dissenting:
In the costs and attorneys’ fees statute before us, Congress has, perhaps deliberately but probably unintentionally, abdicated an important legislative decision to the judiciary. Thus, at Congress’ bidding, the court today — in the majority opinion here and, especially, in Sierra Club v. Gorsuch1 —necessarily plays legislature in its “interpretation” of the statute. I would decline to do so. Moreover, the way in which the court has played legislature has only added to the confusion and constitutional strain created by Congress. In the instant litigation, the majority misapplies whatever standard it has even on its own terms, especially by the refusal to apply the standard issue by issue. I must respectfully dissent.
I. OVERVIEW
We have before us certain applications for fees and costs associated with the litigation culminating in Alabama Power Co. v. Costle.2 In that action several petitioners challenged the validity of final regulations promulgated by the Environmental Protection Agency (EPA) on 19 June 1978. These “PSD” regulations generally aimed at the p revention of significant deterioration of air quality in the nation’s “clean air areas.” They interpreted and began the implementation of various provisions of the Clean Air Act Amendments of 1977.3 Both the litigation itself and our final opinion were lengthy and complicated. On some of the issues petitioners — who included representatives of various industries, environmental groups, and local governments — can fairly be said to have prevailed, on others to have failed, and on still other issues to have neither won nor lost.
Certain petitioners applied for the award of costs and attorneys’ fees under the 1977 Amendments, which provide in relevant part that “[i]n any judicial proceeding under this section, the court may award costs of litigation (including reasonable attorney and expert witness fees) whenever it deter*9mines that such award is appropriate.”4 Congress thus fashioned a statute the meaning of half of which we can determine. We can be reasonably certain that Congress intended for courts to award costs and attorneys’ fees in those suits in which petitioners prevailed. Congress also seemingly intended that such awards were not to be limited to petitioners who prevailed,5 a departure from all previous statutory authorization or previous practice in the courts.
But Congress did not say when awards were to be made to non-prevailing petitioners and, because it is impossible to discern a standard both different from “prevailing” and consistent with principled judicial interpretation, we should not apply this second, shrouded half. Accordingly, I would dismiss the petitions without prejudice so that they might be amended to request attorneys’ fees and costs only to the extent that petitioners prevailed. Moreover, even were I to follow the majority and embrace the “standard” promulgated in Sierra Club v. Gorsuch, I would not join the decision the majority has written in this case today.
In the discussion which follows I shall first analyze the meaning of the word “appropriate” in the costs and fees provision before us by reviewing the applicable statutes and their legislative histories. I shall then apply the interpreted provision to the instant litigation. I submit that the result reached in this dissent will be, besides more firmly grounded constitutionally, also more equitable and reasonable in its award.
II. THE MEANING OF “APPROPRIATE”
A. Applicable Statutes
As already stated, the key statute in this case is § 7607(f), which provides that “[i]n any judicial proceeding under this section [i.e., “Administrative proceedings and judicial review”], the court may award costs of litigation (including reasonable attorney and expert witness fees) whenever it determines that such award is appropriate.” This parallels an earlier statute, which empowered district courts to make similar awards for suits originating there.6 This earlier statute was part of the 1970 Amendments to the Clean Air Act.7
*10The issue which immediately presents itself, and which the statutes themselves do not answer, is what precisely “appropriate” means. On the one hand it is clear from other statutory language8 that, absent specific authorization by Congress, the award of attorneys’ fees, particularly to non-prevailing parties, is to be the exception rather than the rule.9 The Supreme Court, in Alyeska Pipeline Service Co. v. Wilderness Society,10 limited the opportunity to obtain public interest fees in the federal courts almost entirely to those situations where there was specific statutory authority.11 It found that “the approach taken by Congress ... has been to carve out specific exceptions to a general rule that federal courts cannot award attorneys’ fees beyond the limits of 28 U.S.C. § 1923 ....”12 Costs are generally awarded to the party prevailing on appeal.13
On the other hand, it is also clear that Congress intended the awards of costs and fees to be made in a somewhat more liberal fashion in cases falling under the 1970 and 1977 Amendments to the Clean Air Act. It can probably also be fairly implied from the statutes alone that in determining whether or not such awards are “appropriate,” courts were intended to take into account factors besides, though not necessarily excluding, whether or not a given party prevailed. This is so because prevailing vel non was already the standard for costs, and because it would have been a simple matter for Congress to have made this the standard for attorneys’ fees, too — as it has done, for instance, in other statutes.14 It is still unclear, however, what the other criteria for determining whether an award of the costs of litigation should be made are. Given the ambiguity of the statute itself, we turn to its legislative history.
B. Legislative History15
*11In its Report16 on the Clean Air Act Amendments of 1977 the House Committee on Interstate and Foreign Commerce stated that this provision would allow the award of attorneys’ fees and other costs “in the court’s discretion.”17 The House Report went on to state:
The committee bill also contains express authority for the courts to award attorneys fees and expert witness fees in two situations.20 The judicial review proceedings under section 307 [§ 7607] of the act when the court determines such award is appropriate [sic].
In the case of the section 307 judicial review litigation, the purposes of the authority to award fees are not only to discourage frivolous litigation, but also to encourage litigation which will assure proper implementation and administration of the act or otherwise serve the public interest. The committee did not intend that the court’s discretion to award fees under this provision should be restricted to cases in which the party seeking fees was the “prevailing party”. In fact, such an amendment was expressly rejected by the committee, largely on the grounds set forth in NRDC v. EPA, 484 F.2d 1331, 1388 [sic] (1st Cir. 1973).
In adopting this provision concerning fees, the committee intended to meet the requirement for specific authorization imposed by 28 U.S.C. sec. 2412 and by the Supreme Court’s ruling in Alyeska Pipeline Service Co. v. Wilderness Society, 421 U.S. 240, 95 S.Ct. 1612, 44 L.Ed.2d 141 (1975).
Assuming that the citation in the House Report was to page 1338, the passage which the House Report endorsed19 in Natural Resources Defense Council, Inc. v. EPA reads as follows:
We are not impressed by the government’s argument that because some issues were decided adversely to petitioners each party should bear its own costs. The authorizing language of section 304(d) [i.e., § 7604(d)] permits an award “to any party, whenever the court determines such award is appropriate.” This suggests greater latitude even than is found in 28 U.S.C. § 2412, which authorizes awards to “the prevailing party”. We are at liberty to consider not merely “who won” but what benefits were conferred. The purpose of an award of costs *12and fees is not mainly punitive. It is to allocate the costs of litigation equitably, to encourage the achievement of statutory goals. When the government is attempting to carry out a program of such vast and unchartered [sic] dimensions, there are roles for both the official agency and a private watchdog. The legislation is itself novel and complex. Given the implementation dates, its earlier interpretation is desirable. It is our impression, overall, that petitioners, in their watchdog role, have performed a service.
Were we to believe that the litigation were wholly or in substantial part frivolous, we would not, of course, award costs of any description to petitioners. In such cases, indeed, we reserve the right to award costs and fees in favor of the EPA. But the challenges here, even those not sustained, were mainly constructive and reasonable. And petitioners were successful in several major respects: they should not be penalized for having also advanced some points of lesser weight.20
The 1977 Senate Report stated that § 7607(f)’s “purpose ... is to carry out the intent of the committee in 1970 [i.e., when the 1970 Amendments were added] that a court may, in its discretion, award costs of litigation to a party bringing a suit under section [7607] of the Clean Air Act.”21 The 1977 Conference Report said, “The conferees agreed to accept both the House provision, which is identical to one portion of the Senate provision, as well as the entire Senate provision.”22
Because the statutory language and the legislative history of the 1977 Amendments are far from dispositive, courts have also looked to the legislative history of the earlier and, in the final version at least, almost identically-worded § 7604(d).23 In 1970 the House bill did not include a provision for citizen’s suits.24 In the Senate Report,25 however, three relevant passages appear. First:
Concern was expressed that some lawyers would use section 304 [i.e., § 7604] to bring frivolous and harassing actions. The Committee has added a key element in providing that the courts may award costs of litigation, including reasonable attorney and expert witness fees, whenever the court determines that such action is in the public interest. The court could thus award costs of litigation to defendants where the litigation was obvi*13ously frivolous or harassing. This should have the effect of discouraging abuse of this provision, while at the same time encouraging the quality of the actions that will be brought.
The Courts should recognize that in bringing legitimate actions under this section citizens would be performing a public service and in such instances the courts should award costs of litigation to such party. This should extend to plaintiffs in actions which result in successful abatement but do not reach a verdict. For instance, if as a result of a citizen proceeding and before a. verdict is issued, a defendant abated a violation, the court may award litigation expenses borne by the plaintiffs in prosecuting such actions.26
Second, it said, “Government initiative in seeking enforcement under the Clean Air Act has been restrained. Authorizing citizens to bring suits for violations of standards should motivate governmental agencies charged with the responsibility to bring enforcement and abatement proceedings.”27 Third, the Senate Report stated, “The court may award costs of litigation to either party whenever the court determines such an award is in the public interest without regard to the outcome of the litigation.”28
Finally, the 1970 Conference Report stated simply that “[discretionary authority was provided to the court to grant reasonable attorney and expert witness fees.”29 Senator Muskie’s “Summary of the Provisions of Conference Agreement on the Clean Air Act Amendments of 1970,” presented to the Senate on the day after the Conference Report was published, repeats largely verbatim the language of the Senate Report quoted above.30
C. Analysis
The statutes and legislative histories raise more questions than they answer about the definition of “appropriate.”31 All precedent and statute points'to the fair conclusion that there is a general presumption against one side (particularly the winner) paying the other’s attorneys’ fees. Thus, when in doubt, no award will be made. Dramatic deviations from the American rule will be construed against. Similarly, there is a presumption that non-prevailing parties are not entitled to costs.
It must also be kept in mind that “appropriate” is at least in part a reference to the fact that generally attorney’s fees are not awarded even when a party wins. Thus, in the absence of an explicit statutory provision, attorneys’ fees would not be awarded to even a prevailing party. In short, while there is an inference here that, as we discuss below,32 something less than “prevailing” will suffice, there is generally an intimation that something more than “prevailing” is required — explicit statutory authority.
A review of the relevant statutes and legislative history, however, makes clear that at least some departure from the stricter rules discussed above was contemplated by § 7607(f). It is fair to conclude that prevailing is not a necessary condition for the award of attorneys’ fees and costs, though it is sufficient. Congress, that is, generally intended that an award be made when a petitioner prevailed, plus in some other instances as well. The reasons for this conclusion are set out in the next section. In the subsequent section, however, I *14outline why the “plus” in this “prevailing-plus” standard is without judicially cognizable meaning.
1. Congress intended a “prevailing-plus” standard
There is abundant evidence that prevailing should be sufficient. The 1970 Senate Report states, “The Courts should recognize that in bringing legitimate actions under this section citizens would be performing a public service and in such instances the Courts should award costs of litigation to such party.”33 Although the context of this passage makes it unclear whether “legitimate actions” might extend to all non-frivolous suits, it is likely that where the party prevails such action would be considered “legitimate.” If the idea is to encourage “legitimate actions,” “proper implementation,” or suits in the “public interest,” an award to parties who prevail would seem to be required by the legislative history.
Section 7607(f), of course, adopted the same language as § 7604(d) and, thus, presumably intended the same effect. It is also to be presumed that Congress in 1977 was aware of the cases decided under the 1970 Amendment language implying that prevailing was a sufficient condition for an award,34 and that in adopting the same language the holdings in these cases were approved. Finally, the 1977 House Report stated that the citizen’s suit provisions were intended “to encourage litigation which will assure proper implementation and administration of the act or otherwise serve the public interest. The committee did not intend that the court’s discretion to award fees under this provision should be restricted to cases in which the party seeking fees was the ‘prevailing party’.”35 Although this again makes it unclear to what extent Congress wanted awards made to non-prevailing plaintiffs, it is clear that prevailing plaintiffs were to be awarded attorneys’ fees and costs.
The legislative history also indicates that, while prevailing is sufficient, it is not necessary for an award to be made. We have already pointed out the significance of the fact that the statute itself does not invoke the standard of prevailing, though other statutes have. Moreover, as we just quoted, the 1977 House Report explicitly stated, “The committee did not intend that the court’s discretion to award fees under this provision should be restricted to cases in which the party seeking fees was the ‘prevailing party’. In fact, such an amendment was expressly rejected . ...”36 The House Report went on to endorse that part of Natural Resources Defense Council, Inc. v. EPA, which found “greater latitude [in making awards] even than is found in 28 U.S.C. § 2412, which authorizes awards to the ‘prevailing party’.”37 The 1977 Senate Report said simply that “a court may, in its discretion, award costs of litigation to a party bringing a [§ 7607] suit”38 under the Act, making no mention of a need to prevail in that suit, and the 1977 Conference Re*15port indicated that “the conferees agreed to accept”39 both the Senate and House provisions — with, one must assume, their respective legislative histories. Finally, the 1970 Senate Report stated, “The court may award costs of litigation to either party whenever the court determines such an award is in the public interest without regard to the outcome of the litigation.”40
Still, it is also possible that if prevailing were not required, at least “not losing" was41 The sole examples given of when a court might make an award to a nonprevailing party in any of the legislative history are in the 1970 Senate Report and 1977 House Report. The former states:
The Courts should recognize that in bringing legitimate actions under this section citizens would be performing a public service and in such instances the courts should award costs of litigation to such party. This should extend to plaintiffs in actions which result in successful abatement but do not reach a verdict. For instance, if as a result of a citizen proceeding and before a verdict is issued, a defendant abated a violation, the court may award litigation expenses borne by the plaintiffs in prosecuting such actions.42
Similarly, Natural Resources Defense Council, Inc. v. EPA43 endorsed at least in part by the 1977 House Report,44 involved an award to a plaintiff who, while not prevailing on all counts, nevertheless prevailed on many of them. The court there stated that “the challenges here, even those not sustained, were mainly constructive and reasonable. And petitioners were successful in several major respects; they should not be penalized for having also advanced some points of lesser weight.”45
2. The “plus" has no judicially cognizable meaning
In any event, if we assume that awards are not limited to those who prevail — or who at least do not lose — then the obvious next question is what standards are the courts to use in making awards? It is, unfortunately, at this point that the trail is lost and evidence of congressional intent vanishes without a trace. While we can deduce with some certainty that prevailing is sufficient for an award to be made, and are told that other standards besides prevailing exist, we are never given the barest clue of what those standards are.
a. Sierra Club and the “non-frivolous” standard
In Sierra Club v. Gorsuch,46 also decided today and relied upon by the majority, a number47 of standards are conjured up, but *16they are all defective. Some are simply question-beggars: whether “the suit [was] of the type Congress intended to encourage,” what “benefits” were “conferred” by the litigation, and whether the litigation was in the “public interest.” Other criteria clearly have no basis whatever in the statute or its history: whether the suit would have been brought without an award, the need to ensure representation of the environmentalist point of view, and the aid given to congressional reevaluation of the Clean Air Act. What is truly remarkable is that in the seemingly interminable list of criteria explicitly or implicitly endorsed by Sierra Club and the majority, all are defective: they are either unworkable, or unrelated to anything in the statute or its legislative history, or both.
Moreover, even if some of these criteria would be workable if narrowly and carefully construed, a fatal flaw would remain: It is never made clear how the various standards endorsed are to relate to one another. How are they to be balanced, or are they to be balanced? Is it necessary to meet more than one test? Are some to be weighted more heavily than others? Which ones? Not only has the court acted like a legislature, it has not even improved over the inconclusive job done by Congress. If anything, the vagueness of the statute has been augmented to validate any whim of any judge.
Thus the most troubling aspect of Sierra Club and, to the extent it endorses Sierra Club, the majority opinion herein, is their effect of enacting the all-but-the-frivolous standard. That is, I fear that these multifarious and diffuse “standards” amount to no standard at all. It would be difficult to imagine a non-frivolous petitioner who *17could not qualify under the language used. I find it at least troublesome, for instance, that while no petitioner was successful in any appeal of EPA regulations in Sierra Club, all appeals were apparently adjudged to have “substantially eontribute[d] to the goals of the Act.”
The trouble with the non-frivolous standard is that it would dramatically alter the award structure. In the absence of some evidence that Congress intended a dramatic departure, we should not adopt such a construction. Were such an expensive and unprecedented result contemplated, it surely would have been mentioned.48 Yet in none of the myriad House and Senate reports, committee prints, and floor debates cited in this opinion were more than a few lines ever devoted to these sections at all, let alone any mention made of such a change. Indeed, only once was any part of § 7607 mentioned in the floor debates, and that was of a version later rejected.49 Moreover, it must be kept in mind that the whole tenor of the Clean Air Act was, in both 1970 and 1977, to strike a balance among industry, government, and environmentalists. If Congress meant for “appropriate” to mean that all non-frivolous plaintiffs would recover, it surely would have said so. As the Supreme Court observed in Alyeska:
[Congressional utilization of the private-attorney-general concept can in no sense be construed as a grant of authority to the Judiciary to jettison the traditional rule against nonstatutory allowances ... and to award attorneys’ fees whenever the courts deem the public policy furthered by a particular statute important enough to warrant the award.50
Since the approach taken by Congress . .. has been to carve out specific exceptions to a general rule that federal courts cannot award attorneys’ fees beyond the limits of 28 U.S.C. § 1923, those courts are not free to fashion drastic new rules with respect to the allowance of attorneys’ fees ... in federal litigation or to pick and choose among plaintiffs and the statutes under which they sue and to award fees in some cases but not in others, depending on the courts’ assessment of the importance of the public policies involved in particular cases.51
Moreover, it would have been a simple enough matter for Congress to have drafted a statute embodying the non-frivolous standard, or at the very least to have included its intention to adopt such a standard in the legislative history somewhere, but Congress did not do so. Finally, even were we to adopt such a standard for private litigant-versus-private litigant actions, we might still decline to do so for private litigant-versus-government actions. In the latter cases although perhaps the public is afforded some nebulous benefit by having the law clarified, it must pay the very real cost of a lawsuit and of the diversion of EPA’s resources from doing what it is supposed to be doing.52
I am hopeful that the awards by the majority today will be read as generally limited and circumspect, starting the post-Sierra Club cases in the right direction. Perhaps courts will, in spite of Sierra Club’s ambiguities, enact something besides the non-frivolous standard, which Congress clearly did not intend. But the point is that courts ought not to be enacting standards at all, even standards to which Congress might not have objected. We should instead follow congressional guidance and, where there is none, we should wait.
*18b. The “public interest” standard
Besides “prevailing,” the one53 other gauge that was mentioned in the legislative history, though only briefly, is the “public interest.”54 Yet we can hardly embrace this “standard.” It was, significantly, expressly struck from the statute in 1970.55 Moreover, it is nowhere defined. Presumably, it can have meant one of two things, “consistent with valid public law” or “consistent with valid public policy.” Yet neither meaning affords a court any additional guidance. I shall consider in turn the ramifications of adopting either.
If the “public interest” meant suits consistent with valid public law, then it is difficult to see how this standard is different from the prevail/not prevail standard.56 After all, as a matter of valid public law, non-prevailing petitioners have lost. It is especially difficult to see how it is ever consistent with “valid public law” to force the government to spend taxpayers’ dollars defending an EPA action that is eventually adjudicated as lawful.
Nor is it implausible that Congress intended “public interest” to mean “consistent with valid public law.” Note that both the 1977 and 1970 legislative histories gave as a central purpose of the statutes the encouragement of litigation insuring “proper implementation” of the Amendments by the government;57 if the idea was to get the government to do its job, unsuccessful suits have an attenuated utility. This point is buttressed by the 1970 legislative history which spoke of “legitimate actions.” Note too that the 1970 and 1977 legislative histories addressed themselves as much to discouraging frivolous suits as to encouraging valid ones.58 Finally, as discussed above, *19the only examples given of appropriate awards to “non-prevailing” plaintiffs were when those plaintiffs had practically won or at least had not lost.
Courts and litigants have on occasion argued that even though a party did not prevail, it served the public interest and perhaps valid public law — or, as Sierra Club phrased it, made a “substantial contribution” 59 — by exploring a previously uncharted legal issue.60 The resulting “standard” is apparently endorsed by the majority, but it really amounts to awards for all non-frivolous plaintiffs, and I would therefore decline to adopt it for the reasons discussed earlier.61
Similarly, it may be argued that Congress deemed losing suits to be in the “public interest” and consistent with valid public law since it wanted every possible winning suit brought and was willing to pay the losers to ensure this result. This is not an inherently implausible argument, but it too is totally without support in the language of the statute or its legislative history. It would lead to the same non-frivolous standard, which I would reject.
It should also be pointed out that, although the protection of the environment is surely in the public interest, there are other activities which are presumably in an even greater public interest, and thus involve public law even more critically, for which attorneys’ fees are not awarded to all non-frivolous plaintiffs. For instance, in civil rights suits it is only in the court’s discretion to award attorneys’ fees to prevailing plaintiffs, and no mention is made of allowing courts to award costs or fees to non-prevailing plaintiffs.62 Nor are the free-rider effect and other market-failure arguments attenuated in civil rights cases. Indeed, given the socio-economic background of many civil rights plaintiffs on the one hand, and environmentalists on the other, more financial aid in litigation is needed for the former, not the latter. As the Supreme Court pointed out in Alyeska Pipeline Service Co. v. Wilderness Society:
Congress itself presumably has the power and judgment to pick and choose among its statutes and to allow attorneys’ fees under some, but not others. But it would be difficult, indeed, for the courts, without legislative guidance, to consider some statutes important and others unimportant and to allow attorneys’ fees only in connection with the former. If the statutory limitation of right-of-way widths involved in this case is a matter of gravest importance, it would appear that a wide range of statutes would arguably satisfy the criterion of public importance and justify an award of attorneys’ fees to the private litigant. And, if any statutory policy is deemed so important that its enforcement must be encouraged by awards of attorneys’ fees, how would a court deny attorneys’ fees to private litigants ... seeking to enforce constitutional rights?63
Thus, we cannot conclude that Congress believed some suits brought under the 1977 Amendments to be inherently in the public interest, win or lose, if by public interest we mean “consistent with valid public law.”
On the other hand, if “public interest” meant consistent with good public policy, I would a fortiori reach the same result today. In that case Congress would be asking the courts to legislate, and we should refuse to do so. Just as the Supreme Court held *20that courts could not aggrandize legislative power in Alyeska, we should hold that courts cannot accept legislative power here. Congress may direct courts to make awards consistent with good policy only if it first describes what good policy is. It has not done so here. The judiciary cannot be left to define this standard on its own. That is a quintessentially legislative task.
The unfortunate phenomenon of the judiciary acting as legislature is not entirely the fault of judges. Congress must also bear some of the blame because of deliberate or inadvertent abdication of decisionmaking to the courts. In drafting a statute charging the courts to award costs and fees where “appropriate,” it is difficult — and at this point unimportant — to tell whether this vague instruction was the result of the efforts of draftsmen who had an enormously complex and lengthy act to write, or the result of a deliberate compromise in the bill’s acrimonious passage.
“Appropriate,” without more, is not an appropriate word, for its meaning varies with the politics of the. reader. It should be clear that the award of fees and costs might in a given instance seem appropriate to some but inappropriate to others.64 Webster’s Third New International Dictionary defines “appropriate” as “specially suitable: fit, proper.” This definition demonstrates the term’s subjectivity and vagueness — and thus the difficulty of any principled judicial interpretation.
It is true that courts have long had some discretionary authority to award costs and attorneys’ fees. But the authorization here is new and different. The discretionary authority granted to courts in the past has involved' determinations which courts are qualified to make: whether a suit was frivolous, whether a party prevailed, and so forth. Even an award according to “equity” has had an historical gloss. But here we are asked to award fees where “appropriate,” with, in at least some cases, a requirement that we act as legislature. Delegation of such legislative power by Congress to an agency empowered by Congress is one thing, to a separate branch of government another, and to the non-political branch yet a third. It is in the latter case that the delegation is most likely to be unacceptable, and with which we are today confronted. Policymaking of any sort is likely to run afoul of the “case or controversy” requirement of Article III, and the legislative function is more antithetical to the judiciary, more foreign to its very nature.
Professor Charles L. Black, Jr., of Yale Law School, in his Oliver Wendell Holmes Lectures delivered at Harvard Law School in 1979, observed that a
live question has to be whether the materials available as foundations for judicial judgment contain some concepts and terms so vague and ambiguous, so intractable to the normal intellectual processes of law, that a congressional command to apply them, and so, necessarily, to interpret them, must be disobeyed —on the Article III45 ground that the judiciary has been commanded to perform a non-judicial function — as might be the case if the district judges were told to award $1,000, in reasonably lawful money of the United States, to every virtuous person who filed a nice-looking complaint, or to make sure, by the wielding of the mandatory and injunctive powers, that only reverend and discreet merchants were appointed to the Federal Trade Commission.46
Professor Black is echoing a sentiment expressed repeatedly by the Supreme Court in the course of American jurisprudence.66
*21The Justices were faced early on, in 1792, with an attempt by Congress to enlarge the judiciary’s power beyond the scope of Article III. But the members of the Court — albeit separately, in their capacity as Circuit Justices — refused to accept this proffer of power, saying they were agreed:
That by the constitution of the United States, the government thereof is divided into three distinct and independent branches, and that it is the duty of each to abstain from, and to oppose, encroachments on either. That neither the legislative nor the executive branches, can constitutionally assign to the judicial any duties, but such as are properly judicial, and to be performed in a judicial manner.67
One hundred and eighty-four years later these principles remained, and still now remain, vital. In 1976 the Supreme Court, in its per curiam decision in Buckley v. Valeo,68 cited James Madison’s discussion in The Federalist No. 47 of “Montesquieu’s well-known maxim that the legislative, executive, and judicial departments ought to be separate and distinct”:
“ ‘Were the power of judging joined with the legislative, the life and liberty of the subject would be exposed to arbitrary control, for the judge would then be the legislator. Were it joined to the executive power, the judge might behave with all the violence of an oppressor.’ ”69
Even more pertinent to the present case were Hamilton’s words in The Federalist *22No. 78: “To avoid an arbitrary discretion in the courts, it is indispensable that they should be bound down by strict rules and precedents which serve to define and point out their duty in every particular case that comes before them ....”70 Where are the “strict rules and precedents which serve to define and point out [our] duty” in the grant of authority to determine what is “appropriate” in the circumstances here?
The statute as written confuses legislative and judicial decisionmaking in the manner warned against by Montesquieu, Madison, Hamilton, and the Supreme Court in 1792, 1976, and thereafter. The case at hand seems, in particular, to be very much like National Cable Television Association, Inc. v. United States.71 There the Supreme Court was faced with the task of interpreting an act authorizing each federal agency to calculate fees to be charged for its services according to its determination of what is “ ‘fair and equitable taking into consideration direct and indirect cost to the Government, value to the recipient, public policy or interest served, and other pertinent facts ....’”72 In the course of its interpretation, Justice Douglas for the Court pointed out that “[t]he addition of ‘public policy or interest served, and other pertinent facts,’ if read literally, carries an agency far from its customary orbit and puts it in search of revenue in the manner of an Appropriations Committee of the House.”73
Justice Douglas then discussed the pitfalls of such broad delegation, citing A.L.A. Schecter Poultry Corp. v. United States74 and J. W. Hampton, Jr., & Co. v. United States.75 He concluded with this paragraph, which parallels our treatment of a similarly and perilously broad delegation in the instant case:
Whether the present Act meets the requirement of Schechter and Hampton is a question we do not reach. But the hurdles revealed in those decisions, lead us to read the Act narrowly to avoid constitutional problems.76
Like the Supreme Court, we are faced with a delegation of legislative authority so broad that it would require a non-congressional body to make “public policy” determinations that are, constitutionally, the exclusive province of the legislative branch. Indeed, the delegation in the instant case is even more dangerous since it is not to an agency created by Congress but to a separate, and non-political, branch of government. Moreover, it is somewhat easier for us to exorcise “public policy” from the statute since it is only part of the legislative history, not the statute itself. Thus, we should adopt the approach taken by the Supreme Court in National Cable and read the delegation narrowly.77 In doing so we *23would simply follow the conventional practice of construing a statute, where possible, so as to avoid issues of constitutionality.78
Also closely analogous to the case at hand is TVA v. Hill.79 There the Supreme Court was “urged to view the Endangered Species Act ‘reasonably,’ and hence shape a remedy ‘that accords with some modicum of common sense and the public weal.’ ”80 The Supreme Court responded, “But is that our function? We have no expert knowledge on the subject ....”81 In that case the Court thought it clear that the statute allowed for no such weighing, while in the case before us today we have, perhaps, the consent of Congress to do its job. But this is a distinction without a difference since Congress has no more authority to grant power in violation of our constitutional scheme than we have authority to usurp power. Marbury v. Madison82 of course, rested on Chief Justice Marshall’s conclusion that the Supreme Court could not accept a congressional, but unconstitutional, enlargement of its original jurisdiction. As the Court in TVA v. Hill more recently concluded:
... [Tjhese principles [of balancing equities in “the public interest”] take a court only so far. Our system of government is, after all, a tripartite one, with each branch having certain defined functions delegated to it by the Constitution. While “[i]t is emphatically the province and duty of the judicial department to say what the law is,” Marbury v. Madison, 1 Cranch 137, 177 [2 L.Ed. 60] (1803), it is equally — and emphatically — the exclusive province of the Congress not only to formulate legislative policies and mandate programs and projects, but also to establish their relative priority for the Nation. Once Congress, exercising its *24delegated powers, has decided the order of priorities in a given area, it is for the Executive to administer the laws and for the courts to enforce them when enforcement is sought.
We agree with the Court of Appeals that in our constitutional system the commitment to the separation of powers is too fundamental for us to pre-empt congressional action by judicially decreeing what accords with “common sense and the public weal.” Our Constitution vests such responsibilities in the political branches.83
c. Conclusion
Thus, neither meaning of “the public interest” affords a court any additional guidance. Because this was the only candidate endorsed by the legislature for the “plus” in the prevailing-plus standard Congress apparently intended, there is no comprehensible or principled meaning for “appropriate” that extends further than “prevailing.” Accordingly, § 7607(f), as written, enables us to award attorneys’ fees and costs if, and only if, petitioners have prevailed in at least some part of their asserted cause of action.84
It might be objected at this point that it would be anomalous to enforce the “prevailing” half of the statute when, as concluded early on, the clear intent of Congress was prevailing plus something else. The argument would run that, if we know for certain that Congress did not intend just prevailing, we cannot enforce the statute as if it were written that way.
This argument is flawed, however. It might have some force where the “something else” to be added to “prevailing” was intended to limit it. In that case enforcing the prevailing part without the part that was to limit prevailing might indeed be impossible to defend, rather like a contract-enforcing case in which buyer was forced to pay seller since that part of the contract was clear while seller was absolved from delivery to buyer since that part of the contract was too vague. This is not the case here, however. The uncertain part of the statute was to be in addition to “prevailing,” so Congress has given us at least half a loaf. Again drawing an analogy from commercial law, this is rather like a contract wherein certain clauses are independent and severable.
It might also be objected that if we reject the word “appropriate” here, we would be obligated to reject the similarly broad language which pervades much of the rest of our statutory law. This is a non sequitur. While it is true that broad language exists elsewhere in the United States Code, a judicially cognizable meaning can typically be gleaned from the statutory context, legislative history, prior case law, or some other gloss.85 As we have seen, these sources offer no aid here. I emphasize again, too, that the legislative delegation here is to another branch, a non-political branch with neither expertise nor experience in deciding what kinds of suits should be encouraged. And I emphasize again that I would not strike down the statute here anyway, but only construe its language so as to avoid the necessity of making a constitutional ruling. For all of these reasons, the rejection of what Congress did here would have a carefully limited impact elsewhere.
*25III. THE INSTANT LITIGATION
A. General Disposition
In turning now to the specific litigation before us, I hope to make three points. The first is that the majority in this case, whatever standard it relies upon, should have undertaken an issue-by-issue analysis of the appropriateness of an award. The second is that an issue-by-issue application of the prevailing-only standard would result in an equitable and reasonable, besides more legally sound, disposition. Thus, it cannot be argued that my suggested construction leads to an absurd result. The third is that a theory suggested, albeit half-heartedly, by the majority — that a petitioner’s motive be scrutinized by the court in determining the appropriateness of an award — is untenable.
The tabulation at the Appendix summarizes the earlier holding on the merits in this litigation. There were at least twenty-four discrete substantive holdings in Alabama Power. There were eleven holdings favorable to the EPA and eleven unfavorable; the environmental petitioners had victories in five instances and lost in three; the District of Columbia won both its arguments, while the State of Texas won one and lost one; the industry petitioners lost nine and won five. In two instances, even on the narrow issue involved, there was no clear prevailing party.
My major quarrel with the majority’s general disposition of this ease is that it seems to me inconsistent with the statute and even with Sierra Club to conclude summarily that “[w]e encounter little difficulty in granting Sierra attorneys’ fees for their activities as petitioners . ...” 86 No reference is made to the issues they litigated even in general, let alone in the issue-by-issue manner I think necessary for an appropriate award. In fourteen words the majority disposes of the bulk of this case. Symptomatic of the majority’s approach is its conclusion that, on one issue, “Sierra and EPA were adversary parties. We accordingly approve [Sierra’s] claim for attorneys’ fees' in that regard.”87 Why the mere fact that Sierra was opposed to EPA should qualify it, without any discussion of how Sierra’s action redounded to the “public interest,” or whatever, is not explained. If Sierra Club is to be more than just a justification for an award in any non-frivolous suit, I think it is necessary to discuss — as even the per curiam there did — the issues for which an award is being made, not just who the parties are. Whether a suit is “in the public interest” cannot be determined solely by who brings it. It must hinge on the suit itself. Otherwise, we are left with assuming that an avowal by petitioner that it is a public interest group is all that is necessary for an award to be “appropriate.” Is this at all determinative of “whether, litigation by that party has served the public interest by assisting the interpretation or implementation of the Clean Air Act”88?
Thus, it is not at all clear how Sierra Club can be relied upon to make the across-the-board awards the majority makes today. Sierra Club at least pretended to a careful, issue-specific approach. In Natural Resources Defense Council, Inc. v. EPA, the court warned:
Petitioners are not a public agency and are legally responsible to no one but themselves; we must satisfy ourselves that the taxpayers’ money will not be used to support needless or excessive legal items. We must also recognize that petitioners, as surrogate attorneys, for the interests of the public and the EPA, have volunteered and “imposed” their services on “clients” that never contracted for them.89
That the majority nonetheless eschewed a careful analysis confirms my fears that Sierra Club’s pretensions to balance were only pretentious, and that its effect will be simply to award tax dollars to any environmental litigant who can scrawl “7607(f).”
*26The majority justifies declining to undertake an issue-by-issue analysis of its decision to award environmentalists attorneys’ fees by saying, “The number, complexity and frequent interrelationships of the legal problems presented by these consolidated cases as a whole obviates the need for an issue-by-issue explication of our decision to award attorneys’ fees.”90 The obvious question is ... Why? Does the difficulty of reaching the proper result make it less “appropriate” to do so? In the next sentence the majority intimates that issue-by-issue consideration might be necessary if awards on some issues would be inappropriate; the obvious question here is, how do you know if such issues exist if you have not made an issue-by-issue analysis first?91
Conversely, in explaining why an award of costs to industry petitioners — albeit under the general costs statute92 — was not made, the justification seems to be that “in this large and complex body of cases none of the principals clearly prevailed.”93 Besides not explaining why complexity justifies not considering issues separately, making awards to some petitioners, and not making awards to other petitioners, the majority opinion will deprive a party who may have won on many issues of any cost recovery, discourage consolidation of cases, and encourage piecemeal litigation.
As difficult as issue-by-issue analysis may be for making awards, I think it essential for ensuring that such awards will be “appropriate,” however we define it. The proper approach is, then, to consider the separate issues and litigants carefully, and to apply a judicially cognizable standard, which is to say the “prevailing” standard.
One of the parties, the District of Columbia, is easily dealt with. The District brought two suits and won both. It clearly prevailed, and is therefore clearly entitled to attorneys’ fees and costs.94 Also, because of the many parties to be served in this litigation, the District is entitled to reimbursement for even those copies of its briefs over the normal allowance of fifty specified by Local Rule 15. Thus, I concur in the majority’s decision to award costs and fees to the District of Columbia.95
The remaining parties, however, present harder questions. Each of them, while winning on at least one issue, also lost on at least one. Several ideas can be suggested and discarded very quickly. Failing to prevail on one issue ought not necessarily to bar an award for the others — one need not pitch a perfect game to win. On the other hand, prevailing cannot really be defined in terms of a won-lost percentage. In the first place, what figure would we require? .367 is a very good individual hitting percentage, but few teams would make the playoffs with an overall won-lost record like that, even last year. Moreover, all issues should not be weighted the same. Some games are decisive, others strictly warm-up.
The best solution is to grant the fees and costs incurred issue-by-prevailing issue. Since it will be far easier for petitioners to itemize these expenses than for the EPA to do so, they must demonstrate how much they spént on the issues they prevailed on and for which they are claiming fees and costs. A demonstration of how much was spent on each particular issue may be diffi*27cult, and we should therefore be willing to rely heavily on the integrity of counsel in reaching these determinations in this case. In the future, however, litigants planning to ask for an award of costs of fees would be required to keep reasonably detailed accounts on which issues what litigation expenses were incurred.
Because petitioners hoped that they might be awarded costs and fees regardless of whether they prevailed or not, in their requests for attorneys’ fees and costs they understandably did not address the problem of on which issues they prevailed. Accordingly, their current requests should be dismissed without prejudice so that they may refashion them.
Note that the Supreme Court has been fairly liberal in its definition of “prevailing” in the context of civil rights fees,96 and this liberality should be extended to § 7607(f) awards — especially given the apparent intent of Congress to endorse a prevailing-plus standard. Moreover, I would be willing to make awards where, although petitioners did not technically prevail, they can demonstrate that they did not “lose” either.97 For instance, actions which result in successful abatement but do not reach a verdict merit the award of costs and fees.98
It should be noted at this point that only environmental petitioners and the District of Columbia have invoked § 7607(f) — industry petitioners have not.99 However, the majority has implied that were such a claim to be filed, the court may perhaps need to look beyond the suit itself, and into the heart of the petitioner, in deciding whether it is worthy of an award.100 Though the majority only suggests the idea as a possibility, this infant is ugly enough to warrant strangling in the cradle.
We ought not require the court to look into the heart of the petitioner to determine whether his motivations were in the public interest only. This, I submit, is an impossible standard for human beings to apply.101 Moreover, it is at odds with the general purpose of § 7607(f) to encourage proper implementation of the act, which must depend on what suit is brought, not on who brings it.
If “the dominant consideration [in whether an award is appropriate] is whether litigation by that party has served the public interest by assisting the interpretation or implementation of the Clean Air Act,”102 then why would a party invite ineligibility for an award when it has “no more than its own economic interests at heart” 103? What relevance does the former have to the latter? It also follows that there is the implication in the majority opinion that prevailing is not only unnecessary, but not even sufficient, for an award to be made if the wrong party has brought the suit. This seems entirely inconsistent with Congressional intent,104 and will remove one of the few easily workable standards we have. The reliance of the majority on Environ*28mental Defense Fund, Inc. v. EPA,105 an opinion involving the statute and legislative history of an entirely different act, seems to me questionable jurisprudentially. What is appropriate under one statute is not necessarily appropriate under another. Senator Magnuson, on whose remarks alone the majority’s suggestion hinges, was not involved with the.committees considering § 7607(f), so far as the floor debates and Senate reports reveal. The legislative history of § 7607(f) makes clear only that it is to adopt the standards of § 7604(d),106 which in fact preceded the statutes involved in EDF. Moreover, is the majority really prepared to be bound by the legislative histories of all fourteen statutes that use the “appropriate” standard for attorneys’ fees?107 How is this reconciled with Sierra Club’s discounting of even § 7604(d)’s efficacy in interpreting our statute?108
I agree with the majority’s conclusion that “pro-environment” is no criterion.109 It is puzzling, then, how the majority can then flirt a few words later with the even less plausible criterion of some sort of pure-heart test.
The criterion of presence vel non of an economic incentive to bring the suit raises a host of additional objections. How are we to define “economic incentives”? Why should only economic incentives disqualify? Does any economic incentive disqualify a suit? If not, where is the line drawn? The standard would prohibit an award to a tenant farmer who seeks to stop a nearby factory from polluting his water supply, but would allow an award to his amateur fisherman brother-in-law who visits him on weekends. I am not even sure how far I would want to carry the argument that environmental petitioners have no economic interest. This will also add to the already anomalous relationship of standing issues in these cases: petitioners must have enough interest at stake for standing, but not too much for § 7607(f).
On the other hand, suppose a plant is required by regulation, or is sued by environmental petitioners, to use or not use certain equipment. It may not be worth it to the plant to go to court unless attorneys’ fees are awarded it. Yet the majority and Sierra Club apparently might deny an award to the plant even if it prevailed. Indeed, the plant would even be made to pay the other side, which lost. I still see no relationship between the presence or absence of economic incentives and “whether litigation by that party has served the public interest by assisting the interpretation or implementation of the Clean Air Act.” Moreover, the Clean Air Act was intended to strike a balance between environmentalists and industry, so preventing overenforcement of the Act is as much in the public interest as preventing underenforcement.110
It cannot be said that awards to certain petitioners will not aid proper implementation of the Act since these petitioners would have sued anyway. Not only is this speculative, but the prospect of attorneys' fees will provide an additional incentive and will alter the decisionmaker’s calculus of whether to sue or not. If the suit contemplated is “in the public interest,” the award is thus appropriate. Moreover, as I just discussed, just because some economic incentive already exists does not mean that enough exists without attorneys’ fees for the suit to be brought. Thus even on its own terms, the majority’s approach is too simple and, to be done properly, would involve us in complex analyses of economic actors’ decision-making.
*29B. Remaining Issues
There are several additional issues now before us which are amenable to discussion.
1. Environmentalist briefs duplicative of the EPA’s
On certain issues in this litigation environmentalists, as intervenor-respondents, filed briefs in favor of the EPA, and seek fees and costs from the EPA. The EPA opposes this claim. The majority would deny petitioners their request, and on this issue I concur.
It is true that the prevailing party in a suit against the EPA is entitled to costs and attorneys’ fees. Here, however, although an award to intervenor-respondents would be to the “right" party, the payment would be coming from the “wrong" party. The EPA should have to pay intervenor-respondents only if the positions briefed by the latter — and not by the EPA — were those adopted by the court, or were otherwise instrumental in the court’s ultimate disposition of the case.111 That is, the EPA should have to pay intervenor-respondents only if it were really they, and not the EPA, who prevailed. As the EPA puts it in its response to the environmentalists’ motion, “The government has” paid its own attorneys once to brief and argue these points and . .. there are no good reasons why the government should pay twice to defend its position on these points.” 112
The environmentalists argue that an award is appropriate even where the briefs were duplicative and unadopted since, in complex litigation like this, a certain amount of bolstering is to be encouraged. Of course, this is answered by noting that the standard advocated in the preceding paragraph would allow recovery of attorneys’ fees when it is shown to have been helpful. For those arguments that were not demonstrably helpful, however, we should decline to make an award. Note that were these additional arguments costless, or even were the EPA a limitless source of funds, it might be tempting to grant petitioners’ request. However, we live in a world of scarce resources and the question inevitably becomes how best to allocate them. In the issue at hand the question is, who is better able to decide how to allocate the EPA’s resources: the EPA or the Sierra Club? For the money which the EPA must pay the environmentalists for their bolstering cannot be spent by the EPA for those actions it would otherwise undertake in implementing the Clean Air Act. The environmentalists would then be deciding whether the EPA was spending enough money on enforcing a given issue.
Since the environmentalists intervened in favor of the EPA, the issue here is not whether the EPA is properly implementing the Clean Air Act. Yet, as discussed above,113 this was a major rationale of the citizen’s suits provisions. Of particular interest is an, exchange which occurred on the Senate floor between Senators Hruska and Muskie. Senator Hruska declared that he had grave reservations regarding the citizen’s suits provision of the 1970 Amendments, fearing that they would encourage profuse litigation and that it assumed that the EPA would not do its job.114 Senator Muskie responded that such suits were limited to enforcing provisions of the Act and could be brought only after the citizen had brought his intention to the attention of the *30EPA.115 This exchange implies that the EPA is not to be punished except when it neglects to bring suits it should have. Obviously, that is not the case here. Moreover, Senator Muskie’s further remarks 116 indicate that he expected that most citizens would quit their suits once the EPA was prompted to bring its suit; under these circumstances it seems unlikely that he would have foreseen the EPA paying the litigation costs of such a duplicate suit. Thus, while Congress intended to encourage private parties to make sure the Act was enforced, it did not intend to encourage litigation second-guessing the EPA on how best to allocate its (demonstrably correct) enforcement expenditures.117
In Citizens Association of Georgetown v. Washington,118 this court declined to require defendant District of Columbia to pay plaintiffs’ attorneys’ fees under § 7604(d) since that provision “ ‘was hedged by limitations — the confinement to clear-cut violations by polluters’ ”119 and there was no allegation that the District was itself a polluter, only that it had failed to enforce the Clean Air Act. The case at hand, of course, is under § 7607(f), which specifically does contemplate awards against the EPA when it fails to enforce the Act. However, even that allegation is not made here. Petitioners not only concede that the EPA was right, but were in fact on EPA’s side throughout. Thus, the analogy with Citizens Association seems entirely accurate. There we held that, to collect an award, it must at least be alleged that the party to be billed is the “right party” — a party alleged to be at fault under the statute — to be made to pay. Since here the EPA is the “wrong party,” an award is inappropriate under the rationale of Citizens Association.
Sierra Club endorses a number of criteria which bear on the nature of the aid rendered by petitioners: the helpfulness of the contribution, of what quality the briefs were, as well as whether the aid was duplicative.120 As just discussed, I agree in principle with the duplicative standard and, for the same reasons, I suppose I would ceteris paribus want to encourage “helpful” and “quality” briefs, though there seems to be some redundancy here (is a duplicative brief ever helpful?). But all of this begs the more basic question in this case: under what circumstances did Congress intend to encourage any brief — even a helpful, nonduplicative, quality one — in a losing cause? On this issue these second-level criteria are of little use. After all, for example, we would reach the decision of frivolous vel non without taking into account how well-written an allegedly frivolous brief was.
Requiring the EPA to reimburse all intervenor-respondents will have three additional ill effects. First, it will encourage the EPA to withdraw from cases in which there had been intervention. This seems particularly indefensible in light of the purpose of the citizen’s suit provisions to encourage government involvement and enforcement. Second, it will encourage parasitic intervention by parties having nothing to add but eager to collect costs and fees in suits where the EPA was perceived as likely to prevail. Third, it will encourage additional filings of dubious value in suits already of notorious complexity. Limiting awards as we have indicated today would certainly discourage the latter two temptations; as for the first, the EPA would withdraw only if it were making arguments it thought the court less likely to adopt than intervenors’, a palatable result. Intervenor-respondents would then probably make the EPA’s arguments in the alternative, in which case very little would be lost indeed.
A particularly troublesome aspect of the opinion in Sierra Club is its assumption that the environment has no protectors in the *31Environmental Protection Agency. It is “absolutely essential,” opines the panel, “in a case of this dimension that this court have expert and articulate spokesmen for environmental as well as industrial interests,” 121 thereby implying at best that the EPA represents neither, and at worst that it is an industrial stooge. The truth of the matter is that the EPA has been entrusted by Congress to protect the environment and, in the absence of a convincing reason to the contrary, it ill-behooves a court to require the EPA to subsidize those who thought — in this case, generally wrongly— that they could do its job better.
In summary, an award of attorneys’ fees or costs to intervenors is inappropriate — unless intervenor-respondents show that the positions briefed by them, and not the EPA, were those adopted by the court, or that their brief was otherwise instrumental in the court’s disposition of the case.
2. Other Issues
On another issue now before us and amenable to resolution, the EPA argues that it would be inappropriate to award compensation to environmental petitioners for time spent in researching an ex parte issue which was later withdrawn. Clearly, petitioners cannot be said to have prevailed on this matter and, moreover, it seems particularly unlikely that Congress intended to encourage abortive actions of this sort. Accordingly, an award of costs or fees would be inappropriate. Even assuming arguendo that we should apply Sierra Club’s criteria, it is hard to justify the majority’s award here. The ex parte issue was neither novel, nor complex, nor important, nor was the work done by petitioners on the issue particularly helpful or skillful — in short, under any circumstances an award would be inappropriate.
Another specific issue the parties have already brought before us involves a 25 July 1979 order of this court. At that time we ordered environmental petitioners to respond to industry petitioners’ motion for rehearing on the issue of whether any increase in emissions from a major emitting facility constitutes a “modification” subject to PSD. In our final opinion we found that it did,122 and rejected EPA’s 100/250 ton threshold for determining whether a “modification” was subject to PSD review. The court’s decision reflected the environmental petitioners’ position on this issue. Therefore, in light of the court’s 25 July 1979 order and its final ruling on this point, environmental petitioners can fairly be said to have prevailed on this issue. For this reason an award to the Sierra Club reflecting the attorney time and costs associated with its efforts is appropriate, and I concur with the majority’s disposition of the issue.123
The EPA also objects generally to environmental petitioners’ requests for reimbursement for money spent on travel, meals, and postage because no bills were presented. However, it is acceptable here to rely upon the integrity of counsel. A similar issue arises from the Sierra Club’s opposition to the EPA’s petition for a stay in the issuance of mandate. With respect to this matter, the EPA submits “that an award of reasonable counsel fees is not inappropriate, but that the amount requested by the Sierra Club is not reasonable.”124 Here again, however, the integrity of counsel may be relied upon, though it should still be necessary for petitioners to demonstrate that the reimbursement requested stems from expenditures made on issues on which they prevailed.125
One additional, logically last, question is raised by environmental petitioners. They request costs and attorneys’ fees for work *32done in connection with their request for costs and attorneys’ fees. Again, the appropriate response seems to be to allow such an award for expenditures made in pursuit of claims eventually granted — which means, were my dissent controlling, those claims for costs and fees on issues where petitioners prevailed on the merits.
IV. CONCLUSION
It is impossible to review the attorneys’ fees statute and its legislative history without being struck by the absence of any clue as to the meaning of “appropriate.” Congress clearly failed to give us a judicially cognizable definition of the word beyond the “prevailing” standard, and I would decline as inconsistent with the judicial role Congress’ implicit invitation to supply one. The majority here and in Sierra Club v. Gorsuch do their best to conjure criteria from the vacuum left us by Congress, but the effort ultimately fails. Even were I to accept as correct the inclination to try to fashion something from nothing, the Sierra Club approach must be adjudged a failure. The diverse and vague criteria mentioned there would justify an award to nearly anyone, and in all the confusion it is nowhere intimated how the various factors are to be weighted. And whatever criteria are settled upon, surely they should be applied on an issue-by-issue basis.
Accordingly, I dissent.
APPENDIX to dissent
ISSUES NON-PREVAILING PREVAILING PARTY PARTY
(1) potential to emita (small) industry EPA
(2) general exemption for stationary sources emitting less than 50 tons pollutant per year b environmentalists EPA
■protection of increments
(3) a. prevention and correction of violation of increments c EPA industry
(4) b. guidelines for states regarding increments d EPA environmentalists
(5) sources located in non-attainment areas e industry EPA
(6) regulation of fugitive emissionsf environmentalists EPA, industry
monitoring
(7) a. § 165(e)(1) g (mixed outcome: EPA, industry, and environmentalists each got something)
(8) b. § 165(e)(2)h environmentalists EPA
(9) c. state exemption authority guidelines i environmentalists EPA
(10) d. post-construction monitoringj EPA environmentalists
(11) baseline date k industry, District of Columbia, Texas EPA
(12) baseline and voluntary fuel switchesl EPA Texas, industry (chemical manufacturing and utilities)
(13) modeling"m EPA industry
(14) stack heightn EPA industry
source definition
(15) a. “equipment,” “operation,” and “combination thereof” inclusion o industry EPA
(16) b. contiguity and common ownership extension p (mixed outcome: EPA and industry each got something)
(17) c. PSD extension to all sources q EPA industry
(18) major modificationr environmentalists EPA
bubble
(19) a. mandated by statute s EPA environmentalists
(20) b. PSD review when level unchangedt industry EPA
*33ISSUES_ NON-PREVAILING PREVAILING PARTY_PARTY
(21) low emission exemptionu District of Columbia EPA
(22) pollutant types regulated v EPA industry
(23) BACT definition to include visible emission standard w EPA industry (American Iron and Steel Institute)
(24) commenced construction x EPA industry (utilities)
. 672 F.2d 33 (D.C.Cir. 1982). Sierra Club, also decided today, involves the award of attorneys’ fees under the same statute as the one before us.
. 636 F.2d 323 (D.C.Cir.1979).
. Pub.L. No. 95-95, 91 Stat. 685 (1977) (hereinafter “the 1977 Amendments”).
. 42 U.S.C. § 7607(f) (Supp. III 1979) (emphasis added) (hereinafter “§ 7607(f)”).
. The decision of this circuit in Metropolitan Washington Coalition for Clean Air v. District of Columbia, 639 F.2d 802 (D.C.Cir.1981) recognized that Congress intended a prevailing-plus standard. See note 34 infra.
But cf. Cong. Research Serv. Am. Law Div., 97th Cong., 1st Sess., The Clean Air Act in the Courts 143-44 (Comm.Print 1981).
. 42 U.S.C. § 7604(d) (1976) (hereinafter “§ 7604(d)”). The language of § 7604(d) was also repeated at 33 U.S.C. § 1365(d) (1976), the citizen’s suit provision of the Water Pollution Control Act Amendments of 1972. H.R.Rep. No.911, 92d Cong., 2d Sess. 133 (1972) made clear that the adoption of § 7604(d)’s language was intentional. This report made reference only to an intent to discourage frivolous suits, id. at 133-34, though S.Rep.No.414, 92d Cong., 1st Sess. 81 (1971) U.S.Code Cong. & Admin.News 1972, p. 3668 essentially parroted the somewhat broader passage quoted from S.Rep. No.1196, 91st Cong., 2d Sess. 38 (1970) at note 26 infra. Still, the legislative history of the Water Pollution Control Act Amendments indicates that even a prevailing plaintiff might not be entitled to costs and fees. S.Rep.No.414, 92d Cong., 1st Sess. 92, U.S.Code Cong. & Admin.News 1972, p. 3757 (1971) (“The [Senate] Committee [on Public Works] on October 13 approved, 12-3, an amendment offered by Senator Baker to eliminate from the citizen suit provision the mandatory award of litigation costs to a successful plaintiff”). For the reasons discussed at pp. 13-14 infra, however, we believe that in the statute before us today prevailing is sufficient for an award to be “appropriate.”
The “appropriate” standard has been incorporated into a number of other statutes as well. See 16 U.S.C. § 1540(g)(4) (1976); id. 33 U.S.C. § 1365(d); id. § 1415(g)(4); id. 1515(d); id. 42 U.S.C. § 300j-8(d); id. § 4911(d); id. § 5851(e) (Supp. III 1979); id. § 6305(d) (1976); id. § 7622(e) (Supp. III 1979) (also part of the 1977 Clean Air Act Amendments; its legislative history adds nothing to our understanding of § 7607(f)); id. § 8435(d); id. 43 U.S.C. § 1349(a). The interpretation of any one of these statutes, of course, depends on its statutory context, legislative history, role in its area’s scheme of enforcement, etc. None of these statutes or their legislative histories seem to shed any light on the meaning of the statutes before us today.
. Pub.L.No. 91-604, 84 Stat. 1676 (1970).
. 28 U.S.C. §§ 1920, 1923 (1976); Equal Access to Justice Act, Pub.L.No. 96-481, 94 Stat. 2325 (1980).
. The Equal Access to Justice Act, id., liberalizes awards to prevailing parties in suits against the Government. However, the presumption against awards to non-prevailing parties — the only difficult issue before us today— remains. Nor does the Act attenuate the principle in Alyeska Pipeline Serv. Co. v. Wilderness Soc’y, 421 U.S. 240, 95 S.Ct. 1612, 44 L.Ed.2d 141 (1975), of narrowly construing congressional authority to make awards (discussed at pp. 10, 17 and 19 infra), where the award would be to a non-prevailing party. It should also be noted that this Act had not been passed when the statutes here were being written, so Alyeska and pre-Alyeska law are in any event critical to understanding and interpreting the legislation before us.
. 421 U.S. 240, 95 S.Ct. 1612, 44 L.Ed.2d 141 (1975). It cited, inter alia, what is now 28 U.S.C. §§ 1920, 1923, 2412 (1976). Accord: F. D. Rich Co. v. Industrial Lumber Co., 417 U.S. 116, 94 S.Ct. 2157, 40 L.Ed.2d 703 (1974).
. Thus, it reversed Wilderness Soc’y v. Morton, 495 F.2d 1026 (D.C.Cir.1974) (en banc), which had concluded that attorneys’ fees were recoverable for public interest lawsuits even in the absence of statutory authorization. This circuit has since followed Alyeska in making awards under the Clean Air Act. See Citizens Ass’n of Georgetown v. Washington, 535 F.2d 1318, 1320 (D.C.Cir.1976); Montgomery Environmental Coalition v. Costle, 646 F.2d 595 (D.C.Cir.1981).
It was Alyeska, therefore, which provided part of the impetus for the passage of § 7607(f). See H.R.Rep.No.294, 95th Cong., 1st Sess. 337 (1977). The amendment of what is now 42 U.S.C. § 1988 (1976) was for the same reason. S.Rep.No.1011, 94th Cong., 2d Sess. 1 (1976) U.S.Code Cong. & Admin.News 1976, p. 5908. Even five years ago, one commentator estimated that there were by then at least 75 statutory grants of authority by Congress for courts to award attorneys’ fees. Berger, Court Awarded Attorneys’ Fees: What is ‘‘Reasonable’’?, 126 U.Pa.L.Rev. 281, 303-04 (1977).
. 421 U.S. at 269, 95 S.Ct. at 1627.
. Fed.R.Civ.P. 39; 28 U.S.C. § 2412 (1976); Equal Access to Justice Act, Pub.L.No. 96-481, 94 Stat. 2325 (1980).
. See, e.g., 42 U.S.C. § 2000a-3(b) (1976); id. § 2000e-5(k); Pub.L.No. 96-481, 94 Stat. 2325 (1980). See also 5 U.S.C. § 552(a)(4)(E) (1976) (“substantially prevail[ing]”).
. Most of the relevant legislative history for the 1977 and 1970 Amendments is conveniently collected in, respectively, Cong. Research Serv. Env. Policy Div., 95th Cong., 2d Sess., A Legislative History of the Clean Air Act Amendments of 1977 (Comm. Print 1978) (8 volumes) *11and Cong. Research Serv. Env. Policy Div., 93d Cong., 2d Sess., A Legislative History of the Clean Air Amendments of 1970 (Comm.Print 1974) (2 volumes). In this section I generally limit my discussion to the House and Senate Reports: the relevant floor debates are discussed in the next section (“C. Analysis,” pp. 13-24 infra).
It should be borne in mind that although we must of necessity rely heavily on the legislative history in interpreting the meaning of the word “appropriate,” the language there will not outweigh an actual statute. Thus, for instance, the statutes discussed earlier establishing a presumption against the award of attorneys’ fees to non-prevailing parties will not be overridden by the legislative history per se. Instead, legislative history is useful only to the extent that it defines the meaning of “appropriate.”
Such fees are already authorized to be awarded in suits brought under Section 304 [i.e., § 7604] of the Act.18
. H.R.Rep.No.294, 95th Cong., 1st Sess. (1977).
. Id. at 28.
. Id. at 337.
. Note that NRDC was endorsed only for the reasons it gave for why the simple “prevailing” standard should be rejected. Id. Even this limited endorsement was qualified and muddled by the word “largely.” Id. Thus, it is quite unclear to what extent Congress endorsed the “standards” that were adopted in NRDC, particularly those which seemed not to bear on that court’s reasons for its rejection of the prevailing standard (see, e.g., the second paragraph from the passage next quoted in the text). At any rate, to the extent that workable standards might be derived from the rather vague language of NRDC (“benefits ... conferred,” “allocate the costs of litigation equitably,” “encourage the achievement of statutory goals,” “earlier interpretation”), it is consistent with this dissent. See, e.g., pp. 18-21 infra. The majority’s and Sierra Club’s apparent interpretation of some of the vague language here in its broadest possible scope — amounting to a “non-frivolous standard” — is unacceptable for the reasons discussed at pp. 15-17 infra.
. NRDC v. EPA, 484 F.2d 1331, 1338 (1st Cir.1973).
On other grounds this circuit had “reluctantly” declined to follow the result of this case. NRDC v. EPA, 512 F.2d 1351, 1354 (D.C.Cir.1975).
. S.Rep.No.127, 95th Cong., 1st Sess. 99 (1977).
. H.R.Rep.No.564, 95th Cong., 1st Sess. 177 (1977) U.S.Code Cong. & Admin.News 1977, pp. 1502, 1557.
. See Metropolitan Washington Coalition for Clean Air v. District of Columbia, 639 F.2d 802 (D.C.Cir.1981). The legislative history of the 1977 Amendments makes clear that § 7607(f) was designed to achieve the same ends as § 7604(d). See especially H.R.Rep.No.294, 95th Cong., 1st Sess. 337 n.20 (1977); S.Rep. No.127, 95th Cong., 1st Sess. 99 (1977). The legislative history of § 7604(d) is therefore highly relevant to these proceedings. The majority and Sierra Club, however, fail to consider this adequately. See Sierra Club, at 35 n.3.
The Senate Report alludes to the judicial confusion over whether § 7604(d) itself could be applied to administrative proceedings. Compare NRDC v. EPA, 484 F.2d 1331 (1st Cir.1973) (such application is contemplated) with NRDC v. EPA, 512 F.2d 1351 (D.C.Cir.1975) (such application is not contemplated).
. See H.R.Rep.No.1783, 91st Cong., 2d Sess. 55 (1970).
. S.Rep.No.1196, 91st Cong., 2d Sess. (1970). It must be kept in mind that S.Rep.No.1196 reported on S.4358, which embodied not the “appropriate” standard but read, instead, “in the public interest.” Id. at 123; see also 116 Cong.Rec. 33104 (1970) (remarks of Sen. Hart). This further muddies the waters in this statutory pond. Are we to take this as evidence that “the public interest” is the intended standard, or precisely that it is not ? There is no evidence one way or the other.
There was yet another version of this bill in the Senate. See id. at 33098 (remarks of Sen. Cooper) (proposed statute provides for “the application of equitable principles in allocating such [attorneys’ fees and] costs to prevent injustice”).
. S.Rep.No.1196, 91st Cong., 2d Sess. 38 (1970).
. Id at 36-37.
. Id. at 65.
. H.R.Rep.No. 1783, 91st Cong., 2d Sess. 55, 56 (1970).
. See note 26 supra; 116 Cong.Rec. 42386 (1970).
. One instance where Congress certainly intended awards to be “appropriate” was to defendants in suits brought frivolously or in bad faith. Indeed, this was a major reason behind the attorneys’ fees and costs provisions in 1970 and 1977. See pp. 18-19 infra. However, in the case before us today respondents do not claim that petitioners’ actions were in any way frivolous or in bad faith.
. See pp. 14-15 infra (“1. Congress intended a ‘prevailing-plus’ standard”).
. S.Rep.No.1196, 91st Cong., 2d Sess. 38 (1970).
. See, e.g., NRDC v. EPA, 512 F.2d 1351 (D.C.Cir.1975); Citizens Comm. for the Columbia River v. Callaway, 494 F.2d 124 (9th Cir. 1974); NRDC v. EPA, 484 F.2d 1331 (1st Cir. 1973) (mentioned, as already discussed at length, by the House in its report); Delaware Citizens for Clean Air, Inc. v. Stauffer Chem. Co., 62 F.R.D. 353 (D.Del.1974), aff'd, 510 F.2d 969 (3d Cir. 1975). Accord: Washington Metropolitan Coalition for Clean Air v. District of Columbia, 639 F.2d 802 (D.C.Cir.1981) (if awards are to be made “whenever the underlying suit is a prudent and desirable effort to achieve an unfulfilled objective of the Act,” id. at 804, how could any successful suit fail to qualify?).
. H.R.Rep.No.294, 95th Cong., 1st Session 337 (1977) U.S.Code Cong. & Admin.News 1977, p. 1416 (emphasis added).
. Id. The “expressly rejected” amendment was in S. 252, 95th Cong., 1st Sess. § 36 (1977) and S. 3219, 94th Cong., 2d Sess. § 35 (1976). See also Staff of the Sen. Subcomm. on Env. Pollution, 95th Cong., 1st Sess., A Section-by-Section Analysis of S.252 and S.253 36-37 (Comm.Print 1977); S.Rep.No.717, 94th Cong., 2d Sess. 82-83, 225 (1976); 122 Cong.Rec. 23834 (1977) (remarks of Sen. Buckley).
. 484 F.2d at 1338.
. S.Rep.No.127, 95th Cong., 1st Sess. 99 (1977).
. H.R.Rep.No.564, 95th Cong., 1st Sess. 177 (1977) U.S.Code Cong. & Admin.News 1977, p. 1557.
. S.Rep.No.1196, 91st Cong., 2d Sess. 65 (1970).
. Cf. Cong. Research Serv. Am. Law Div., 97th Cong., 1st Sess., The Clean Air Act in the Courts 143-44 (Comm.Print 1981) (indicating that prevailing vel non is the major factor in determining “appropriate”).
. S.Rep.No.1196, 91st Cong., 2d Sess. 38 (1970) (emphasis added).
. 484 F.2d 1331 (1st Cir. 1973).
. H.R.Rep.No.294, 95th Cong., 1st Sess. 337 (1977). See note 19 supra.
. 484 F.2d at 1338 (emphasis added).
. 672 F.2d 33 (D.C.Cir. 1982).
. “A number” is an understatement. Here are some of the criteria for awards listed in Sierra Club:
—“ ‘encouraging] litigation which will assure proper implementation and administration of the act or otherwise serve the public interest.’ ” Id. at 35.
—“ ‘allocating] the costs of litigation equitably, to encourage the achievement of statutory goals ... [especially when] the government is attempting to carry out a program of such vast and uncharted dimensions ... [,] the legislation is itself complex and novel ... [, or] its early interpretation is desirable’.” Id. at 36.
—“not . .. merely . . . ‘who won’ ... [but the] benefits conferred by the litigation.... ” Id. at 36.
—“whether the litigation — successful or not — furthered the goals of the Act.” Id. at 36.
—“ ‘whether the suit was of the type Congress intended to encourage ... [or] was a *16prudent and desirable effort to achieve an unfulfilled objective of the Act.’ ” Id. at 36-37. See also id. at 38 n.7.
—“where the non-prevailing parties had a ‘well-founded expectation’ of success when their suit was brought — the issues were not frivolous but substantial — and where the appeal furthered the gpals of the Act by facilitating the prompt resolution of the important and complex issues confronting this court involving the Act’s interpretation ... [, or where], by assisting judicial interpretation .... agency implementation and Congressional reevaluation [was aided] .... ” Id. at 37.
—“whether [petitioners] have served the goals of the Act.” Id. at 38.
—“encouraging] the participation of ‘public interest’ groups in resolving complex technical questions and important and difficult questions of statutory interpretation, and in monitoring the prompt implementation of the Act.” Id. at 38.
—“[whether the parties] substantially contribute[d] to the goals of the Act; [whether:] the issues they addressed were important, complex and novel; their assistance in the resolution of the issues was substantial and not duplicative of the efforts of other parties; and the caliber of their written and oral presentations was exemplary.” Id. at 38-39.
—“The Importance of the Case and the Issues Involved” Id. at 39-40.
—“[whether or not the issues were] remotely frivolous [or] deserved to have been aired, and having been aired will contribute both to the agency’s future efforts to implement the . .. Act and to Congress’ ongoing review of the Act.” Id. at 40.
—“The Substantial Nature of the Petitioners’ Assistance” Id. at 40-41.
—“In conclusion, we find that the express goals of the Clean Air Act — prompt resolution of serious questions of statutory interpretation and citizen participation in monitoring administration of the Act through enforcement suits — require that substantial contributions to significant litigation in furtherance of these goals be compensated.... Given the complexity of the subject matter, without competent representatives of environmental interests, the process of judicial review might have been fatally skewed.” Id. at 41. See also id. at 35 n.3.
—“The questions raised by [petitioners] needed to be resolved; yet no other party had a sufficient economic interest at stake to represent them.... [0]ne cannot expect that contributions as substantial as those made by [petitioners] would be made by public interest groups without some form of compensation.” Id. at 41.
Mercifully, Sierra Club makes a belated— and, I hope, successful — attempt to limit its damage by saying that it really embraces only one standard, that of “substantial contributfion]” to the goals of the Clean Air Act. Yet it immediately explains that this nebulous standard of uncertain origin itself embraces all the other “relevant factors,” thereby redoing the ephemerally undone damage. At 42 n.10.
. This conclusion was reached by another court in Delaware Citizens for Clean Air, Inc. v. Stauffer Chem. Co., 62 F.R.D. 353, 356 (D.Del.1974), aff'd, 510 F.2d 969 (3d Cir. 1975).
. 122 Cong.Rec. 23834 (1976) (remarks of Sen. Buckley).
. 421 U.S. at 263, 95 S.Ct. at 1624.
. Id. at 269, 95 S.Ct. at 1627 (emphasis added).
. Cf. pp. 29-31 infra (“1. Environmentalist briefs duplicative of the EPA’s’’).
. There is perhaps a second, discussed at note 57 and accompanying text infra.
. H.R.Rep.No.294, 95th Cong., 1st Sess. 337 (1977) U.S.Code Cong. & Admin.News 1977, p. 1416; S.Rep.No.1196, 91st Cong., 2d Sess. 38, 65 (1970). But note that the version of § 7604(d) reported on in the 1970 Senate Report actually read “in the public interest” instead of “appropriate.” See note 25 supra.
. See id.
. See Citizens Comm. for the Columbia River v. Callaway, 494 F.2d 124 (9th Cir. 1974); Berger, Court Awarded Attorneys’ Fees: What is “Reasonable’’?, 126 U.Pa.L.Rev. 281 (1977). Mr. Berger talks at length about how the award of attorneys’ fees may best serve the public interest, and there is never any suggestion that awards be made to non-prevailing parties. This is also true of Council for Public Interest Law, Balancing the Scales of Justice: Financing Public Interest Law in America (1976), which similarly limits its suggestion for awards to prevailing parties. Id. at 312-23, 348-49.
. H.R.Rep.No.294, 95th Cong., 1st Sess. 337 (1977) (emphasis added). See also S.Rep.No. 1196, 91st Cong., 2d Sess. 36-37 (1970) (“The provision in the proposed bill is carefully restricted to actions where violations of standards and regulations or a failure on the part of officials to act are alleged. ... [Citizen’s suits] should motivate governmental agencies charged with the responsibility to bring enforcement and abatement proceedings”); 116 Cong.Rec. 32903 (1970) (remarks of Sen. Muskie) (“Although the committee does not advocate these suits as the best way to achieve enforcement, it is clear they should be an effective tool”); id. at 32919-20 (remarks of Sen. Spong) (suits are “to assure enforcement” and “to complement and encourage the abatement activities of governmental agencies”); id. at 32925-26 (remarks of Sen. Hruska); id. at 32927 (remarks of Sen. Muskie) (“Citizens can be a useful instrument for detecting violations and bringing them to the attention of the enforcement agencies and courts alike”); id. at 33101-02 (statement of Sen. Scott).
The majority apparently concludes that “proper implementation” is part of what Congress wanted to add to “prevailing” to define “appropriate.” But, as discussed in the text, it seems impossible for a court to derive a distinction between the “proper implementation” standard and “prevailing.” Theoretically, Congress might have been so anxious to see that every valid portion of the Act was enforced that it was willing to err on the side of paying off losers just to encourage suits by potential winners. But there is nothing in the statute or its history to lend credence to this counter-intuitive interpretation. Seep. 19 infra.
. H.R.Rep.No.294, 95th Cong., 1st Sess. 337 (1977); S.Rep.No.1196, 91st Cong., 2d Sess. 38 (1970); 11 Cong.Rec. 32919 (1970) (remarks of Sen. Spong); id. at 32927 (remarks of Sen. Muskie); id. at 33103 (staff memorandum of Sen. Muskie); id. at 33104 (remarks of Sen. Hart). Indeed, the 1977 Amendments were specifically aimed at, inter alia, “authorizpng] courts to award reasonable attorneys fees to any party against whom EPA acts unreasonably in initiating an enforcement action.” H.R.Rep.No.564, 95th Cong., 1st Sess. 177 (1977) *19U.S.Code Cong. & Admin.News 1977, p. 1557. See also NRDC v. EPA, 484 F.2d 1331, 1337-38 (1st Cir. 1973); Delaware Citizens for Clean Air, Inc. v. Stauffer Chem. Co., 62 F.R.D. 353, 355 n.6 (D.Del.1974), aff'd, 510 F.2d 969 (3d Cir. 1975).
. At 35 n.3, 41.
. See, e.g., North Slope Borough v. Andrus, 515 F.Supp. 961, 964 (D.D.C.1981); North Slope Borough v. Andrus, 507 F.Supp. 106, 108 (D.D.C.1981).
. See pp. 15-17 supra (“a. Sierra Club and the non-frivolous standard").
. See 42 U.S.C. §§ 1988, 2000a-3(b), 2000e-5(k) (1976).
. 421 U.S. 240, 263-64, 95 S.Ct. 1612, 1624-1625, 44 L.Ed.2d 141 (1975) (emphasis added).
Article III of the Constitution establishes and distributes the “judicial power of the United States.”
These would perhaps not be “judicial” functions, because the judges would have no intelligible standards by which to decide.65
. Congressmen of such divergent views as David Stockman and Toby Moffett were on the 1977 bill’s House committee.
. C. Black, Decision According to Law 42-43 (1981) (emphasis in original).
. In addition to the cases discussed in the text, many other cases, involving a wide array of *21Supreme Court Justices in a variety of constitutional contexts, have alluded to the dangers, difficulties, and unconstitutionality of using the courts to make public policy. See, e.g., Sinking-Fund Cases, 99 U.S. 700, 718, 25 L.Ed. 496 (1879) (Waite, C. J.) (“One branch of the government cannot encroach on the domain of another without danger. The safety of our institutions depends in no small degree on a strict observance of this salutary rule”); Flast v. Cohen, 392 U.S. 83, 131, 88 S.Ct. 1942, 1968, 20 L.Ed.2d 947 (1968) (Harlan, J., dissenting) (“The powers of the federal judiciary will be adequate for the great burdens placed upon them only if they are employed prudently, with recognition of the strengths as well as the hazards that go with our kind of representative government”); United States v. Richardson, 418 U.S. 166, 192-93, 94 S.Ct. 2940, 2954, 41 L.Ed.2d 678 (1974) (Powell, J., concurring) (“It merits noting how often and how unequivocally the Court has expressed its antipathy to efforts to convert the Judiciary into an open forum for the resolution of political or ideological disputes about the performance of government. [Numerous decisions cited ranging from 1900 to 1937.] These holdings and declarations reflect a wise view of the need for judicial restraint if we are to preserve the Judiciary as the branch ‘least dangerous to the political rights of the Constitution .... ’ Federalist No. 78 ... ”); United States v. New York Tele. Co., 434 U.S. 159, 179, 98 S.Ct. 364, 375, 54 L.Ed.2d 376 (1977) (Stevens, J., dissenting) (“The principle of limited federal jurisdiction is fundamental .... [T]he Court’s rush to achieve a logical result must await congressional deliberation”); Cannon v. University of Chicago, 441 U.S. 677, 743, 99 S.Ct. 1946, 1981, 60 L.Ed.2d 560 (1979) (Powell, J., dissenting):
Rather than confronting the hard political choices involved [in adopting legislation], Congress is encouraged to shirk its constitutional obligation and leave the issue to the courts to decide. When this happens, the legislative process with its public scrutiny and participation has been bypassed, with attendant prejudice to everyone concerned. Because the courts are free to reach a result different from that which the normal play of political forces would have produced, the intended beneficiaries of the legislation are unable to ensure the full measure of protection their needs may warrant. For the same reason, those subject to the legislative constraints are denied the opportunity to forestall through the political process potentially unnecessary and disruptive litigation. Moreover, the public generally is denied the benefits that are derived from the making of important societal choices through the open debate of the democratic process.
. Hayburn’s Case, 2 U.S. (2 Dall.) 408, 410 n.(a), 1 L.Ed. 436 (1792) (emphasis in original). This case involved a refusal by members of the Court to give extrajudicial advice to Congress and the Secretary of War on pension applications. See also the Supreme Court’s celebrated refusal under Chief Justice Jay to advise President Washington informally on questions relating to the neutral status of the United States in the European war 1793. Correspondence of the Justices, Letter from Chief Justice John Jay and the Associate Justices to President George Washington, 8 August 1793.
. 424 U.S. 1, 96 S.Ct. 612, 46 L.Ed.2d 659 (1976).
. Id. at 120 (emphasis in The Federalist No. 47).
. The Federalist No. 78, at 471 (New American Library ed. 1961) (A. Hamilton).
. 415 U.S. 336, 94 S.Ct. 1146, 39 L.Ed.2d 370 (1974). Accord: FPC v. New England Power Co., 415 U.S. 345, 94 S.Ct. 1151, 39 L.Ed.2d 383 (1974), affirming 467 F.2d 425 (D.C.Cir.1972); Alumet v. Andrus, 607 F.2d 911 (10th Cir.1979); Electronic Ind. Ass’n v. FCC, 554 F.2d 1109 (D.C.Cir.1976); Nevada Power Co. v. Watt, 515 F.Supp. 307 (D.Utah 1981). See also National Cable Television Ass’n, Inc. v. FCC, 554 F.2d 1094 (D.C.Cir.1976).
. 415 U.S. at 337, 94 S.Ct. at 1147 (quoting what is now 31 U.S.C. § 483a (1976)).
. Id. at 341, 94 S.Ct. at 1149 (emphasis in original). Compare the words Justice Douglas was discussing with the “standards” enunciated in Sierra Club, relied on by the majority here.
. 295 U.S. 495, 55 S.Ct. 837, 79 L.Ed. 1570 (1935).
. 276 U.S. 394, 48 S.Ct. 348, 72 L.Ed. 624 (1928).
. 415 U.S. at 342, 94 S.Ct. at 1149 (emphasis added).
. The Court there found the words “public policy” not “relevant” to its case “whatever may be their ultimate reach.” 415 U.S. at 343, 94 S.Ct. at 1150. Our task is more difficult since those words would be relevant to this case — but by reading the word “appropriate” broadly to include them we would be creating an unconstitutional delegation. Following National Cable’s admonition, this we decline to do.
FEA v. Algonquin SNG, Inc., 426 U.S. 548, 96 S.Ct. 2295, 49 L.Ed.2d 49 (1976), distinguished National Cable, but in such a way as to underscore the latter’s applicability to our case:
Respondents rely on our decision in National Cable Television Assn. v. United States, ... to support their delegation doc*23trine argument. But we find that case clearly distinguishable from the one before us today. In National Cable Television, we held that the fees to be imposed on community antenna television systems should be measured by the “value to the recipient” even though the language of the general statute allowing fee setting by federal agencies [citation omitted] permits consideration not only of “value to the recipient” but also of “public policy or interest served, and other pertinent facts.” The Court’s conclusion that the words of the last-quoted phrase were not relevant to the CATV situation was apparently motivated by a desire to avoid any delegation doctrine problem that might have been presented by a contrary conclusion.... But what might be considered the open-ended nature of the phrase “public policy or interest served, and other pertinent facts” stands in contrast to [the present statute’s] more limited authorizations of the President to act only to the extent necessary to eliminate a threat of impairment to the national security, and [its] articulation of standards to guide the President in making the decision whether to act.
Id. at 559-60 n.19, 96 S.Ct. at 2302. Clearly, the statute before us more closely resembles the “open-ended” statute in National Cable than the “more limited” one in Algonquin.
. See, e.g., Greene v. McElroy, 360 U.S. 474, 79 S.Ct. 1400, 3 L.Ed.2d 1377 (1959); Kent v. Dulles, 357 U.S. 116, 78 S.Ct. 1113, 2 L.Ed.2d 1204 (1958); Ashwander v. TVA, 297 U.S. 288, 348, 56 S.Ct. 466, 483, 80 L.Ed. 688 (1936) (Brandeis, J., concurring) (and cases cited therein); Crowell v. Benson, 285 U.S. 22, 62, 52 S.Ct. 285, 296, 76 L.Ed. 598 (1932) (“When the validity of an act of the Congress is drawn in question, and even if a serious doubt of constitutionality is raised, it is a cardinal principle that this Court will first ascertain whether a construction of the statute is fairly possible by which the question may be avoided”) (footnote omitted); Sinking-Fund Cases, 99 U.S. 700, 718, 25 L.Ed. 496 (1870) (“It is our duty, when required in the regular course of judicial proceedings, to declare an act of Congress void if not within the legislative power of the United States; but this declaration should never be made except in a clear case. Every possible presumption is in favor of a statute, and this continues until the contrary is beyond a rational doubt”).
. 437 U.S. 153, 98 S.Ct. 2279, 57 L.Ed.2d 117 (1978). Accord: Busic v. United States, 446 U.S. 398, 100 S.Ct. 1747, 64 L.Ed.2d 381 (1980); Patrick v. Marshall, 460 F.Supp. 23 (N.D.Calif.1978).
. 437 U.S. at 194, 98 S.Ct. at 2301.
. Id. We recognize that courts do have expertise in awarding costs and fees, but we have been, and are, able to exercise that expertise only with clearly defined standards — e.g., who prevailed? See p. 20 supra. What are we to do if, as is quite plausible, we find that the other side has a valid claim that it is acting in the public interest, too?
. 5 U.S. (1 Cranch) 137, 2 L.Ed. 60 (1803).
. 437 U.S. at 194-95, 98 S.Ct. at 2301-2302.
. Cf. Citizens Comm. for the Columbia River v. Callaway, 494 F.2d 124 (9th Cir. 1974); Delaware Citizens for Clean Air, Inc. v. Stauffer Chem. Co., 62 F.R.D. 353 (D.Del.1974), aff'd, 510 F.2d 969 (3d Cir. 1975).
. See, e.g., Central Fla. Enters., Inc. v. FCC, 598 F.2d 37, 60-61 n.18 (D.C.Cir.1978), cert. dismissed, 441 U.S. 957, 99 S.Ct. 2189, 60 L.Ed.2d 1062 (1979):
We recall that the [Federal Communications] Commission’s license to define the public interest, although broad, is not unbounded. See NAACP v. FPC, 425 U.S. 662, 669 [96 S.Ct. 1806, 1811, 48 L.Ed.2d 284] ... (1976) (“the use of the words ‘public interest’ in a regulatory statute is not a broad license to promote the general public welfare,” instead these words “take meaning from the purposes of the regulatory legislation”).
. At 11.
. Id. (footnote omitted).
. Majority opinion, at 3 (footnote omitted).
. 484 F.2d at 1338-39.
. At 11 n.16.
. Perhaps the majority has made an issue-by-issue analysis and simply decided that an award is appropriate on each issue. If this is the case, then I think the majority should have outlined the reasons for its decision so that future courts might be guided. Here again, the number and complexity of the issues does not obviate the need for or value of explication— indeed, it heightens them.
. 28 U.S.C. § 2412 (1976). But see note 99 infra.
. At 6-7.
. The District of Columbia’s bill of. costs should, however, be revised so that it seeks costs from respondent Anne M. Gorsuch not in her individual capacity but rather in her capacity as administrator of the EPA.
. However, the amount awarded may be open to some dispute.
. See, e.g., Maher v. Gagne, 448 U.S. 122, 100 S.Ct. 2570, 65 L.Ed.2d 653 (1980); New York Gaslight Club, Inc. v. Carey, 447 U.S. 54, 100 S.Ct. 2024, 64 L.Ed.2d 723 (1980).
. Seep. 15 supra.
. See S.Rep.No. 1196, 91st Cong., 2d Sess. 38 (1977); Friends of the Earth v. Carey, 535 F.2d 165, 173 (2d Cir. 1976).
. After today’s opinions, it is unlikely that any petitioner will make this mistake again. The day of reckoning for the majority — on which they will have to explain why industry petitioners are or are not entitled to an award on the issues they lost — has been mercifully postponed. Even in the case before us, according to Respondents’ Response to Industry and Environmental Petitioners’ Motion for Attorneys’ Fees at 7, an industry petitioner at one point sought “to leave open the option of filing for attorneys’ fees pursuant to [§ 7607(f)] . ... ”
. At 5-7.
. The majority undertakes such an examination, id. at 5, but its diagnosis is assertive rather than informative.
. Id. at 3.
. Id. at 5-6.
. See p. 14 supra. Note particularly the difficulty of reconciling such a holding with last year’s Metropolitan Washington Coalition for Clean Air v. District of Columbia, 639 F.2d 802 (D.C.Cir.1981). See note 34 supra.
. 672 F.2d 42 (D.C.Cir. 1982). I note that in this case petitioners prevailed on two of the three issues. Moreover, the fact that petitioner did lose on even one issue reduced its award. Slip op. at 35. The majority’s reliance on EDF is thus rather tenuous. ous.
. See note 23 supra.
. See note 6 supra.
. Sierra Club, At 35 n.3.
. At 5. But cf. Sierra Club, at 28, 41 (implying that awards to environmentalists may be favored).
. Accord: majority opinion, at 5.
. See Delta Air Lines, Inc. v. CAB, 505 F.2d 386, 388 (D.C.Cir.1974) (appropriateness of awarding costs to intervenors generally governed by, “[b]eyond the obvious initial determination of whether the intervenor was on the winning or losing side, such other factors as the relative merit of the intervenor’s contribution, the novelty of the issues, [and] the necessity of intervention and the public interest, to name a few ...” (emphasis added); award made since “[i]n the instant case the intervenors were prevailing parties who substantially contributed to our resolution of the issues presented"). For reasons already discussed, the other factors mentioned in Delta are unsuitable for the case before us.
. Respondents’ Response to Industry and Environmental Petitioners’ Motion for Attorneys’ Fees at 10 (footnote omitted).
. See pp. 18-19 supra.
. 116 Cong.Rec. 32925-26 (1970).
. Id. at 32926-27; see also id. at 33103.
. Id. at 32927.
. Cf. Delaware Citizens for Clean Air, Inc. v. Stauffer Chem. Co., 62 F.R.D. 353, 357 (D.Del.1974), aff'd, 510 F.2d 969 (3d Cir. 1975).
. 535 F.2d 1318 (D.C.Cir.1976).
. Id. at 1321 (emphasis in original).
. See note 47 supra.
. At 41.
. 636 F.2d at 399-400.
. The EPA has stated that “respondents would not oppose an award to Sierra Club” on this specific matter. Respondents’ Response to Sierra Club’s Supplemental Motion for Allowance of Counsel Fees at 2.
. Id. at 3 (footnote omitted).
. The standards for this demonstration would be, however, liberal. See p. 27 supra.
636 F.2d at 352-55.
Id. at 355-61.
Id. at 361-63.
Id. at 363-64.
Id. at 364-68.
Id. at 368-70.
Id. at 371-72.
Id. at 372.
Id. at 372-73.
Id. at 373.
Id. at 374-76.
Id. at 376-81.
Id. at 381-88.
Id. at 388-94.
Id. at 395-96.
Id. at 396-98.
Id. at 398-99.
Id. at 399-400.
Id. at 400-02.
Id. at 402-03.
Id. at 403-05.
Id. at 405-06.
Id. at 407-09.
Id. at 409-11.