These consolidated appeals present questions of federal jurisdiction and constitutional law arising out of an attempt by the State of Illinois to prevent the shipment of spent nuclear fuel into the state for storage.

The fuel for nuclear electric-power reactors — enriched uranium — becomes depleted after a few years in the reactor and has to be replaced. Because the spent fuel is highly radioactive and its radioactivity very long-lived, the question what to do with it is a troublesome one. The spent fuel cannot just be thrown away, because wherever it is thrown will be contaminated. It used to be thought that the spent fuel would be reprocessed to make new fuel; but reprocessing of nuclear fuel has become intensely controversial and its prospects currently are uncertain. At the moment, then, spent nuclear fuel has to be stored indefinitely. But there is as yet no acceptable method of permanent, safe storage; and meanwhile the spent fuel is being stored for the most part in “swimming pools” at the reactor sites. However, some reactors may run out of on-site storage space by the middle of this decade.

The only away-from-site facility in the United States that is accepting spent nuclear fuel for storage is General Electric’s facility at Morris, Illinois, sixty miles southwest of Chicago. Spent fuel from reactors both within and outside Illinois is stored at the Morris facility. But in December 1980 Illinois enacted the Spent Fuel Act, Ill.Rev. Stat. ch. 111V4, § 230.1 et seq., which provides that “no person may dispose of, store, or accept any spent nuclear fuel which was used in any power generating facility located outside this State, or transport into this State for disposal or storage any spent nuclear fuel which was used in any power generating facility located outside this State . . . . ” § 230.22. The Act makes an exception for imports of spent nuclear fuel from states that have a storage facility like Illinois’ and have entered into a reciprocity agreement with Illinois, but it is a meaningless exception because, as mentioned, there is no other facility in the United States that is accepting spent nuclear fuel for storage away from the reactor site. The Act imposes a civil penalty of $10,000 per violation— $1,000 per day for a continuing violation— and also authorizes state officials to bring suits to enjoin violations. §§ 230.23, 230.24.

Three weeks after the passage of the Act, General Electric, joined by an out-of-state utility, Southern California Edison, which had shipped spent nuclear fuel to the Morris facility in the past and had a contract to ship additional amounts beginning early in 1981, brought suit in federal district court under 28 U.S.C. § 2201 (Declaratory Judgment Act) and 28 U.S.C. § 1331 (federal-question jurisdiction) against the state officials responsible for enforcing the Spent Fuel Act. The suit sought a declaration that the Act violates the supremacy and commerce clauses of the U.S. Constitution. Just a few hours after this suit (No. 81-2778 on our docket) was filed, the Attorney General of Illinois brought a suit (No. 81— 2768) in state court under the Illinois Spent Fuel Act and other Illinois statutes and the common law to enjoin Southern California Edison from shipping spent nuclear fuel into Illinois and General Electric from receiving it. The defendants removed the state’s suit to federal court, where it was consolidated with the federal suit, and moved for summary judgment in both suits. The motion was granted. The state has appealed, challenging the final judgment in both cases and also the district court’s refusal to remand No. 81-2768 to the state court.

The district court erred in not remanding No. 81-2768. The removal of a state court action to federal court is authorized only if the cause of action arises under federal law, see 28 U.S.C. § 1441, and the cause of action in No. 81-2768 did not; it arose under Illinois law. The fact that Illinois law may be unconstitutional as applied to the defendants’ activities is a matter of defense, not a ground for removal, even if the defendants show that the state laws sued on by the plaintiff are totally preempted by federal law. Illinois v. Kerr-McGee Chem. Corp., 677 F.2d 571 (7th Cir. 1982).

*209The state argues that there also is no federal jurisdiction over No. 81-2778. It says, first, that the companies failed to allege an actual controversy within the meaning of Article III of the Constitution and the Declaratory Judgment Act. See Skelly Oil Co. v. Phillips Petroleum Co., 339 U.S. 667, 671-74, 70 S.Ct. 876, 878-80, 94 L.Ed. 1194 (1950). When the companies’ suit was filed the state had not brought its suit, and it argues that until that happened no one could be certain that the Spent Fuel Act would actually affect the Morris facility.

It might appear to be a matter of indifference, or at most a technical defect readily curable by amending the complaint, whether there was an actual controversy when the federal suit was filed or not until a few hours later, when the state suit was filed. But it could make a difference, even a practical one. If there was no federal jurisdiction until the state suit was filed, the state’s argument, discussed later in this opinion, that even if the federal court had jurisdiction it should have abstained from exercising it might be strengthened. More important, the mere fact that the state sought injunctive relief against a threatened violation of the Act by the companies would not necessarily establish an actual controversy in the Article III sense; not being bound by Article III a state might allow a suit to be maintained in its courts though there was no actual controversy in that sense. Finally, technicality though it may be, there is considerable doubt, to say the least, whether a federal court can acquire jurisdiction retroactively. Mansfield, Coldwater & Lake Michigan Ry. v. Swan, 111 U.S. 379, 381-82, 4 S.Ct. 510, 511, 28 L.Ed. 462 (1884), holds that it cannot. If there was no jurisdiction when the complaint was filed the federal court had no power to act — even, perhaps, to the extent of allowing the complaint to be amended when a basis for federal jurisdiction came into being.

But we need not pursue these byways further in this case, because there was, we think, federal jurisdiction at the moment the federal complaint was filed, even though the filing of the state complaint came later. To constitute an actual controversy “the disagreement must not be nebulous or contingent but must have taken on fixed and final shape so that a court can see what legal issues it is deciding, what effects its decision will have on the adversaries, and some useful purpose to be achieved in deciding them.” Public Serv. Comm’n v. Wycoff Co., 344 U.S. 237, 244, 73 S.Ct. 236, 240, 97 L.Ed. 291 (1952). The disagreement between the state and the companies was not “nebulous or contingent” when the federal complaint was filed. The complaint alleged that Southern California Edison was about to ship spent nuclear fuel from California to the Morris facility. That shipment, and General Electric’s receipt of it, would have been in patent violation of the statute and would have made the companies liable to pay heavy civil penalties. Of course the companies could have gone ahead and violated the law and then have set up its alleged unconstitutionality as a defense in any civil penalty or other enforcement action brought by the state — in No. 81-2768, for example. But that is true with any statute alleged to be unconstitutional: you can violate it and then defend your conduct in state court on constitutional grounds. If it followed that there was no actual controversy until a violation occurred, the federal courts would never have jurisdiction to enjoin the threatened enforcement of state legislation alleged to violate the Constitution (except possibly at the behest of someone who had violated the statute but not yet been prosecuted for the violation) — a proposition that cannot be maintained in the face of Ex parte Young, 209 U.S. 123, 28 S.Ct. 441, 52 L.Ed. 714 (1908), and innumerable cases following it. See, e.g., Lake Carriers Ass’n v. McMullan, 406 U.S. 498, 506-08, 92 S.Ct. 1749, 1755-56, 32 L.Ed.2d 257 (1972); Steffel v. Thompson, 415 U.S. 452, 459, 94 S.Ct. 1209, 1215, 39 L.Ed.2d 505 (1974); Supreme Court of Va. v. Consumers Union of the U.S., 446 U.S. 719, 737,100 S.Ct. 1967, 1977, 64 L.Ed.2d 641 (1980).

It is true that where as in this case enforcement is merely threatened, a dispute *210may in fact lack sufficient concreteness and inevitability to be deemed an actual and not merely hypothetical controversy; it may be “unripe.” In limiting the jurisdiction of the federal courts to “cases or controversies” Article III has been understood to express a policy, one that makes especially good sense in regard to constitutional adjudication, against rendering judgments that are either unnecessary to resolve a real dispute or unlikely to be made in a competent fashion. When a suit is brought against a state before the state has actually done anything to the plaintiff, not only may the suit not be necessary to protect the plaintiff’s interests — because the state may never in fact do anything to him — but the lack of an immediate conflict may deprive the court of the facts it needs to make an informed decision; it may also — a related point — reduce the parties’ incentives to make a real contest of the case. See, e.g., Regional Rail Reorganization Cases, 419 U.S. 102, 143-45, 95 S.Ct. 335, 358-59, 42 L.Ed.2d 320 (1974). So if it were uncertain whether the Spent Fuel Act really was applicable to the Morris facility, or to Southern California Edison’s shipment of spent nuclear fuel to the facility, or whether Southern California Edison was really planning to ship spent nuclear fuel to Morris in the near future, or whether the state would actually enforce the statute, we might well conclude that there was no actual controversy between the state and the companies when they brought the suit.

The case would then be like Wycoff, supra, see 344 U.S. at 244, 73 S.Ct. at 240, or our circuit’s recent decision in Nuclear Engineering Co. v. Scott, 660 F.2d 241, 252-53 (7th Cir. 1981). The Attorney General of Illinois had announced at a press conference that he intended to bring lawsuits against violators of Illinois’ environmental protection laws and one of those violators, he said, was Nuclear Engineering Company, which operated a facility for storing hazardous wastes. A few days after the press conference the company sued the state in federal court, alleging that the Illinois laws in question could not constitutionally be applied to its facility. This court held that the threat of enforcement did not create an actual controversy between the state and the company. Nuclear Engineering was not challenging the validity of the state laws that the Attorney General was seeking to enforce against it. On the contrary, it claimed to be in full compliance with them, and if so it had nothing to fear from the carrying out of the Attorney General’s threat — it was not a present or prospective violator. If the threat itself had had an adverse impact on Nuclear Engineering’s business this might have created an actual controversy between it and the state, but the court’s “careful examination of the record ... reveal[ed] not a shred of evidence in support of NEC’s allegation of immediate business injury.” Id. at 252. The company was asking the federal court for an advisory opinion that if the Illinois laws were construed — incorrectly in the company’s view — in such a way as to force it to close its hazardous waste facility, the state would be violating the Constitution.

The present case is different. Not only is it certain that General Electric and Southern California Edison would be violating the Spent Fuel Act if Southern California Edison shipped spent nuclear fuel to Morris, as it intended (but for the state law) to do; but as the Act has only one conceivable target — the Morris facility — it is extremely unlikely that the state would overlook the violation. Adverse legal action was a distinct possibility when Nuclear Engineering filed its suit but it was a virtual certainty when General Electric and Southern California Edison filed theirs. While the difference is one of degree rather than of kind, it is enough we think to show that the controversy between the state and the companies in this case has real and immediate and not merely hypothetical or remote consequences for them.

The state makes the related jurisdictional argument that the companies’ cause of action did not arise under federal law, as required for federal jurisdiction under 28 U.S.C. § 1331, the statutory basis of federal jurisdiction in this case. (The complaint also alleges diversity of citizenship, treating the named defendant, the Attorney General *211of Illinois, as a citizen of Illinois within the meaning of 28 U.S.C. § 1332; but the companies reserve rather than argue this basis for federal jurisdiction on this appeal, in view of its rejection in Nuclear Engineering, supra, 660 F.2d at 250-51.)

In Nuclear Engineering, after the company brought its suit the Attorney General of Illinois carried out his threat and sued the company. That created, in this court’s view, an actual controversy and the court proceeded to consider whether there was a statutory basis for federal jurisdiction over it. Purists might consider this part of Nuclear Engineering an advisory opinion (the court having held it had no jurisdiction), but however that may be its teaching is inapplicable to the present case. Since the only thing that created an actual controversy in Nuclear Engineering was the filing of the state’s lawsuit, it seemed to this court that any federal suit which Nuclear Engineering might subsequently have filed would just have been an attempt to raise in federal court its federal defenses to the state court action. See 660 F.2d at 253-54. The company could not argue that its rights were being infringed by a state law that was inconsistent with one or more provisions of the Constitution, because it did not contend that the state law was invalid, and in any event it claimed to be complying with that law. All it could plead in its federal complaint was that if the state court accepted the Attorney General’s proposed construction of state law, that law would violate its constitutional rights. The conditional assertion of a federal right is in reality the assertion of a federal defense. If this could be the basis of a federal suit, the rule mentioned earlier that forbids the removal of a state suit to federal court on the basis of a federal defense to that suit would be circumvented. But in the present case the companies are claiming an unconditional right under the Constitution to ship spent nuclear fuel into, and to receive it in, Illinois for storage notwithstanding the Spent Fuel Act. They can vindicate that right in a suit under section 1331. The commerce and supremacy clauses of the Constitution create rights enforceable in equity proceedings in federal court. See, e.g., Lake Carriers’ Ass’n, supra, 406 U.S. at 502-03, 92 S.Ct. at 1753; Jones v. Rath Packing Co., 430 U.S. 519, 97 S.Ct. 1305, 51 L.Ed.2d 604 (1977).

There is, it is true, a dictum in the Wycoff case, supra, which if it is the law could put the plaintiffs in this case right out of court: “Where the complaint in an action for declaratory judgment seeks in essence to assert a defense to an impending or threatened state court action, it is the character of the threatened action, and not of the defense, which will determine whether there is federal-question jurisdiction in the District Court.” 344 U.S. at 248, 73 S.Ct. at 242. The impending state court action here was based on an alleged violation of state rather than federal law. Now at one level the dictum is an innocuous if perhaps over-broad statement of the principle that you cannot get around the limitations on the removal jurisdiction of the federal courts by asserting your defenses in a federal declaratory judgment action. But at another level, if understood to require federal claimants always to litigate their claims as defenses in state court if they can, it must be wrong, and though lower federal courts have followed it from time to time the Supreme Court has not. For a perceptive discussion see Note, Federal Jurisdiction over Declaratory Suits Challenging State Action, 79 Colum.L.Rev. 983 (1979). Since the impending state action will almost always be based on state law alone, the dictum, read broadly, would overrule Ex parte Young and every case that has ever followed it. If not wrong, such a reading would still be an inappropriate flight of fancy for an inferior federal court to take.

Still, the companies here could, and did, raise their constitutional claims in defense of the state court action brought by the Attorney General; and the state argues, as an alternative ground for dismissal of the federal suit, that the district court, even if it had jurisdiction, should have abstained from exercising it and forced the companies to litigate exclusively in the *212state courts. There must be something wrong with this argument, because if it were accepted it would prevent people from ever challenging . the constitutionality of state legislation in federal court. Every time someone brought a suit to have a state law declared unconstitutional the state would immediately file an action in state court to enjoin violation of the statute and then move the federal district court to stay or dismiss the federal suit, thereby forcing the federal plaintiff to litigate his federal claims by way of defense in the state court action. If to avoid being forced into state court in this way the federal claimant violated the statute and then brought a federal suit to enjoin its enforcement, his suit would almost certainly be dismissed under the doctrine of Younger v. Harris, 401 U.S. 37, 91 S.Ct. 746, 27 L.Ed.2d 669 (1971)—that federal courts ordinarily will not interfere in pending state prosecutions — provided only that the state brought some sort of enforcement proceeding before the federal suit was well advanced, see Hicks v. Miranda, 422 U.S. 322, 349, 95 S.Ct. 2281, 2291, 45 L.Ed.2d 223 (1975). For the term “prosecution” has been broadened to include civil enforcement actions, clearly including the type of civil penalty suit that the state could have brought under the Spent Fuel Act if General Electric and Southern California Edison had gone ahead and violated the Act. See, e.g., Trainor v. Hernandez, 431 U.S. 434, 443-44, 97 S.Ct. 1911, 1917-18, 52 L.Ed.2d 486 (1977); Middlesex Cty. Ethics Comm. v. Garden State Bar Ass'n, - U.S. -, 102 S.Ct. 2515, 73 L.Ed.2d 116 (1982).

The state is unable to cite any precedent for the contraction of federal jurisdiction that it is advocating. Hicks v. Miranda, supra, on which the state relies heavily, holds only “that where state criminal proceedings are begun against the federal plaintiffs after the federal complaint is filed but before any proceedings of substance on the merits have taken place in the federal court, the principles of [Younger] should apply in full force.” The plaintiffs in Hicks had already violated the statute; that was what exposed them to state criminal prosecution. Younger holds that the state’s interest in enforcing its laws without interference from a federal court takes precedence over the federal claimant’s interest in litigating his claim in a federal forum. But the federal plaintiffs in the present case did not violate state law, so the state could not have brought a penalty action against them. The state’s interest in maintaining law and order through enforcement of its penal statutes was not impaired by the federal lawsuit and the principle of the Younger case is therefore inapplicable. It is not the order of the suits that is important to the Younger doctrine but whether a state statute has been violated and the violator is being sued in state court.

The state also cites Louisville Area Inter-Faith Comm. for United Farm Workers v. Nottingham Liquors, Ltd., 542 F.2d 652 (6th Cir. 1976). The defendant had obtained a temporary restraining order in state court against picketing by the plaintiff union. Rather than ask the state court to dissolve the order the union brought a federal suit to enjoin its enforcement as a violation of the union’s First Amendment rights. The federal district court dismissed the suit and the court of appeals affirmed. The court of appeals cited the Younger case — which is a bit surprising, since the state was not a party to the state court suit that the plaintiff in Nottingham was trying to enjoin— but did not hold that federal courts as a general rule will not exercise jurisdiction in a case where the plaintiff could raise his federal claim in a pending state court proceeding. The state judge who had issued the temporary restraining order was the logical person to consider, in the first instance at least, the union’s claim that the order should be dissolved; judicial economy therefore required that the federal court stay its hand until the union applied to the state judge to dissolve his order. There is nothing like that in the present case. No state suit had even been filed when the federal suit was filed, let alone had proceeded to the point where the exercise of federal jurisdiction would have interrupted an active state court litigation.

*213In essence the state is asking us to adopt a rule of abstention that whenever parallel state and federal suits are pending, the federal suit must be stayed and the federal claimant remitted to state court. The cases do not support such a rule, and it would be hard to square with 28 U.S.C. § 1331(a), which explicitly grants the federal courts jurisdiction over civil suits arising under the Constitution. In every case that the state has cited the federal claimant had gone ahead and violated state law, thereby bringing into play the state’s interest in redressing violations of its laws in its own courts. The message of Younger and of the recent cases that have expanded its principle to civil proceedings is that one who decides to violate a state law that he believes to be unconstitutional may find that he has thereby submitted himself to the jurisdiction of the state courts. Only if he goes to court before violating the law can he be assured of a federal forum, and then only if he satisfies the stringent jurisdictional requirements, discussed earlier in this opinion, for bringing a federal suit to enjoin merely threatened state enforcement action. If the state’s position were correct, virtually no federal suit could ever be brought to challenge the constitutionality of a state statute, at least if the statute was enforceable by the state itself. The state, upon learning that a federal suit had been filed, would have only to bring its own declaratory or injunctive suit to force the federal court to abstain from exercising jurisdiction. While there are arguments in favor of this result, for among other things it is unseemly to allow a single federal district judge to enjoin a state statute, it would be too radical a break with precedent to be decreed by a lower court.

The principle of Younger is of course only one part of the abstention doctrine. Nottingham illustrates another. When the constitutionality of a state statute depends on how the statute is construed, there is a third ground for abstention; but in the present case, as in Zwickler v. Koota, 389 U.S. 241, 249, 88 S.Ct. 391, 396, 19 L.Ed.2d 444 (1967), the statute is unambiguous and — as we are about to see — no saving construction is possible.

We come at long last to the merits. General Electric and Southern California Edison challenge the constitutionality of the Spent Fuel Act on two grounds. The first is that it violates the commerce clause because it discriminates against interstate commerce. There is no dispute that the Spent Fuel Act applies only to spent fuel brought in from other states and that the Morris facility stores spent nuclear fuel from nuclear reactors located in Illinois as well as from out-of-state utilities such as Southern California Edison; in fact about half the spent fuel currently being stored comes from reactors in Illinois. However compelling the state’s interest in safeguarding its residents from the hazards of radioactivity, that interest is unaffected by the origin of the radioactive material. Even if it is the movement rather than the storage of the spent nuclear fuel that is the source of the hazards about which the state is concerned, all spent nuclear fuel stored in the Morris facility was brought there from somewhere else yet the statute does not limit or forbid any intrastate movements, however lengthy or hazardous, nor even interstate movements so long as they do not end up at the Morris facility.

To pass laws that arbitrarily burden interstate commerce, by forbidding shipments merely because they originate out of state, violates the commerce clause, see, e.g., Hughes v. Oklahoma, 441 U.S. 322, 336-37, 99 S.Ct. 1727, 1736, 60 L.Ed.2d 250 (1979), and it is irrelevant that the traffic is in “bads” rather than goods. The efficient disposal of wastes is as much a part of economic activity as the production that yields the wastes as a byproduct, and to impede the interstate movement of those wastes is as inconsistent with the efficient allocation of resources as to impede the interstate movement of the product that yields them. If Illinois may not arbitrarily prevent the importation of electricity generated by nuclear reactors in other states, or of nuclear fuel to power its own reactors, no more may it arbitrarily prevent the waste products of nuclear electricity gener*214ation from being imported for storage in the state. Nuclear wastes have to be stored somewhere, and the place of storage should be chosen without regard to the parochial interests of the states.

Yet contrary to what we have just said some early cases, the “quarantine cases” on which the state heavily relies, treat interstate commerce in “bads” as not commerce at all, see, e.g., Bowman v. Chicago & Nw. Ry., 125 U.S. 465, 489, 8 S.Ct. 689, 700, 31 L.Ed. 700 (1888), and thus hold that a state has a right to ban the importation of “convicts, paupers, idiots, and lunatics, and persons likely to become a public charge, as well as persons afflicted by contagious or infectious diseases; a right founded ... in the sacred law of self-defense.” Hannibal & St. Joseph R. R. v. Husen, 95 U.S. 465, 471, 24 L.Ed. 527 (1878). See also Asbell v. Kansas, 209 U.S. 251, 28 S.Ct. 485, 52 L.Ed. 778 (1908) (diseased cattle). Shapiro v. Thompson, 394 U.S. 618, 89 S.Ct. 1322, 22 L.Ed.2d 600 (1969), which forbids the states to deny welfare benefits to new residents, casts a big shadow over Husen ; but Shapiro is a Fourteenth Amendment “right to travel” case, not a commerce clause case (though there are glancing references in the opinions to a right to travel based on that clause, see 394 U.S. at 630 n.8, 666, 89 S.Ct. at 1329 n.8, 1348), and in any event does not apply to a nonhuman import, such as spent nuclear fuel.

The quarantine cases are helpful to the State of Illinois because the Supreme Court did not ask whether the state was treating intrastate “bads” the same way it was treating interstate ones. But maybe it did not ask because it thought the answer obvious. A state that bans the importation of paupers or diseased cattle probably does not do so out of affection for its local paupers and diseased cattle. The hostility is to the thing itself, not to merely interstate shipments of the thing; and an undiscriminating hostility is at least nondiscriminatory. But that is not the case here. The State of Illinois is quite willing to allow the storage and even the shipment for storage of spent nuclear fuel in Illinois, provided only that its origin is intrastate.

This case is governed not by the quarantine cases but by City of Philadelphia v. New Jersey, 437 U.S. 617, 98 S.Ct. 2531, 57 L.Ed.2d 475 (1978), which held that New Jersey could not confine the use of its landfill waste dumps to New Jersey residents. The Court did distinguish the quarantine cases by stating that New Jersey’s objection was not to the shipment but to the storage of wastes, see id. at 628-29, 98 S.Ct. at 2537-38, and the State of Illinois seizes on this distinction and argues that it does object to the shipment as well as storage of spent nuclear fuel. Its objection is unconvincing because it allows intrastate shipments, and even interstate shipments so long as they are not for the purpose of storage in Illinois. But a more satisfactory basis of distinction is, as we have suggested, that the quarantine cases involve state efforts to limit or prevent “bads” regardless of whether they come from in or outside the state. That evenhandedness is missing in this case as in City of Philadelphia. and it is not supplied by the hypothetical possibility, unexplored on this record, that the state might try to use its environmental protection laws to deal with the intrastate shipment or storage of spent nuclear fuel.

Though it may seem superfluous to consider whether the Spent Fuel Act is also unconstitutional under the supremacy clause, we shall do so in order to assist the Supreme Court should it decide to review this case and in doing so disagree with our analysis of the commerce clause issue. Such a discussion may also assist the Court in deciding whether to review the case at all, in the event that review is sought and the Court agrees with us that federal jurisdiction exists and was properly exercised.

The Atomic Energy Act, 42 U.S.C. §§ 2011 et seq., sets up a comprehensive scheme of federal regulation of atomic energy, administered by the Nuclear Regulatory Commission. The Act does not refer explicitly to spent nuclear fuel, but it does refer to the constituents of that fuel, see 42 U.S.C. §§ 2014(e), (z), (aa); 10 C.F.R. § 72.3(v); and the state does not, and could *215not, see 42 U.S.C. §§ 2073, 2111, question the Commission’s authority to regulate the storage of spent nuclear fuel. Pursuant to that authority the Commission recently issued detailed regulations for the licensing of storage facilities — both at the reactor site and, as in the case of the Morris facility, away from the reactor site — for spent nuclear fuel. See 10 C.F.R. Part 72 (1981). These regulations forbid the acquisition, receipt, or possession of spent nuclear fuel for storage except under license from the Commission; and again its authority to require licensing is unquestioned. The Morris facility was licensed under an earlier set of regulations, 10 C.F.R. Part 70 (1980), and its license is up for renewal under the new regulations (Part 72, supra), see 45 Fed. Reg. 74698-99 (1980), but the validity of its license is not challenged.

The regulations, however, say nothing about shipment across state lines; and, curiously, there is no copy of the license for the Morris facility in the record, so that we do not know whether it expressly authorizes the receipt of spent nuclear fuel from other states. Although no one suggests that General Electric may be violating the license by receiving interstate shipments of spent nuclear fuel for storage at the Morris facility, it appears that neither the Act, the regulations under the Act, nor the license itself expressly authorizes such receipt — -though there are some very broad hints in the regulations that it is authorized. See 10 C.F.R. §§ 70.42(b)(5), 71.5. But this hardly matters. Northern States Power Co. v. Minnesota, 447 F.2d 1143 (8th Cir. 1971), aff’d without opinion, 405 U.S. 1035, 92 S.Ct. 1307, 31 L.Ed.2d 576 (1972), holds that the Atomic Energy Act, in the words of a subsequent Supreme Court decision, “created a pervasive regulatory scheme, vesting exclusive authority to regulate the discharge of radioactive effluents from nuclear power plants in the [NRC], and preempting the States from regulating such discharges.” Train v. Colorado Public Interest Research Group, 426 U.S. 1, 16, 96 S.Ct. 1938, 1945, 48 L.Ed.2d 434 (1976). The analysis of the structure and legislative history of the Act in Northern States compels the conclusion that the Act equally preempts state regulation of the storage, and shipment for storage, interstate and intrastate alike, of spent nuclear fuel.

We do not think the State of Illinois would quarrel with anything we have said thus far on the issue of preemption. Its argument rather is that Congress, by passing and later amending section 116 of the Clear Air Act, 42 U.S.C. § 7416, has authorized the state to nullify the Commission’s exercise of authority, presumably even to the extent of shutting down the' Morris facility completely. Section 116 provides that “nothing in this Act shall preclude or deny the right of any State ... to adopt or enforce . . . any requirement respecting control or abatement of air pollution.” And “air pollution” is defined in the Clean Air Act Amendments of 1977 (amending section 302(g) of the Clean Air Act, 42 U.S.C. § 7602(g)) as including any “radioactive (including source material, special nuclear material, and byproduct material) substance or matter which is emitted into or otherwise enters the ambient air.” Since source material, special nuclear material, and byproduct material are the principal constituents of spent nuclear fuel, section 116 read in light of the amended section 302(g) appears to authorize the states to regulate air pollution caused by facilities such as that at Morris. But it does not, we believe, authorize the Spent Fuel Act.

Section 116 of the Clean Air Act qualifies federal powers only under that act. It does not purport to limit federal powers under other statutes, such as the Atomic Energy Act, under which the Morris facility was licensed. At least that is what section 116 says, although an argument can be made from the legislative history that Congress intended to allow the states to adopt stricter standards than those imposed by the NRC under the Atomic Energy Act. See H.Rep.No. 294, 95th Cong., 1st Sess. 43 and n.8 (1977), U.S.Code Cong. & Admin.News 1977, p. 1077; Consolidated Edison Co. of N.Y., Inc., 7 N.R.C. 31, 34 and n.13 (1978); Pacific Legal Foundation v. State Energy Resources Conservation & Development *216Comm’n, 659 F.2d 903, 927 (9th Cir. 1981). But even so, for this expansive view of section 116 to be applicable to the Spent Fuel Act there would have to be some basis for regarding the Act as a measure for controlling (radioactive) air pollution; and we cannot find any. The Act places no limitation on the level of radioactive emissions from the Morris facility or from shipments to the facility. And while one effect of the Act may be to reduce those emissions by eliminating interstate shipments to the facility, that is not enough to make the Act a rational pollution-control measure, especially when, for aught that appears, the facility emits no radioactivity into the air. We cannot believe that Congress in promulgating the Clean Air Act Amendments meant the states to have carte blanche to enact any statutes, and promulgate any regulations, that might as a side-effect reduce the level of radioactive emissions in the state, regardless of how much the statute or regulation disrupted the federal atomic energy program, which includes as we have noted a program for disposing of nuclear wastes. That would be too much to read into those amendments.

The district court’s judgment in No. 81-2768 and its order denying the state’s motion to remand that case to state court are reversed, and the case is remanded to the district court with instructions to remand it to the state court where it began. The judgment in No. 81-2778 is affirmed. The parties shall bear their own costs in this court.

So Ordered.