test. I join Parts I and III of the Court's opinion, and concur in the judgment of the Court.

[1] Allan B. Taylor, James J. Tancredi, and Kirk D. Tavtigian, Jr., filed a brief for the Connecticut Bankers Association et al. as amici curiae urging reversal.

[2] THE CHIEF JUSTICE, JUSTICE BLACKMUN, JUSTICE KENNEDY, and JUSTICE SOUTER join Parts I, II, and III of this opinion, and JUSTICE SCALIA joins Parts I and III.

[3]The complete text of § 52-278e reads:

"Allowance of prejudgment remedy without hearing. Notice to defendant. Subsequent hearing and order. Attachment of real property of municipal officers. (a) The court or a judge of the court may allow the prejudgment remedy to be issued by an attorney without hearing as provided in sections 52-278c and 52-278d upon verification by oath of the plaintiff or of some competent affiant, that there is probable cause to sustain the validity of the plaintiff's claim and (1) that the prejudgment remedy requested is for an attachment of real property; or (2) that there is reasonable likelihood that the defendant (A) neither resides in nor maintains an office or place of business in this state and is not otherwise subject to jurisdiction over his person by the court, or (B) has hidden or will hide himself so that process cannot be served on him or (C) is about to remove himself or his property from this state or (D) is about to fraudulently dispose of or has fraudulently disposed of any of his property with intent to hinder, delay or defraud his creditors or (E) has fraudulently hidden or withheld money, property or effects which should be liable to the satisfaction of his debts or (F) has stated he is insolvent or has stated he is unable to pay his debts as they mature.

"(b) If a prejudgment remedy is granted pursuant to this section, the plaintiff shall include in the process served on the defendant the following notice prepared by the plaintiff: YOU HAVE RIGHTS SPECIFIED IN THE CONNECTICUT GENERAL STATUTES, INCLUDING CHAPTER 903a, WHICH YOU MAY WISH TO EXERCISE CONCERNING THIS PREJUDGMENT REMEDY. THESE RIGHTS INCLUDE: (1) THE RIGHT TO A HEARING TO OBJECT TO THE PREJUDGMENT REMEDY FOR LACK OF PROBABLE CAUSE TO SUSTAIN THE CLAIM; (2) THE RIGHT TO A HEARING TO REQUEST THAT THE PREJUDGMENT REMEDY BE MODIFIED, VACATED OR DISMISSED OR THAT A BOND BE SUBSTITUTED; AND (3) THE RIGHT TO A HEARING AS TO ANY PORTION OF THE PROPERTY ATTACHED WHICH YOU CLAIM IS EXEMPT FROM EXECUTION.

"(c) The defendant appearing in such action may move to dissolve or modify the prejudgment remedy granted pursuant to this section in which event the court shall proceed to hear and determine such motion expeditiously. If the court determines at such hearing requested by the defendant that there is probable cause to sustain the validity of the plaintiff's claim, then the prejudgment remedy granted shall remain in effect. If the court determines there is no such probable cause, the prejudgment remedy shall be dissolved. An order shall be issued by the court setting forth the action it has taken."

[4] Three other plaintiffs joined Doehr, challenging § 52-278e(a)(1) out of separate instances of attachment by different defendants. These other plaintiffs and defendants did not participate in the Court of Appeals and are no longer parties in this case.

[5] The Court of Appeals invited Connecticut to intervene pursuant to 28 U. S. C. § 2403(b) after oral argument. The State elected to intervene in the appeal and has fully participated in the proceedings before this Court.

[6] Our summary affirmance in Spielman-Fond, Inc. v. Hanson's, Inc., 417 U. S. 901 (1974), does not control. In Spielman-Fond, the District Court held that the filing of a mechanic's lien did not amount to the taking of a significant property interest. 379 F. Supp. 997, 999 (Ariz. 1973) (three-judge court) (per curiam). A summary disposition does not enjoy the full precedential value of a case argued on the merits and disposed of by a written opinion. Edelman v. Jordan, 415 U. S. 651, 671 (1974). The facts of Spielman-Fond presented an alternative basis for affirmance in any event. Unlike the case before us, the mechanic's lien statute in Spielman-Fond required the creditor to have a pre-existing interest in the property at issue. 379 F. Supp., at 997. As we explain below, a heightened plaintiff interest in certain circumstances can provide a ground for upholding procedures that are otherwise suspect. Infra, at 15.

[7] The parties vigorously dispute whether a defendant can in fact receive a prompt hearing. Doehr contends that the State's rules of practice prevent the filing of any motion—including a motion for the mandated postattachment hearing—until the return date on the complaint, which in this case was 30 days after service. Connecticut Practice Book § 114 (1988). Under state law at least 12 days must elapse between service on the defendant and the return date. Conn. Gen. Stat. § 52-46 (1991). The State counters that the postattachment hearing is available upon request. See Fermont Division, Dynamics Corp. of America v. Smith, 178 Conn. 393, 397-398, 423 A. 2d 80, 83 (1979) ("Most important, the statute affords to the defendant whose property has been attached the opportunity to obtain an immediate postseizure hearing at which the prejudgment remedy will be dissolved unless the moving party proves probable cause to sustain the validity of his claim"). We assume, without deciding, that the hearing is prompt. Even on this assumption, the State's procedures fail to provide adequate safeguards against the erroneous deprivation of the property interest at stake.

[8] One State, Pennsylvania, has not had an attachment statute or rule since the decision in Jonnet v. Dollar Savings Bank of New York City, 530 F. 2d 1123 (CA3 1976).

[9] Ordinarily we will not address a contention advanced by a respondent that would enlarge his or her rights under a judgment, without the respondent filing a cross-petition for certiorari. E. g., Trans World Airlines, Inc. v. Thurston, 469 U. S. 111, 119, n. 14 (1985). Here the Court of Appeals rejected Doehr's argument that § 52-278e(a)(1) violates due process in failing to mandate a preattachment bond. Nonetheless, this case involves considerations that in the past have prompted us "to consider the question highlighted by respondent." Berkemer v. McCarty, 468 U. S. 420, 435-436, n. 23 (1984). First, as our cases have shown, the notice and hearing question and the bond question are intertwined and can fairly be considered facets of the same general issue. Thus, "[w]ithout undue strain, the position taken by respondent before this Court . . . might be characterized as an argument in support of the judgment below" insofar as a discussion of notice and a hearing cannot be divorced from consideration of a bond. Ibid. Second, this aspect of prejudgment attachment "plainly warrants our attention, and with regard to which the lower courts are in need of guidance." Ibid. Third, "and perhaps most importantly, both parties have briefed and argued the question." Ibid.

[10]Section 52-568(a)(1) provides:

"Any person who commences and prosecutes any civil action or complaint against another, in his own name, or the name of others, or asserts a defense to any civil action or complaint commenced and prosecuted by another (1) without probable cause, shall pay such other person double damages, or (2) without probable cause, and with a malicious intent unjustly to vex and trouble such other person, shall pay him treble damages."

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.

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."