ROCKWELL INT’L CORP. v. UNITED STATES (No. 05-1272)
92 Fed. Appx. 708, reversed.
Syllabus

Opinion
[Scalia]
Dissent
[Stevens]
HTML version
PDF version
HTML version
PDF version
HTML version
PDF version

Justice Scalia, Opinion of the Court

NOTICE: This opinion is subject to formal revision before publication in the preliminary print of the United States Reports. Readers are requested to notify the Reporter of Decisions, Supreme Court of the United States, Washington, D. C. 20543, of any typographical or other formal errors, in order that corrections may be made before the preliminary print goes to press.

SUPREME COURT OF THE UNITED STATES

ROCKWELL INTERNATIONAL CORP., et al.,
PETITIONERS v. UNITED STATES et al.

on writ of certiorari to the united states court ofappeals for the tenth circuit


[March 27, 2007]

Justice Scalia delivered the opinion of the Court.

The False Claims Act, 31 U. S. C. §§3729–3733, eliminates federal-court jurisdiction over actions under §3730 of the Act that are based upon the public disclosure of allegations or transactions “unless the action is brought by the Attorney General or the person bringing the action is an original source of the information.” §3730(e)(4)(A). We decide whether respondent James Stone was an original source.

I

The mixture of concrete and pond sludge that is the subject of this case has taken nearly two decades to seep, so to speak, into this Court. Given the long history and the complexity of this litigation, it is well to describe the facts in some detail.

A

From 1975 through 1989, petitioner Rockwell International Corp. was under a management and operating contract with the Department of Energy (DOE) to run the Rocky Flats nuclear weapons plant in Colorado. The most significant portion of Rockwell’s compensation came in the form of a semiannual “ ‘award fee,’ ” the amount of which depended on DOE’s evaluation of Rockwell’s performance in a number of areas, including environmental, safety, and health concerns. United States ex rel. Stone v. Rockwell Int’l, Corp., 92 Fed. Appx. 708, 714 (CA10 2004).

From November 1980 through March 1986, James Stone worked as an engineer at the Rocky Flats plant. In the early 1980’s, Rockwell explored the possibility of disposing of the toxic pond sludge that accumulated in solar evaporation ponds at the facility, by mixing it with cement. The idea was to pour the mixture into large rectangular boxes, where it would solidify into “pondcrete” blocks that could be stored onsite or transported to other sites for disposal.

Stone reviewed a proposed manufacturing process for pondcrete in 1982. He concluded that the proposal “would not work,” App. 175, and communicated that conclusion to Rockwell management in a written “Engineering Order.” As Stone would later explain, he believed “the suggested process would result in an unstable mixture that would later deteriorate and cause unwanted release of toxic wastes to the environment.” Ibid. He believed this because he “foresaw that the piping system” that extracted sludge from the solar ponds “would not properly remove the sludge and would lead to an inadequate mixture of sludge/waste and cement such that the ‘pond crete’ blocks would rapidly disintegrate thus creating additional contamination problems.” Id., at 290.

Notwithstanding Stone’s prediction, Rockwell proceeded with its pondcrete project and successfully manufactured “concrete hard” pondcrete during the period of Stone’s employment at Rocky Flats. It was only after Stone was laid off in March 1986 that what the parties have called “insolid” pondcrete blocks were discovered. According to respondents, Rockwell knew by October 1986 that a substantial number of pondcrete blocks were insolid, but DOE did not become aware of the problem until May 1988, when several pondcrete blocks began to leak, leading to the discovery of thousands of other insolid blocks. The media reported these discoveries, 3 Appellants’ App. in Nos. 99–1351, 99–1352, 99–1353 (CA10), pp. 889–38 to 889–39; and attributed the malfunction to Rockwell’s reduction of the ratio of concrete to sludge in the mixture.

In June 1987, more than a year after he had left Rockwell’s employ, Stone went to the Federal Bureau of Investigation (FBI) with allegations of environmental crimes at Rocky Flats during the time of his employment. According to the court below, Stone alleged that

“contrary to public knowledge, Rocky Flats accepted hazardous and nuclear waste from other DOE facilities; that Rockwell employees were ‘forbidden from discussing any controversies in front of a DOE employee’; that although Rocky Flats’ fluid bed incinerators failed testing in 1981, the pilot incinerator remained on line and was used to incinerate wastes daily since 1981, including plutonium wastes which were then sent out for burial; that Rockwell distilled and fractionated various oils and solvents although the wastes were geared for incineration; that Stone believed that the ground water was contaminated from previous waste burial and land application, and that hazardous waste lagoons tended to overflow during and after ‘a good rain,’ causing hazardous wastes to be discharged without first being treated.” App. to Pet. for Cert. 4a.

Stone provided the FBI with 2,300 pages of documents, buried among which was his 1982 engineering report predicting that the pondcrete-system design would not work. Stone did not discuss his pondcrete allegations with the FBI in their conversations.1

Based in part on information allegedly learned from Stone, the Government obtained a search warrant for Rocky Flats, and on June 6, 1989, 75 FBI and Environmental Protection Agency agents raided the facility. The affidavit in support of the warrant included allegations (1) that pondcrete blocks were insolid “due to an inadequate waste-concrete mixture,” App. 429, (2) that Rockwell obtained award fees based on its alleged “ ‘excellent’ ” management of Rocky Flats, id., at 98, and (3) that Rockwell made false statements and concealed material facts in violation of the Resource Conservation and Recovery Act of 1976 (RCRA), 90 Stat. 2811, 42 U. S. C. §6928, and 18 U. S. C. §1001. Newspapers published these allegations. In March 1992, Rockwell pleaded guilty to 10 environmental violations, including the knowing storage of insolid pondcrete blocks in violation of RCRA. Rockwell agreed to pay $18.5 million in fines.

B

In July 1989, Stone filed a qui tam suit under the False Claims Act.2 That Act prohibits false or fraudulent claims for payment to the United States, 31U. S. C. §3729(a), and authorizes civil actions to remedy such fraud to be brought by the Attorney General, §3730(a), or by private individuals in the Government’s name, §3730(b)(1). The Act provides, however, that “[n]o court shall have jurisdiction over an action under this section based upon the public disclosure of allegations or transactions … from the news media, unless the action is brought by the Attorney General or the person bringing the action is an original source of the information.” §3730(e)(4)(A). An “original source” is “an individual who has direct and independent knowledge of the information on which the allegations are based and has voluntarily provided the information to the Government before filing an action under this section which is based on the information.” §3730(e)(4)(B).

Stone’s complaint alleged that Rockwell was required to comply with certain federal and state environmental laws and regulations, including RCRA; that Rockwell committed numerous violations of these laws and regulations throughout the 1980’s3; and that, in order to induce the Government to make payments or approvals under Rockwell’s contract, Rockwell knowingly presented false and fraudulent claims to the Government in violation of the False Claims Act, 31 U. S. C. §3729(a). As required under the Act, Stone filed his complaint under seal and simultaneously delivered to the Government a confidential disclosure statement describing “substantially all material evidence and information” in his possession, §3730(b)(2). The statement identified 26 environmental and safety issues, only one of which involved pondcrete. With respect to that issue, Stone explained in his statement that he had reviewed the design for the pondcrete system and had foreseen that the piping mechanism would not properly remove the sludge, which in turn would lead to an inadequate mixture of sludge and cement.

In December 1992, Rockwell moved to dismiss Stone’s action for lack of subject-matter jurisdiction, arguing that the action was based on publicly disclosed allegations and that Stone was not an original source. The District Court denied the motion because, in its view, “Stone had direct and independent knowledge that Rockwell’s compensation was linked to its compliance with environmental, health and safety regulations and that it allegedly concealed its deficient performance so that it would continue to receive payments.” App. to Pet. for Cert. 61a.

The Government initially declined to intervene in Stone’s action, but later reversed course, and in November 1996, the District Court granted the Government’s intervention. Several weeks later, at the suggestion of the District Court, the Government and Stone filed a joint amended complaint. As relevant here, the amended complaint alleged that Rockwell violated RCRA by storing leaky pondcrete blocks, but did not allege that any defect in the piping system (as predicted by Stone) caused insolid pondcrete.4 Respondents clarified their allegations even further in a statement of claims which became part of the final pretrial order and which superseded their earlier pleadings. This said that the pondcrete’s insolidity was due to “an incorrect cement/sludge ratio used in pondcrete operations, as well as due to inadequate process controls and inadequate inspection procedures.” App. 470. It continued:

“During the winter of 1986, Rockwell replaced its then pondcrete foreman, Norman Fryback, with Ron Teel. Teel increased pondcrete production rates in part by, among other things, reducing the amount of cement added to the blocks. Following the May 23, 1988 spill, Rockwell acknowledged that this reduced cement-to-sludge ratio was a major contributor to the existence of insufficiently solid pondcrete blocks on the storage pads.” Id., at 476–477.

The statement of claims again did not mention the piping problem asserted by Stone years earlier.

Respondents’ False Claims Act claims went to trial in 1999. None of the witnesses Stone had identified during discovery as having relevant knowledge testified at trial. And none of the documents Stone provided to the Government with his confidential disclosure statement was introduced in evidence at trial. Nor did respondents allege at trial that the defect in the piping system predicted by Stone caused insolid pondcrete. To the contrary, during closing arguments both Stone’s counsel and the Government’s counsel repeatedly explained to the jury that the pondcrete failed because Rockwell’s new foreman used an insufficient cement-to-sludge ratio in an effort to increase pondcrete production.

The verdict form divided the False Claims Act count into several different claims corresponding to different award-fee periods. The jury found in favor of respondents for the three periods covering the pondcrete allegations (April 1, 1987, to September 30, 1988), and found for Rockwell as to the remaining periods. The jury awarded damages of $1,390,775.80, which the District Court trebled pursuant to 31 U. S. C. §3729(a).

Rockwell filed a postverdict motion to dismiss Stone’s claims under §3730(e)(4), arguing that the claims were based on publicly disclosed allegations and that Stone was not an original source. In response, Stone acknowledged that his successful claims were based on publicly disclosed allegations, but asserted original-source status. The District Court agreed with Stone. The United States Court of Appeals for the Tenth Circuit affirmed in relevant part, but remanded the case for the District Court to determine whether Stone had disclosed his information to the Government before filing his qui tam action, as §3730(e)(4)(B) required. On remand, the District Court found that Stone had produced the 1982 engineering order to the Government, but that the order was insufficient to communicate Stone’s allegations. The District Court also found that Stone had not carried his burden of proving that he orally informed the FBI about his allegations before filing suit. On appeal, the Tenth Circuit disagreed with the District Court’s conclusion and held (over the dissent of Judge Briscoe) that the 1982 engineering order sufficed to carry Stone’s burden of persuasion. 92 Fed. Appx. 708. We granted certiorari, 548 U. S. ____ (2006), to decide whether Stone was an original source.

II

Section 3730(e)(4)(A) provides that

“[n]o court shall have jurisdiction over an action under this section based upon the public disclosure of allegations or transactions in a criminal, civil, or administrative hearing, in a congressional, administrative, or Government Accounting Office report, hearing, audit, or investigation, or from the news media, unless the action is brought by the Attorney General or the person bringing the action is an original source of the information.” (Footnote omitted.)

As discussed above, §3730(e)(4)(B) defines “original source” as “an individual who [1] has direct and independent knowledge of the information on which the allegations are based and [2] has voluntarily provided the information to the Government before filing an action under this section which is based on the information.” As this case comes to the Court, it is conceded that the claims on which Stone prevailed were based upon publicly disclosed allegations within the meaning of §3730(e)(4)(A). The question is whether Stone qualified under the original-source exception to the public-disclosure bar.

We begin with the possibility that little analysis is required in this case, for Stone asserts that Rockwell conceded his original-source status. Rockwell responds that it conceded no such thing and that, even had it done so, the concession would have been irrelevant because §3730(e)(4) is jurisdictional. We agree with the latter proposition. It is true enough that the word “jurisdiction” does not in every context connote subject-matter jurisdiction. Noting that “jurisdiction” is “ ‘a word of many, too many, meanings,’ ” we concluded in Steel Co. v. Citizens for Better Environment, 523 U. S. 83 (1998) , that establishing the elements of an offense was not made a jurisdictional matter merely because the statute creating the cause of action was phrased as providing for “jurisdiction” over such suits. Id., at 90 (quoting United States v. Vanness, 85 F. 3d 661, 663, n. 2 (CADC 1996)). Here, however, the issue is not whether casting the creation of a cause of action in jurisdictional terms somehow limits the general grant of jurisdiction under which that cause of action would normally be brought, but rather whether a clear and explicit withdrawal ofjurisdiction withdraws jurisdiction. It undoubtedly does so. Just last Term we stated that, “[i]f the Legislature clearly states that a threshold limitation on a statute’s scope shall count as jurisdictional, the courts and litigants will be duly instructed and will not be left to wrestle with the issue.” Arbaugh v. Y & H Corp., 546 U. S. 500, 515–516 (2006) (footnote omitted). Here the jurisdictional nature of the original-source requirement is clear ex visceribus verborum. Indeed, we have already stated that §3730(e)(4) speaks to “the power of a particular court” as well as “the substantive rights of the parties.” Hughes Aircraft Co. v. United States ex rel. Schumer, 520 U. S. 939, 951 (1997) .

Stone’s contrary position rests entirely on dicta from a single Court of Appeals decision, see United States ex rel. Fallon v. Accudyne Corp., 97 F. 3d 937, 940–941 (CA7 1996). Accudyne thought it significant that jurisdiction over False Claims Act cases is conferred by 28 U. S. C. §§1331 and 1345 (the federal-question and United-States-as-plaintiff provisions of the Judicial Code) and 31 U. S. C. §3732(a) (the provision of the False Claims Act establishing federal-court venue and conferring federal-court jurisdiction over related state-law claims), rather than §3730, which is the “section” referenced in §3730(e)(4). To eliminate jurisdiction, the court believed, it is those jurisdiction-conferring sections that would have to be referenced. We know of nothing in logic or authority to support this. The jurisdiction-removing provision here does not say “no court shall have jurisdiction under this section,” but rather “no court shall have jurisdiction over an action under this section.” That is surely the most natural way to achieve the desired result of eliminating jurisdiction over a category of False Claims Act actions—rather than listing all the conceivable provisions of the United States Code whose conferral of jurisdiction is being eliminated. (In addition to the provisions cited by the Accudyne court, one might also have to mention the diversity-jurisdiction provision, 28 U. S. C. §1332, and the supplemental-jurisdiction provision, §1367.) Accudyne next observed that the public-disclosure bar limits only who may speak for the United States on a subject and who if anyone gets a financial reward, not the “categories of disputes that may be resolved (a real ‘jurisdictional’ limit).” 97 F. 3d, at 941. But this is a classic begging of the question, which is precisely whether there has been removed from the courts’ jurisdiction that category of disputes consisting of False Claims Act qui tam suits based on publicly disclosed allegations as to which the relator is not an original source of the information. Nothing prevents Congress from defining the “category” of excluded suits in any manner it wishes. See, e.g., 28 U. S. C. §1500 (no jurisdiction over “any claim for or in respect to which the plaintiff . . . has pending in any other court any suit … against the United States”). Lastly, Accudyne asserted that “the Supreme Court had held that a similar reference to jurisdiction in the Norris-LaGuardia Act, 29 U. S. C. §§101, limits remedies rather than subject-matter jurisdiction.” 97 F. 3d, at 941 (citing Burlington Northern R. Co. v. Maintenance of Way Employes, 481 U. S. 429, 444–446 (1987) ). But the language of the Norris-LaGuardia Act is in fact not similar. It provides that “[n]o court of the United States shall have jurisdiction to issue any restraining order or temporary or permanent injunction in any case involving or growing out of any labor dispute … .” 29 U. S. C. §104 (emphasis added). It is facially a limitation upon the relief that can be accorded, not a removal of jurisdiction over “any case involving or growing out of a labor dispute.” Here, by contrast, the text says “[n]o court shall have jurisdiction over an action under this section.”

Whether the point was conceded or not, therefore, we may, and indeed must, decide whether Stone met the jurisdictional requirement of being an original source.

III

We turn to the first requirement of original-source status, that the relator have “direct and independent knowledge of the information on which the allegations are based.” 31 U. S. C. §3730(e)(4)(B). Because we have not previously addressed this provision, several preliminary questions require our attention.

A

First, does the phrase “information on which the allegations are based” refer to the information on which the relator’s allegations are based or the information on which the publicly disclosed allegations that triggered the public-disclosure bar are based? The parties agree it is the former. See Brief for Petitioners 26, n. 13; Brief for United States 24, and n. 8; Brief for Respondent Stone 15, 21. But in view of our conclusion that §3730(e)(4) is jurisdictional, we must satisfy ourselves that the parties’ position is correct.

Though the question is hardly free from doubt,5 we agree that the “information” to which subparagraph (B) speaks is the information upon which the relators’ allegations are based. To begin with, subparagraph (B) standing on its own suggests that disposition. The relator must have “direct and independent knowledge of the information on which the allegations are based,” and he must “provid[e] the information to the Government before filing an action under this section which is based on the information.” Surely the information one would expect a relator to “provide to the Government before filing an action … based on the information” is the information underlying the relator’s claims.

Subparagraph (A) complicates matters. As described earlier, it bars actions based on the “public disclosure of allegations or transactions” and provides an exception for cases brought by “an original source of the information.” If the allegations referred to in subparagraph (B)’s phrase requiring “direct and independent knowledge of the information on which the allegations are based,” are the same “allegations” referred to in subparagraph (A), then original-source status would depend on knowledge of information underlying the publicly disclosed allegations. The principal textual difficulty with that interpretation is that subparagraph (A) does not speak simply of “allegations,” but of “allegations or transactions.” Had Congress wanted to link original-source status to information underlying the public disclosure, it would surely have used the identical phrase, “allegations or transactions”; there is no conceivable reason to require direct and independent knowledge of publicly disclosed allegations but not of publicly disclosed transactions.

The sense of the matter offers strong additional support for this interpretation. Section 3730(e)(4)(A) bars actions based on publicly disclosed allegations whether or not the information on which those allegations are based has been made public. It is difficult to understand why Congress would care whether a relator knows about the information underlying a publicly disclosed allegation (e.g., what a confidential source told a newspaper reporter about insolid pondcrete) when the relator has direct and independent knowledge of different information supporting the same allegation (e.g., that a defective process would inevitably lead to insolid pondcrete). Not only would that make little sense, it would raise nettlesome procedural problems, placing courts in the position of comparing the relator’s information with the often unknowable information on which the public disclosure was based. Where that latter information has not been disclosed (by reason, for example, of a reporter’s desire to protect his source), the relator would presumably be out of court. To bar a relator with direct and independent knowledge of information underlying his allegations just because no one can know what information underlies the similar allegations of some other person simply makes no sense.

The contrary conclusion of some lower courts rests on the following logic: The term “information” in subparagraph (B) must be read in tandem with the term “information” in subparagraph (A), and the term “information” in subparagraph (A) refers to the information on which the publicly disclosed allegations are based. See, e.g., United States ex rel. Laird v. Lockheed Martin Eng. & Science Servs. Co., 336 F. 3d 346, 354 (CA5 2003). The major premise of this reasoning seems true enough: “information” in (A) and (B) means the same thing. The minor premise, however—that “information” in (A) refers to the information underlying the publicly disclosed allegations or transactions—is highly questionable. The complete phrase at issue is “unless … the person bringing the action is an original source of the information.” It seems to us more likely (in light of the analysis set forth above) that the information in question is the information underlying the action referred to a few words earlier, to-wit, the action “based upon the public disclosure of allegations or transactions” referred to at the beginning of the provision. On this interpretation, “information” in subparagraph (A) and “information on which the allegations are based” in subparagraph (B) are one and the same, viz., information underlying the allegations of the relator’s action.

B

Having determined that the phrase “information on which the allegations are based” refers to the relator’s allegations and not the publicly disclosed allegations, we confront more textual ambiguity: Which of the relator’s allegations are the relevant ones? Stone’s allegations changed during the course of the litigation, yet he asks that we look only to his original complaint. Rockwell argues that Stone must satisfy the original-source exception through all stages of the litigation.

In our view, the term “allegations” is not limited to the allegations of the original complaint. It includes (at a minimum) the allegations in the original complaint as amended. The statute speaks not of the allegations in the “original complaint” (or even the allegations in the “complaint”), but of the relator’s “allegations” simpliciter. Absent some limitation of §3730(e)(4)’s requirement to the relator’s initial complaint, we will not infer one. Such a limitation would leave the relator free to plead a trivial theory of fraud for which he had some direct and independent knowledge and later amend the complaint to include theories copied from the public domain or from materials in the Government’s possession. Even the Government concedes that new allegations regarding a fundamentally different fraudulent scheme require reevaluation of the court’s jurisdiction. See Brief for United States 40; Tr. of Oral Arg. 40.

The rule that subject-matter jurisdiction “depends on the state of things at the time of the action brought,” Mollan v. Torrance, 9 Wheat. 537, 539 (1824), does not suggest a different interpretation. The state of things and the originally alleged state of things are not synonymous; demonstration that the original allegations were false will defeat jurisdiction. Anderson v. Watt, 138 U. S. 694, 701 (1891) ; Morris v. Gilmer, 129 U. S. 315, 326 (1889) . So also will the withdrawal of those allegations, unless they are replaced by others that establish jurisdiction. Thus, when a plaintiff files a complaint in federal court and then voluntarily amends the complaint, courts look to the amended complaint to determine jurisdiction. See Wellness Community-Nat. v. Wellness House, 70 F. 3d 46, 49 (CA7 1995); Boelens v. Redman Homes, Inc., 759 F. 2d 504, 508 (CA5 1984).6

Here, we have not only an amended complaint, but a final pretrial order that superseded all prior pleadings and “controll[ed] the subsequent course of the action,” Fed. Rule Civ. Proc. 16(e). See Curtis v. Loether, 415 U. S. 189, 190, n. 1 (1974) (where a claim was not included in the complaint, but was included in the pretrial order, “it is irrelevant that the pleadings were never formally amended” (citing Fed. Rules Civ. Proc. 15(b), 16)); Wilson v. Muckala, 303 F. 3d 1207, 1215 (CA10 2002) (“[C]laims, issues, defenses, or theories of damages not included in the pretrial order are waived even if they appeared in the complaint and, conversely, the inclusion of a claim in the pretrial order is deemed to amend any previous pleadings which did not include that claim”); Syrie v. Knoll Int’l, 748 F. 2d 304, 308 (CA5 1984) (“[I]ncorporation of a [new] claim into the pre-trial order … amends the previous pleadings to state [the new] claim”). In these circumstances, we look to the allegations as amended—here, the statement of claims in the final pretrial order—to determine original-source status.

The Government objects that this approach risks driving a wedge between the Government and relators. It worries that future relators might decline to “acquiesc[e]” in the Government’s tactical decision to narrow the claims in a case if that would eliminate jurisdiction with respect to the relator. Brief for United States 44. Even if this policy concern were valid, it would not induce us to determine jurisdiction on the basis of whether the relator is an original source of information underlying allegations that he no longer makes.

IV

Judged according to the principles set forth above, Stone’s knowledge falls short. The only false claims ultimately found by the jury (and hence the only ones to which our jurisdictional inquiry is pertinent to the outcome) involved false statements with respect to environmental, safety, and health compliance over a one-and-a-half-year period between April 1, 1987, and September 30, 1988. As described by Stone and the Government in the final pretrial order, the only pertinent problem with respect to this period of time for which Stone claimed to have direct and independent knowledge was insolid pondcrete. Because Stone was no longer employed by Rockwell at the time, he did not know that the pondcrete was insolid; he did not know that pondcrete storage was even subject to RCRA; he did not know that Rockwell would fail to remedy the defect; he did not know that the insolid pondcrete leaked while being stored onsite; and, of course, he did not know that Rockwell made false statements to the Government regarding pondcrete storage.

Stone’s prediction that the pondcrete would be insolid because of a flaw in the piping system does not qualify as “direct and independent knowledge” of the pondcrete defect. Of course a qui tam relator’s misunderstanding of why a concealed defect occurred would normally be immaterial as long as he knew the defect actually existed. But here Stone did not know that the pondcrete failed; he predicted it. Even if a prediction can qualify as direct and independent knowledge in some cases (a point we need not address), it assuredly does not do so when its premise of cause and effect is wrong. Stone’s prediction was a failed prediction, disproved by Stone’s own allegations. As Stone acknowledged, Rockwell was able to produce “concrete hard” pondcrete using the machinery Stone said was defective. According to respondents’ allegations in the final pretrial order, the insolidity problem was caused by a new foreman’s reduction of the cement-to-sludge ratio in the winter of 1986, long after Stone had left Rocky Flats.

Stone counters that his original-source status with respect to his spray-irrigation claim (which related to a time period different from that for his pondcrete claim, App. 492) provided jurisdiction with respect to all of his claims. We disagree. Section 3730(e)(4) does not permit jurisdiction in gross just because a relator is an original source with respect to some claim. We, along with every court to have addressed the question, conclude that §3730(e)(4) does not permit such claim smuggling. See United States ex rel. Merena v. SmithKline Beecham Corp., 205 F. 3d 97, 102 (CA3 2000); Hays v. Hoffman, 325 F. 3d 982, 990 (CA8 2003); Wang ex rel. United States v. FMC Corp., 975 F. 2d 1412, 1415–1416, 1420 (CA9 1992). As then-Judge Alito explained, “[t]he plaintiff’s decision to join all of his or her claims in a single lawsuit should not rescue claims that would have been doomed by section (e)(4) if they had been asserted in a separate action. And likewise, this joinder should not result in the dismissal of claims that would have otherwise survived.” SmithKline Beecham, supra, at 102.

Because Stone did not have direct and independent knowledge of the information upon which his allegations were based, we need not decide whether Stone met the second requirement of original-source status, that he have voluntarily provided the information to the Government before filing his action.

V

Respondents contend that even if Stone failed the original-source test as to his pondcrete allegations, the Government’s intervention in his case provided an independent basis of jurisdiction. Section 3730(e)(4)(A) permits jurisdiction over an action based on publicly disclosed allegations or transactions if the action is “brought by the Attorney General.” Respondents say that any inquiry into Stone’s original-source status with respect to amendments to the complaint was unnecessary because the Government had intervened, making this an “action brought by the Attorney General.”7 Even assuming that Stone was an original source of allegations in his initial complaint, we reject respondents’ “intervention” argument.

The False Claims Act contemplates two types of actions. First, under §3730(a), “[i]f the Attorney General finds that a person has violated or is violating section 3729, the Attorney General may bring a civil action under this section against the person.” Second, under §3730(b), “[a] person may bring an action for a violation of section 3729 for the person and for the United States Government.” When a private person brings an action under §3730(b), the Government may elect to “proceed with the action,” §3730(b)(4)(A), or it may “declin[e] to take over the action, in which case the person bringing the action shall have the right to conduct the action,” §3730(b)(4)(B). The statute thus draws a sharp distinction between actions brought by the Attorney General under §3730(a) and actions brought by a private person under §3730(b). An action brought by a private person does not become one brought by the Government just because the Government intervenes and elects to “proceed with the action.” Section 3730 elsewhere refers to the Government’s “proceed[ing] with an action brought by a person under subsection (b)”—which makes crystal clear the distinction between actions brought by the Government and actions brought by a relator where the Government intervenes but does not oust the relator.

Does this conclusion cast into doubt the courts’ jurisdiction with respect to the Government as well? After all, §3730(e)(4)(A) bars jurisdiction over any action brought under §3730, as this one was, unless the action is brought (1) by the Attorney General or (2) by an original source; and we have concluded that this is brought by neither. Not even petitioners have suggested the bizarre result that the Government’s judgment must be set aside. It is readily enough avoided, as common sense suggests it must be, by holding that an action originally brought by a private person, which the Attorney General has joined, becomes an action brought by the Attorney General once the private person has been determined to lack the jurisdictional prerequisites for suit. The outcome would be similar to that frequently produced in diversity-jurisdiction cases, where the “courts of appeals … have the authority to cure a jurisdictional defect by dismissing a dispensable nondiverse party.” Grupo Dataflux v. Atlas Global Group, L. P., 541 U. S. 567, 573 (2004) (citing Newman-Green, Inc. v. Alfonzo-Larrain, 490 U. S. 826, 837 (1989) ); see United States Steel Corp. v. EPA, 614 F. 2d 843, 845 (CA3 1979) (“[T]here are instances when an intervenor’s claim does not rise and fall with the claim of the original party”); 7C C. Wright, A. Miller, & M. Kane, Federal Practice and Procedure §1920, p. 491 (2d ed. 1986) (“[A]n intervenor can proceed to decision after a dismissal of the original action … if there are independent grounds for jurisdiction of the intervenor’s claim”). What is cured here, by the jurisdictional ruling regarding Stone’s claim, is the characterization of the action as one brought by an original source. The elimination of Stone leaves in place an action pursued only by the Attorney General, that can reasonably be regarded as being “brought” by him for purposes of §3730(e)(4)(A).

*  *  *

We hold that the District Court lacked jurisdiction to enter judgment in favor of Stone. We reverse the Tenth Circuit’s judgment to the contrary.

It is so ordered.

Justice Breyer took no part in the consideration or decision of this case.


Notes

1 Stone claimed the contrary, but the District Court found that he had failed to establish that fact.

2 Qui tam is short for “qui tam pro domino rege quam pro se ipso in hac parte sequitur,” which means “who pursues this action on our Lord the King’s behalf as well as his own.”

3 The laws and regulations allegedly violated included DOE Order Nos. 5480.2 (Dec. 1982), 5483.1 as superseded by 5483.1A (June 22, 1983), and 6430.1 (Dec. 12, 1983) (DOE General Design Criteria Manual); Colo. Rev. Stat. §25–5–501 et seq. (1982) (Hazardous Substances), 25–7–101 et seq. (1982 and Supp. 1988) (Air Quality Control Program), 25–7–501 et seq. (Asbestos Control), 25–15–101 (Hazardous Waste Management Act, 25–8–201 (1982) (Water Quality Control Act), 25–11–101 (1982 and Supp. 1988) (Radiation Control), 29–22–101 (Hazardous Substance Incidents), 25–5–503 (1982), 25–8–506, 25–8–608, 25–15–308 through 25–15–310, and 29–22–108; the Occupational Safety and Health Act of 1970, 29 U. S. C. §651 et seq.; the Atomic Energy Act of 1954, as amended, 42 U. S. C. §2011 et seq.; the Energy Reorganization Act of 1974, 42 U. S. C. §5801 et seq.; the Water Pollution Prevention and Control Act, 33 U. S. C. §1251 et seq.; the Clean Air Act, 42 U. S. C. §7401 et seq.; the Safe Drinking Water Act, 42 U. S. C. §300f et seq.; and regulations promulgated under these statutes.

4 In addition to the pondcrete allegations, respondents charged Rockwell with concealing problems with “saltcrete” (a mixture of cement and salt from liquid waste treatment processes) and “spray irrigation” (a method of disposing of waste water generated by the sewage treatment plant at Rocky Flats).

5 The Courts of Appeals have divided over the question. See United States ex rel. Laird v. Lockheed Martin Eng. & Science Servs. Co., 336 F. 3d 346, 353–355 (CA5 2003) (describing the Courts of Appeals’ divergent approaches). Only by demoting the actual text of §3730(e)(4) to a footnote and then paraphrasing the statute in a way that assumes his conclusion can Justice Stevens assert (without further analysis) that the statute’s meaning is “plain.” See post, at 1, 2 (dissenting opinion).

6 It is true that, when a defendant removes a case to federal court based on the presence of a federal claim, an amendment eliminating the original basis for federal jurisdiction generally does not defeat jurisdiction. See Carnegie-Mellon Univ. v. Cohill, 484 U. S. 343, 346, 357 (1988) ; St. Paul Mercury Indemnity Co. v. Red Cab Co., 303 U. S. 283, 293 (1938) . But removal cases raise forum-manipulation concerns that simply do not exist when it is the plaintiff who chooses a federal forum and then pleads away jurisdiction through amendment.

7 The Government includes a significant caveat: In its view, intervention does not cure any pre-existing defects in Stone’s initial complaint; it only cures defects resulting from amendments to the pleadings.