dissenting:
PROLOGUE
This is a bizarre case: Every reasonable consideration furnishing a basis for judgment favors the appellant; yet, the majority cannot bring itself to reach the result that is compelled by the record before it.
We reject the decision of the District Court, for it fails to comport with the terms of the statute_ We also recognize that the FDA — the agency to which we owe deference if this case is analyzed under “step two” of Chevron U.S.A., Inc. v. NRDC, 467 U.S. 837, 104 S.Ct. 2778, 81 L.Ed.2d 694 (1984) — does not subscribe to the rationale underlying the District Court’s decision.... We also reject the FDA’s interpretation, because it, too, is at odds with the language of the statute and does not otherwise reflect a permissible reading of the disputed legislation.... We agree with the District Court that Abbott’s interpretation is supported by the plain language of the statute, although the majority finds this result unpalatable_ And, finally, our decision today is in direct conflict with a judgment of a sister circuit. See Glaxo Operations UK Ltd. v. Quigg, 894 F.2d 392, 395 (Fed.Cir.1990).
Despite all of the foregoing, the majority has decided to remand this case to the agency to give it another shot at the issue. This makes no sense to me.
In 1978, the Food and Drug Administration (“FDA”) approved a new anticonvul-sant drug, called “Depakene,” that had been developed by the appellant, Abbott Laboratories (“Abbott”). Depakene was effective in controlling seizures but caused harmful gastrointestinal side-effects in some patients. To avoid these side-effects, Abbott subsequently developed a new drug, called “Depakote,” which was also effective in controlling seizures but which used a different “active ingredient” that was less prone to irritating patients’ digestive tracts. Thus, although the two drugs bear chemical similarities, Depakote represents a significant improvement over the earlier drug.
Abbott filed a new drug application for approval to market Depakote in December of 1981. The FDA approved the new drug on March 10, 1983. Thereafter, on September 24, 1984, Congress passed the Drug Price Competition and Patent Term Restoration Act, Pub.L. No. 98-417, 98 Stat. 1585 (1984) (the “Act” or “the Hatch-Waxman Amendments”), which was intended to allow a manufacturer to obtain approval to market a drug without performing clinical tests normally required to support a full new drug application and, also, to provide market exclusivity for innovative drugs approved by the FDA. Under Title I of the Act, there are five categories of drugs eligible for varying periods of market exclusivity ranging from two years to 10 years. Relying on 21 U.S.C. § 355(j)(4)(D)(i) (hereinafter “subsection (i)”), which applies to drugs approved by the FDA between January 1, 1982, and September 24, 1984, Abbott sought 10 years of market exclusivity for Depakote.
Subsection (i) applies to “a drug, no active ingredient (including any ester or salt of the active ingredient) of which has been approved in any other application”; and Abbott claimed that the prior approval of Depakene did not bar 10-year exclusivity for Depakote, because the “active ingredient” for the two drugs is different and because the latter drug’s active ingredient (including any ester or salt of it) had never before been approved by the FDA. The FDA rejected this claim, applying instead the two-year market exclusivity protection of 21 U.S.C. § 355(j)(4)(D)(v) (hereinafter “subsection (v)”), because Depakene and Depakote “shared the same active moiety.” Brief for Appellee at 9. In other words,
*991the FDA ignored the literal terms of the statute (focusing on “active moiety” in place of “active ingredient”), because, in the agency’s view, Congress intended “to provide the maximum period of market protection only to the most innovative drugs which require the largest expenditure of resources to develop.” Id. at 14. The agency concedes that “active ingredient” and “active moiety” are not the same; and the agency can point to no statutory provision supporting the gloss that it has placed on subsection (i).
The majority, while shunning the FDA’s statutory construction, agrees with the conclusion that Depakote should obtain only the shorter, two-year period of market protection. In reaching this judgment, the majority, like the FDA, ignores the plain meaning of the statute because it believes that, if enforced as written, the statute would produce arbitrary results. I dissent because I believe the majority’s conclusion is based on largely groundless conjecture, and because it is our task as judges to apply the statute as it is written, not as we think it ought to have been written.
I.
As the majority correctly notes, the Hatch-Waxman Amendments were the product of compromise. The Act emerged from Congress’ efforts to balance two conflicting policy objectives: to induce name-brand pharmaceutical firms to make the investments necessary to research and develop new drug products, while simultaneously enabling competitors to bring cheaper, generic copies of those drugs to market. See Mead Johnson Pharmaceutical Group v. Bowen, 838 F.2d 1332, 1333 (D.C.Cir.1988); H.R.Rep. No. 857, 98th Cong., 2d Sess., pt. 1, at 14-15, reprinted in 1984 U.S.Code Cong. & Admin.News 2647, 2647-48. The statutory provisions directly at issue here reflect one of the incentives Congress held out to name-brand pharmaceutical concerns in the trade-off, namely, market exclusivity for the pioneer drugs they bring to market. During the period of market exclusivity, the FDA will withhold approval of any generic copy.
In assessing the applicability of subsection (i) (the 10-year provision) versus subsection (v) (the two-year provision),1 the critical question is whether the new drug for which market exclusivity is sought contains an “active ingredient (including any ester or salt of the active ingredient)” that has been previously approved by the FDA. See 21 U.S.C. §§ 355(j)(4)(D)(i) & (v) (1988), If the new drug’s “active ingredient (including any ester or salt of the active ingredient)” has been previously approved by the FDA, then the new drug is entitled only to two years’ protection under subsection (v); if it has not, then the drug may claim the full 10-year period of market exclusivity under subsection (i), the rationale being that the drug likely represents a more significant innovation.
All parties agree that Depakote, the drug for which Abbott seeks market exclusivity, bears many similarities to Abbott’s previously approved Depakene. Both drugs are prescribed to control seizures and both do so by way of a common “active moiety,” i.e., the pharmacologically significant agent that actually accomplishes the intended therapeutic effect in the human body. The “active moiety” for both Depakote and De-pakene is a substance known as valproic acid. The chief innovative development in Depakote is a modification of Depakene’s initial dosage form, i.e., its “active ingredient,” that avoids undesirable side-effects associated with Depakene. Depakene uses valproic acid as its “active ingredient” — in effect introducing the “active moiety” directly into the body. This approach was effective in controlling seizures but resulted in gastrointestinal side-effects, leading *992Abbott to develop Depakote. Depakote uses a different substance, known as dival-proex sodium, as its active ingredient. Divalproex sodium avoids the side-effects associated with the direct ingestion of val-proic acid. Once divalproex sodium is ingested, however, the body converts it into valproic acid, which then accomplishes the intended effect of controlling seizures.
Although valproic acid was previously approved in Abbott’s earlier application for Depakene, it is conceded by all parties that neither divalproex sodium nor any salt or ester of divalproex sodium had ever been previously approved by the FDA.
Abbott sought the maximum, 10-year period of market exclusivity under the statute based on the fact that neither the active ingredient of its new drug (divalproex sodium) nor any salt or ester of that active ingredient had ever before won FDA approval. The FDA concluded, however, that Depakote was entitled only to two years’ market exclusivity. It justified this conclusion by a rather far-fetched reading of the statute: by using the word “including” in the statute, the agency argued, Congress meant to introduce an illustrative list of the sorts of chemical derivations of an active ingredient that, if previously approved, would deprive a new drug of the maximum, 10-year protection. The FDA was thus entitled, the argument goes, to insert the term “active moiety” into the statutory scheme in order to effectuate the alleged broader intentions of Congress to distinguish between drugs that represented entirely new chemical entities and those that were derivations of previously approved entities. By this reading, Depakote would be entitled to only two years’ protection because its active moiety (valproic acid) had been previously approved in Depakene.
II.
I agree with the majority that the FDA’s reading of the statute, which turns entirely on a purported ambiguity or implied delegation in the word “including,” is implausible and must be rejected. I also agree with the majority that the reading offered by the District Court, which concluded that there is an irreconcilable conflict between the two obviously complementary statutory provisions, must also be rejected. I disagree, however, with the majority’s discovery of a new purported ambiguity in the statute, a discovery it uses to justify a remand to the FDA (along with a strong hint about how the agency should construe subsections (i) and (v)). Rather, I believe this case is properly resolved at the first step of the familiar Chevron doctrine, under which the court must adhere to the clear intention of Congress expressed on the face of the statute. See Chevron U.S.A., Inc. v. NRDC, 467 U.S. 837, 842-43, 104 S.Ct. 2778, 2781-82, 81 L.Ed.2d 694 (1984).
The parties are in agreement that the terms used by the statute — active ingredient, salt and ester — have plain and established meanings in scientific and regulatory parlance. See Glaxo Operations UK Ltd. v. Quigg, 894 F.2d 392, 395 (Fed.Cir.1990). The only ambiguity asserted by the FDA pertains to the statute’s use of the word “including,” a claim this court properly dismisses. Thus, by the plain terms of the statute, Abbott’s Depakote is entitled to 10 years of market exclusivity because the FDA had never previously approved Depa-kote’s active ingredient (divalproex sodium) or any salt or ester of that ingredient.
The majority rejects the literal terms of the Act, not because the statutory language is actually ambiguous, but rather because it fears that applying the statute as written might produce arbitrary results. Although the majority acknowledges that Abbott’s interpretation of the statute is the more obvious one, it rejects this view as whimsical because of the supposed ability of a drug manufacturer to manipulate the length of market exclusivity it obtains under the statute by altering the sequence in which it markets certain drugs. For example, it is conceded that if Abbott had invented and won FDA approval for Depakote before it invented Depakene, the latter drug would be entitled only to two years’ protection. This is because divalproex sodium, arguably a salt of Depakene’s active ingredient, valproic acid, would then have *993been previously approved by the FDA, pushing the second drug into the compass of subsection (v). By reversing the order in which Government approval is sought, however, the manufacturer could obtain 10 years’ protection for both drugs under Abbott’s “plain language” reading of the statute. Such a result could not have been within Congress’ intention, the majority concludes, and the statute’s plain meaning must therefore be ignored.
There are at least two answers to the majority’s argument. First, we have no basis for knowing whether it would be scientifically plausible to reverse the order of pharmaceutical invention in a case of this sort, or economically feasible to reverse the order in which Government approval is sought. Intuition suggests that a “derivative” drug is likely to be invented after the “parent” drug, rather than the other way around. Indeed, it would seem capricious, even absurd, for a pharmaceutical company to develop and attempt to market a drug (like Depakene) with harmful side-effects subsequent to the development of a better drug (like Depakote) that avoided those side-effects. And in this case there is clearly no danger that Abbott orchestrated its new drug applications to exploit the statute given that the statutory provisions at issue here apply only retrospectively to conduct that occurred before the statute was enacted.
Furthermore, and more to the point, the parties offered no evidence either to support or refute the majority’s hypothesis. There is simply no basis in the record for the majority’s blithe assumption that it is entirely serendipitous that Abbott invented Depakene before Depakote. I am loath to ignore the plain words of Congress based upon a counterintuitive supposition that is apparently groundless.
Second, even if the sequence of invention and approval could be manipulated to the manufacturer’s advantage, there is nothing so outrageous about such a result to justify the majority’s refusal to enforce the statute as Congress wrote it. While it is true that “a court should go beyond the literal language of a statute if reliance on that language would defeat the plain purpose of the statute,” Bob Jones Univ. v. United States, 461 U.S. 574, 586, 103 S.Ct. 2017, 2025, 76 L.Ed.2d 157 (1983), the Supreme Court ordinarily applies this canon only where it is clear that adherence to a statute’s literal terms would produce a result flagrantly at odds with legislative purpose.2 The result that would ensue from literal application of the market exclusivity provisions — granting the maximum economic reward to a drug that had to endure less than a maximum amount of regulatory delay and expense — does not seem likely to “defeat the plain purpose” of the statute, which was, after all, to strike a rough compromise between accelerating generic entry into the market and providing sufficient economic incentives for continued name-brand innovation.
*994That a statute yields a somewhat anomalous result does not mean that Congress could not have intended it. Many statutes, being products of innumerable and sometimes hasty and pragmatic compromises, permit seemingly arbitrary results. See Board of Governors of the Federal Reserve Sys. v. Dimension Fin. Corp., 474 U.S. 361, 374, 106 S.Ct. 681, 689, 88 L.Ed.2d 691 (1986). That fact is not cause, however, for rewriting them either in the agencies or in the courts. See id. Contrary to the tenor of the majority’s analysis, Congress is not subject to the “arbitrary and capricious” constraints of the Administrative Procedure Act when it drafts and enacts statutes; and courts are not free to withhold enforcement of congressional enactments merely because they seem arbitrary or anomalous. Cf. Motor Vehicle Mfrs. Ass’n v. State Farm Mut. Auto. Ins. Co., 463 U.S. 29, 43 n. 9, 103 S.Ct. 2856, 2866 n. 9, 77 L.Ed.2d 443 (1983) (presumption of rationality accorded a statute is much greater than presumption of regularity accorded an agency action).
The majority labors mightily to discover a possible ambiguity in the statute in order to justify its refusal to adhere to the plain meaning of Congress’ words. It concedes that Abbott’s is the more obvious construction of the actual language used in the statute, but finds that reading unpalatable because of its suspicion that “Congress did not focus on the reverse sequence” problem. Judicial tinkering with statutes in order to correct suspected congressional oversights is, however, a dangerous business:
In these rules [of statutory interpretation], we courts assume that Congress always knows the particulars whereof it speaks; we do not assert that as an empirical fact, for we all understand, as Horace said, that “sometimes even excellent Homer nods.” We hold to this course, however, because there is no other less dangerous way; if courts were free to “correct” what they believe to be congressional oversights by construing unambiguous statutes to the contrary of their plain meaning — apart from that rare case in which specific legislative history compels such a result — even a good faith attempt to further Congress’s goals would open the way to judicial hijacking of the power to legislate.
Consolidated Rail Corp. v. United States, 896 F.2d 574, 579 (D.C.Cir.1990); cf. Dow Jones & Co. v. Department of Justice, 908 F.2d 1006, 1009 (D.C.Cir.1990) (Silberman, J.) (“It may well be true that if Congress had thought about this question, the [statute] would have been drafted more broadly.... But Congress did not, and the words simply will not stretch to cover this situation_”), reh’g denied, No. 89-5353 (D.C.Cir. Oct. 5, 1990).
There is nothing in the legislative history of the market exclusivity provisions to suggest the sort of “clearly expressed legislative intention to the contrary” that is normally required before a court will act to override a statute’s plain meaning. See Burlington Northern R.R. Co. v. Oklahoma Tax Comm’n, 481 U.S. 454, 461, 107 S.Ct. 1855, 1860, 95 L.Ed.2d 404 (1987) (quoting United States v. James, 478 U.S. 597, 606, 106 S.Ct. 3116, 3121, 92 L.Ed.2d 483 (1986) (quoting Consumer Prod. Safety Comm’n v. GTE Sylvania, Inc., 447 U.S. 102, 108, 100 S.Ct. 2051, 2056, 64 L.Ed.2d 766 (1980))). Therefore, I cannot concur in the majority’s decision to override the words Congress chose to express its intention in this matter — particularly based upon judicial suspicion that Congress “did not focus” on a supposed anomalous eventuality that is, at any rate, without factual support in the record.
III.
Underlying the majority’s analysis is the assumption that if one can perceive any ambiguity in a statute, however remote, slight or fanciful, the statute must be pushed into the second step of Chevron analysis, in which the court must defer to a reasonable agency interpretation. This fundamentally misconceives the point of Chevron analysis. Chevron does not suggest that courts are to search statutes, overturning linguistic rocks and brush, in the hope of discovering some arguable ambiguity, which would then justify deference *995to an administrative construction. The court’s responsibility under the first step of Chevron analysis is considerably more serious: the court is to “employ[] traditional tools of statutory construction” to ascertain whether “Congress had an intention on the precise question at issue.” Chevron, 467 U.S. at 843 n. 9, 104 S.Ct. at 2781 n. 9. Minor ambiguities or occasional imprecision in language may be brooked under Chevron ’s first step, so long as “traditional tools of statutory construction” reveal Congress’ intentions. See NLRB v. United Food & Commercial Workers Union, Local 23, 484 U.S. 112, 123, 108 S.Ct. 413, 421, 98 L.Ed.2d 429 (1987) (“On a pure question of statutory construction, our first job is to try to determine congressional intent, using ‘traditional tools of statutory construction.’ If we can do so, then that interpretation must be given effect_”). Advancement to Chevron’s second step is not appropriate merely where the court stumbles across a perceptible ambiguity; rather, given that the judiciary remains “the final authority on issues of statutory construction,” Chevron, 467 U.S. at 843 n. 9, 104 S.Ct. at 2781 n. 9, abdication of that authority and deference to an administrative construction is legitimate only where the court confronts a gap in the statute that cannot be bridged by way of traditional tools of statutory construction and which can properly be characterized as an express or implied delegation of authority by Congress to an agency, see id. at 843-44, 104 S.Ct. at 2781-82.
In this case, the majority has alerted upon a claimed ambiguity in the statute that apparently has occurred to no other reader — not to the District Court, not to Abbott, not to the agency charged with administering the statute and not to the Government’s lawyers who appeared before this court on behalf of the agency. But the majority’s announcement that it can perceive an ambiguity in the statute is alone not grounds for bucking the whole business back to the agency. The real question is whether, notwithstanding some arguable imprecision in language, Congress’ intentions cannot be satisfactorily discerned from the terms of the statute. See K Mart Corp. v. Cartier, Inc., 486 U.S. 281, 291, 108 S.Ct. 1811, 1817, 100 L.Ed.2d 313 (1988) (“In ascertaining the plain meaning of the statute, the court must look to the particular statutory language at issue, as well as the language and design of the statute as a whole.”). Indeed, to accept the majority’s proposition that Chevron step-one analysis is finished as soon as a court can perceive some ambiguity in a statute would mean that all statutes must be analyzed under Chevron’s, deferential second step, for clever lawyers — and clever judges — will always be capable of perceiving some ambiguity in any statute, no matter how clearly Congress struggles to emblazon its intentions on the face of the statute. That is clearly not what the Supreme Court intended in Chevron and it would hardly leave the judiciary, in any meaningful sense, as “the final authority” on issues of statutory construction. In INS v. Cardozo-Fonseca, 480 U.S. 421, 446-48, 107 S.Ct. 1207, 1221-22, 94 L.Ed.2d 434 (1987), Justice Stevens (who wrote the opinion for the Court in Chevron) strongly indicates that the interpretation of a statutory provision remains “a pure question of statutory construction for the courts to decide” even when the provision at issue admits of some ambiguity. Thus, the second step of the Chevron test is applicable only when a court is unable to discern congressional intent after employing traditional tools of statutory construction.
On the facts of this case, Congress’ intention in the market exclusivity provisions is absolutely clear. “Active ingredient” is an unambiguous term of art and it does not include “active moiety.” The agency’s suggestion to the contrary flies in the face of the clear terms of the Act. Nor is there any genuine uncertainty about Congress’ intention in the statute to refer to the active ingredient of the new drug for which market exclusivity is sought. The majority’s discovery of a purported ambiguity in the statute’s use of the articles “an” and “the” is at least as strained as the agency’s finding of ambiguity in the word “including.” Abbott is entitled to 10 years of market exclusivity for Depakote *996because neither its active ingredient, dival-proex sodium, nor any ester or salt of divalproex sodium, has ever before been approved by the FDA.
Under Chevron, we are required to adhere to a statute’s plain terms. Some courts have held that a statute’s plain meaning may be ignored if there is a clear direction to the contrary in the legislative history or if strict adherence would defeat the clear purpose of Congress. However, neither possible exception applies in this case. Therefore, I would reverse the judgment of the District Court and apply the statute’s plain terms in Abbott's favor.
. Subsection (i) and subsection (v), at issue here, apply only to new drugs approved by the FDA during a transitional "window” from January 1, 1982, through September 24, 1984. Drugs approved after September 24, 1984, are entitled to generally shorter periods of market exclusivity under subsections (ii), (iii) and (iv) of 21 U.S.C. § 355(j)(4)(D) (1988). Abbott's Depakote, the drug whose claim to market exclusivity is contested in this action, was approved on March 10, 1983, and its right to market protection is therefore governed by either subsection (i) or subsection (v) of the Act.
. In Bob Jones, for example, the Court upheld an IRS interpretive ruling that restricted tax-exempt status only to "charitable” institutions even though the relevant statute contained no such literal restriction. The Court held that it was proper for the IRS to depart from the "plain language” of the statute where strict application of that language would provide governmental aid to an organization that was itself at odds with "a most fundamental national public policy” of racial equality. 461 U.S. at 593, 103 S.Ct. at 2029. The Court in that case found “unmistakable evidence that, underlying all relevant parts of the Code, is the intent that entitlement to tax exemption depends on meeting certain common-law standards of charity — namely, that an institution seeking tax-exempt status must serve a public purpose and not be contrary to established public policy.” Id. at 586, 103 S.Ct. at 2025.
In a more recent case, Green v. Bock Laundry Mach. Co., 490 U.S. 504, 109 S.Ct. 1981, 104 L.Ed.2d 557 (1989), the Court upheld a non-literal interpretation of Rule 609(a)(1) of the Federal Rules of Evidence where a strictly literal approach would lead to “an unthinkable disposition” and might render the Rule unconstitutional. See id. 109 S.Ct. at 1994 (Scalia, J., concurring in the judgment).
In Public Citizen v. United States Dep’t of Justice, 491 U.S. 440, 109 S.Ct. 2558, 105 L.Ed.2d 377 (1989), the Court rejected a literal application of a statute that would have led to the “outlandish” conclusion that Members of Congress had intended to "subject[ ] their own political parties to bureaucratic intrusion and public oversight.” Id., 109 S.Ct. at 2566 n. 9.