BREYER, Circuit Judge.

The Puerto Rican Cement Co. (the "Company") wishes to build a new cement kiln, replacing older kilns that it now operates at about 60 percent of their capacity. If operated to achieve about the same level of production, the new kiln will pollute far less than the older kilns; but, if the Company operates the new kiln at significantly higher production levels, it will emit more pollutants than did the older kilns. The Environmental Protection Agency, noting that it is possible that the new kiln will produce more pollution, has held that the Company cannot build it without obtaining a special kind of EPA approval, required when one wishes to "construct" a "major emitting facility" in a place where the air is particularly clean. (The facility must meet "prevention of significant deterioration" ("PSD") requirements. See 42 U.S.C. § 7475.) The Company appeals. We find that EPA's determination is lawful.

I. [2] Background

1. Factual: The Company's cement plant contains six kilns, which produce a fine powder called "clinker." In 1987 the Company decided to convert Kiln No. 6 from a "wet," to a "dry," cement-making process, and to combine that kiln with Kiln No. 3. At that time, Kilns 3 and 6 were operating at about 60 percent of their combined capacity, producing about 424,000 tons of clinker per year. The converted kiln would have a total capacity of 961,000 tons of clinker per year, or about 35 percent more than the 705,000 ton capacity of Kilns 3 and 6. At any given level of production, the new kiln would emit less air polluting substance than the two older kilns combined, and would use less fuel to boot. However, if the Company decided to operate the new kiln close to its capacity, it might produce both more clinker and more pollution than the old kilns produced when operated at 60 percent of their capacity. In particular, information submitted by the Company suggests the following:

Pounds of Emissions per Ton of Clinker Produced

Fig. 1: Comparative Emissions Rates

Fig. 2: Comparative Emissions Amounts

These charts show the rate and amount of emissions of three pollutants: nitrogen oxides, sulfur dioxide, and particulate matter. The "Actual" rate of production is the average rate for Kilns 3 and 6 for the years 1985-86, or 424,000 tons; the "Potential" rate equals 705,000 tons of clinker per year for the old wet process and 911,000 tons of clinker per year for the new dry process. The emboldened numbers are those used by EPA in comparing actual emissions of the old kilns with potential emissions of the proposed new kiln. The charts make clear that emissions will increase only if the company operates the new kiln at significantly higher production levels.

2. Legal: Since the cement plant is located near Ponce, Puerto Rico, where the air quality is better than national ambient air quality standards, new construction is subject to PSD provisions contained in Part C of Title I of the Clean Air Act. See 42 U.S.C. §§ 7470- 7479. That part of the Act says that "[n]o major emitting facility . . . may be constructed in any [such] area" without various specified studies, reviews, demonstrations of compliance with certain substantive standards, and the issuance of a permit. See 42 U.S.C. § 7475 (emphasis added). The Act defines "major emitting facility" as a "stationary source of air pollutants," including Portland Cement plants that "emit, or have the potential to emit, one hundred tons per year or more of any air pollutant" (such as the facilities at issue here). 42 U.S.C. § 7479(1). It defines "construction" to include "modification," which it says

means any physical change in, or change in the method of operation of, a stationary source which increases the amount of any air pollutant emitted by such source or which results in the emission of any air pollutant not previously emitted.

42 U.S.C. §§ 7411(a)(4), 7479(2)(C). The Act also provides that EPA itself must review the construction proposal and provide the necessary approvals where, as here, no EPA-approved "state implementation plan" is in effect. See 42 U.S.C. § 7478; 40 CFR 52.21(a).

Because the permitting process is costly and time-consuming, EPA has developed an informal system for determining whether or not a particular construction proposal does, or does not, fall within the scope of the PSD permit law. If EPA decides that PSD review is unnecessary, it issues a "non-applicability determination" (known as a "NAD").

3. Proceedings: On July 9, 1987, the Company asked EPA for a NAD. It submitted information to EPA over an eight-month period. On August 30, 1988, EPA denied the Company the NAD. The Company has appealed EPA's determination to this court. Subsequent to the docketing of this appeal the Company and EPA agreed that, if the Company loses this appeal, it will operate its new facility at a sufficiently low capacity to prevent any actual increase in emissions levels. EPA will then issue a NAD, see 40 CFR 52.21(b)(4) (federally enforceable limitations on emissions will be taken into consideration in determining "potential to emit"), but the Company will lose its right to ask for a PSD permit, thereby giving up the possibility of obtaining EPA's approval for an increase of emissions.

II. [9] Jurisdiction

The Company can appeal the EPA's decision denying a NAD only if that decision is a "final action of the administrator." 42 U.S.C. § 7607(b)(1); cf. 5 U.S.C. § 704 (specifying actions reviewable under the Administrative Procedure Act). As other courts have recognized, see Hawaiian Elec. Co. v. EPA, 723 F.2d 1440, 1442-44 (9th Cir. 1984), one might question the "finality" of such a decision either 1) because the agency must take further action to obtain an enforceable order (a problem of "ripeness"), or 2) because the Company can take further administrative steps ( i.e., it can invoke the PSD review process) and thereby perhaps obtain the permission to build that it seeks (a problem of "exhaustion of administrative remedies").

The first of these problems — that of "ripeness" — is not particularly serious here. Even though the NAD denial does not, by itself, order the Company to refrain from building (EPA would have to bring an enforcement action to stop the Company from building, see 42 U.S.C. § 7477), it is well established that "ripeness" turns not upon such formal considerations, but rather upon such functional considerations as "the fitness of the issues for judicial decision and the hardship to the parties of withholding court consideration." Abbott Laboratories v. Gardner, 387 U.S. 136, 149, 87 S.Ct. 1507, 1515, 18 L.Ed.2d 681 (1967). (But compare Justice Brandeis' now-outdated description of finality in United States v. Los Angeles Salt Lake R.R. Co., 273 U.S. 299, 309-310, 47 S.Ct. 413, 414, 71 L.Ed. 651 (1927)). Here, the EPA's position on the legal question (of PSD applicability) is final and authoritative; court review will not "deprive the agency of the opportunity to refine, revise or clarify the . . . matter at issue." Roosevelt Campobello Int'l Park Comm'n v. EPA, 684 F.2d 1034, 1040 (1st Cir. 1982). Moreover, the fact-based record makes the legal issue "sufficiently concrete" to permit a court's focused attention. Id. At the same time, to withhold review would work considerable hardship on the Company, forcing it either to abandon its building plans, to compromise them by agreeing to emissions limitations, or to engage in a long, costly PSD review process. Under these circumstances, we consider the issue sufficiently "ripe." See Abbott Laboratories, 387 U.S. at 148-49, 87 S.Ct. at 1515-16; Hawaiian Elec. Co., 723 F.2d at 1443. Cf. Roosevelt, 684 F.2d at 1040 (issue not "ripe" where agency may well take legal action that would moot the controversy).

The second problem is more serious. The Company, in a sense, may not yet have "exhausted" its agency remedies; in principle it could, by following the PSD review procedures, possibly obtain from EPA permission to build the new kiln and to operate it at whatever levels it wishes. Of course, it is most unlikely that EPA, in the process, will reverse its determination that PSD review applies to the kiln. But, that fact does not end the matter, both because the Company may obtain a form of building permission and because the Supreme Court has held that an "interlocutory" agency decision may not be sufficiently "final" to warrant review. The Court held that a roughly analogous type of agency decision — a Federal Trade Commission decision to initiate an expensive, time consuming agency proceeding against a company — was "interlocutory" and not "final" for review purposes despite the "substantial burden" that forced participation in the administrative proceeding would impose upon the company. FTC v. Standard Oil Co., 449 U.S. 232, 244, 101 S.Ct. 488, 495, 66 L.Ed.2d 416 (1980). And, in so holding, the Court noted that "'the expense and annoyance of litigation is "part of the social burden of living under government."'" Standard Oil, 449 U.S. at 244, 101 S.Ct. 495 (quoting Petroleum Exploration, Inc. v. Public Service Comm'n, 304 U.S. 209, 222, 58 S.Ct. 834, 841, 82 L.Ed. 1294 (1938) (quoting Bradley Lumber Co. v. NLRB, 84 F.2d 97, 100 (5th Cir. 1936))).

While we recognize the possible analogy to Standard Oil, we also recognize that legal doctrines concerning "finality," particularly those related (closely or distantly) to "exhaustion of remedies," are riddled with exceptions. See, e.g., 4 K. Davis, Administrative Law Treatise § 26:1, at 414 (1983) (The doctrine that administrative remedies must be exhausted is "false almost as often as [it is] true."). We believe this is a case to apply the exception, not the rule.

First, the legal question at issue — the applicability of PSD review — is plainly separable from, and therefore collateral to, all the matters that the agency would consider in a PSD review itself. The collateral nature of the issue diminishes the likelihood that further agency proceedings will make it unnecessary for a court to decide the issue (as does the fact that PSD review cannot give the Company precisely what it wants without a few conditions, such as a use of "best available" anti-pollution technology, that it may not want). See, e.g., City of New York v. Heckler, 742 F.2d 729, 736-37 (2d Cir. 1984) (court may waive statutory exhaustion requirement when factors such as collaterality, futility, and irreparable harm indicate waiver would be appropriate), aff'd, 476 U.S. 467, 106 S.Ct. 2022, 90 L.Ed.2d 462 (1986); Kuehner v. Schweiker, 717 F.2d 813, 822-25 (3d Cir. 1983) (Becker, J., concurring) (judicial waiver of statutory exhaustion requirement appropriate when unexhausted claim substantially collateral, agency has taken final position on claim, and requiring exhaustion would cause substantial hardship), vacated, 469 U.S. 977, 105 S.Ct. 376, 83 L.Ed.2d 312 (1984) (remanding case for reconsideration in light of new statute).

Second, the agency itself can waive "exhaustion" requirements. See Mathews v. Diaz, 426 U.S. 67, 76-77, 96 S.Ct. 1883, 1889-90, 48 L.Ed.2d 478 (1976); Mathews v. Eldridge, 424 U.S. 319, 328, 96 S.Ct. 893, 899, 47 L.Ed.2d 18 (1976); Dugan v. Ramsay, 727 F.2d 192, 194 (1st Cir. 1984). The EPA here has created an administratively separate agency decision making process for granting or denying NADs. The EPA has agreed with the Company to grant a NAD if and only if the Company not only promises not to increase emissions but also promises not to seek permission, through the PSD process, to increase emissions levels. And, the EPA has not raised any objection to our reviewing this case. We therefore find a waiver of whatever exhaustion requirements might otherwise apply.

Together these considerations mean that, whether one views the statutory "finality" problem through the lens of "ripeness," of "exhaustion of remedies," or of "interlocutory decision," the EPA determination before us is sufficiently "final" to warrant review under 42 U.S.C. § 7607(b)(1). See Hawaiian Elec. Co., 723 F.2d at 1442-44 (holding that the determination that a proposed change is a "major modification" requiring PSD review is a "final action" under § 7607(b)(1)).

III. [17] The Merits A. [18] Interpreting EPA's Regulations

The statute applies its PSD requirements to the Company's proposed modification of its kilns only if the modification will "increase the amount of any air pollutant emitted." 42 U.S.C. §§ 7411(a)(4), 7479(2)(C). In deciding whether or not the kiln conversion would result in such an increase, EPA calculated the actual historical amount of pollutants that Kilns 3 and 6 emitted in the past (which, under the regulations, equals the average emissions over the past two years, see 40 CFR § 52.21(b)(21)(ii)) and compared that with the amount of pollutants that the converted kiln would be capable of emitting in the future. Since the Company operated the kilns at only 60 percent of their capacity in 1985-86, the new kiln, though cleaner and more efficient, is obviously capable of emitting significantly more pollutants.

The Company argues that EPA's application of this "actual/potential" method of measurement to its proposed kiln modification represents an improper, arbitrary, and contradictory interpretation of EPA's own regulations. After reading the regulations themselves, we disagree.

First, the language and expressed intent of the regulations both support EPA's interpretation. The regulations provide that a "major modification," subject to PSD review, includes "any physical change in or change in the method of operation of a major stationary source that would result in a significant net emissions increase of any pollutant. . . ." 40 CFR § 52.21(b)(2)(i) (emphasis added). They go on to define "net emissions increase" as the amount by which the "sum of . . . any increase in actual emissions" (plus or minus other "contemporaneous" changes in emissions) "exceeds zero." 40 CFR § 52.21(b)(3) (emphasis added). And, most importantly for present purposes, they define the words "actual emissions" in a special way.

They state that

"[a]ctual emissions" means the actual rate of emissions of a pollutant from an emissions unit, as determined in accordance with paragraphs . . . (ii) through (iv) [below].

40 CFR § 52.21(b)(21)(i) (emphasis added). Paragraph (ii) says that

[i]n general, actual emissions as of a particular date shall equal the average rate, in tons per year, at which the unit actually emitted the pollutant during [the preceding] two-year period.

40 CFR § 52.21(b)(21)(ii). But, paragraph (iv) adds that

[f]or any emissions unit which has not begun normal operations on the particular date, actual emissions shall equal the potential to emit of the unit on that date.

40 CFR § 52.21(b)(21)(iv) (emphasis added). The regulations also define "emissions unit" to include " any part of a stationary source which . . . would have the potential to emit any pollutant." 40 CFR § 52.21(b)(7) (emphasis added).

The Company's proposed modified kiln is "part of a stationary source" and it has the "potential to emit" a pollutant. 40 CFR § 52.21(b)(7). EPA considered it to be an "emissions unit which has not begun normal operations." 40 CFR § 52.21(b)(21)(iv). It therefore counted as its "actual emissions," the modified kiln's "potential to emit" pollution, id., namely, in the case of SO[2], 1927 tons per year. See p. 293, supra. It counted the "actual emissions" of the existing kilns as "the average rate . . . at which" they "actually emitted the pollutant during the [preceding] two year period," 40 CFR § 52.21(b)(21)(ii), namely, in the case of SO[2], 1340 tons per year. See p. 293, supra. It therefore found an increase in what the regulations call "actual emissions" (1927 minus 1340 equals 587 tons per year). And, after setting off allowable contemporaneous changes, it found that the net increase was significantly greater than zero. See 40 CFR §§ 52.21(b)(2)(i), 52.21(b)(3)(i).

EPA's application of its regulation to the facts of this case complies with the expressed intent of the regulation's writers as well. In a preamble to the regulation, EPA says that, when calculating whether a physical change will bring about a significant net increase in emissions, "the source owner must [first] quantify the amount of the proposed emissions increase. This amount will generally be the potential to emit of the new or modified unit." 45 Fed.Reg. 52,677 (emphasis added).

In considering the lawfulness of an agency's interpretation of its own regulations, courts often give that interpretation "'controlling weight unless it is plainly erroneous or inconsistent with the regulation.'" Udall v. Tallman, 380 U.S. 1, 16-17, 85 S.Ct. 792, 801-02, 13 L.Ed.2d 616 (1965) (quoting Bowles v. Seminole Rock Sand Co., 325 U.S. 410, 414, 65 S.Ct. 1215, 1217, 89 L.Ed. 1700 (1945)); accord Donovan v. A. Amorello Sons, Inc., 761 F.2d 61, 63 (1st Cir. 1985). In this case, EPA needs little help from this principle, for both language and expressed purpose indicate that EPA applied the regulations properly.

Second, the Company argues that EPA's interpretation of the regulation is arbitrary — that the interpretation makes little sense because it would significantly discourage the Company, and others like it, from installing more efficient machinery that, at any production level, emits significantly less pollution. But we cannot agree. EPA has simply taken account of, and given controlling weight to, a different consideration: the fact that a firm's decision to introduce new, more efficient machinery may lead the firm to decide to increase the level of production, with the result that, despite the new machinery, overall emissions will increase. Indeed, EPA points out that a firm introducing such machinery can escape PSD review simply by promising that it will ensure its actual emissions do not in fact increase (that is, by promising that it will not run the machinery at such a rate as to create an actual increase in emissions levels). See 40 CFR 52.21(b)(4) (federally enforceable physical or operational limitations which effect emissions will be taken into consideration in determining "potential to emit").

One can imagine circumstances that might test the reasonableness of EPA's regulation. An electricity company, for example, might wish to replace a peak load generator — one that operates only a few days per year — with a new peak load generator that the firm could, but almost certainly will not, operate every day. And, uncertainties about the precise shape of future electricity peak demand might make the firm hesitate to promise EPA it will never increase actual emissions (particularly since EPA insists, as a condition of accepting the promise and issuing the NAD, that the firm also promise not to apply for permission for an actual increase under the PSD review process). Whatever the arguments about the "irrationality" of EPA's interpretation in such circumstances, however, those circumstances are not present here. The Company is not interested in peak load capacity; it operated its old kilns at low levels in the past; its new, more efficient kiln might give it the economic ability to increase production; consequently, EPA could plausibly fear an increase in actual emissions were it to provide the NAD. Thus, this seems the very type of case for which the regulations quoted above were written. We can find nothing arbitrary or irrational about EPA applying those regulations to the Company's proposal.

Finally, the Company points to another regulation with which, it argues, EPA's interpretation conflicts. That regulation says that

a physical change or change in the method of operation shall not include . . . an increase in the hours of operation or in the production rate.

40 CFR § 52.21(b)(2)(iii)(f). The Company notes that, given this regulation, it could increase production at its old kilns to 100 percent of capacity, thereby vastly increasing actual emissions; why, it argues, should it not be permitted to do the same by building a more efficient kiln and then increasing output?

The answer to this question likely lies in the statute itself, for the statute refers to the "construction" of facilities, not to increased use of existing facilities. See 42 U.S.C. § 7479(2)(C). It may also lie in a prediction that, as a general rule, new building will more likely lead to increased emissions levels. Regardless, there is no logical contradiction in rules that, on the one hand, permit firms using existing capacity simply to increase their output and, on the other, use the potential output of new capacity as a basis for calculating an increase in emissions levels. And, we can find no policy conflict sufficiently serious for a court to override the policymaking authority that Congress has entrusted to the agency.

B. [34] Inconsistency

The Company argues that EPA has interpreted its regulations inconsistently; it says that sometimes EPA has interpreted the words "emissions unit which has not begun normal operations" to include only new units, while here it has interpreted those words to include modified units as well. The Company points to the well-established legal doctrine that an agency "'must either follow its own precedents or explain why it departs from them.'" Shaw's Supermarkets, Inc. v. NLRB, 884 F.2d 34, 36 (1st Cir. 1989) (quoting 2 K. Davis, Administrative Law Treatise § 8:9 at 198 (1979)). And, it argues that EPA has provided no explanation for any such departure here.

We have examined the relevant agency materials that the parties have submitted, however, and we cannot find any significant conflict. First, the more official EPA documents — the regulations and the written materials explaining them — make clear that EPA does mean to include "modified units" in the category of units that have "not begun normal operations." The preamble to which we earlier referred, for example, says that the "amount of the proposed emissions increase" will "generally be the potential to emit of the new or modified unit." 45 Fed.Reg. 52,677 (emphasis added). Second, a number of EPA internal memoranda concerning specific projects clearly follow this interpretation. Third, two or three internal memoranda and NAD letters are ambiguous about whether modified units are, or are not, included. Fourth, as EPA concedes, one NAD letter clearly contains a contrary interpretation.

In our view, these materials do not show a significant, legally recognizable "conflict" within the agency for two reasons. First, the "deviant" interpretation occurs but once. EPA materials written both before, and after, the deviant letter are consistent with its present interpretation. As the Fifth Circuit stated in NLRB v. Sunnyland Packing Co., 557 F.2d 1157 (5th Cir. 1977):

[O]ne swallow doesn't make a summer, and one inconsistent precedent does not entitle an agency litigant to demand that the [agency] ignore prior and subsequent holdings which have followed a uniform approach. . . . [Plaintiff] must do more than point to a single potentially deviant precedent before the reviewing court can find such inconsistency in agency action as to constitute arbitrary treatment of litigants.

Sunnyland, 557 F.2d at 1160-61. Second, the NAD letters and internal memoranda were written by different regional administrators and division directors on different occasions. No large agency can guarantee that all its administrators will react similarly, or interpret regulations identically, throughout the United States. The purpose of the "consistency" doctrine in administrative law is not so much to assure that lower level officials will properly interpret an agency's well-established pre-existing policies as to prevent the agency itself from significantly changing those policies without conscious awareness of, and consideration of the need for, change. See, e.g., Atchison, Topeka Santa Fe Ry. Co. v. Wichita Bd. of Trade, 412 U.S. 800, 808, 93 S.Ct. 2367, 2375, 37 L.Ed.2d 350 (1973) (plurality opinion) (ground for departure from prior norms must be clearly set forth so that the reviewing court may understand the basis of the agency's action and judge the consistency of that action with the agency's mandate); Shaw's Supermarkets, 884 F.2d at 41 ("Unless an agency either follows or consciously changes the rules developed in its precedent, those subject to the agency's authority cannot use its precedent as a guide for their conduct; nor will that precedent check arbitrary agency action."); Chisholm v. Defense Logistics Agency, 656 F.2d 42, 47 (3d Cir. 1981) (obligation to explain departures from precedent is an aspect of requirement that agency not act arbitrarily or capriciously); Miner v. FCC, 663 F.2d 152, 157 (D.C. Cir. 1980) (agencies must explain departures from prior precedent because "'the Rule of Law requires that agencies apply the same basic standard of conduct to all parties appearing before them'") (quoting Teamsters Local Union 769 v. NLRB, 532 F.2d 1385, 1392 (D.C. Cir. 1976)). The material we have described shows no such change in EPA policy.

C. [39] Lawfulness of the Regulations

The Company argues that EPA's regulations, insofar as they apply the "actual/potential" method to plant modifications, fall outside the scope of the statute's regulation-writing authority. However, judicial review under these circumstances is governed by 42 U.S.C. § 7607(b), which provides that "[a]ny petition for review [of the lawfulness of a regulation] shall be filed within 60 days from the date notice of [the regulation's] promulgation . . . appears in the Federal Register." 42 U.S.C. § 7607(b)(1). EPA promulgated the regulations in question in 1980, see 45 Fed.Reg. 52,735 (1980); other parties have challenged their lawfulness in a suit filed in the Court of Appeals for the District of Columbia, see Chemical Mfrs. Ass'n v. EPA, No. 79-1112 (D.C. Cir.). The Company has not tried to intervene in that suit (which is still pending). It seems obviously too late for the Company to mount an independent legal challenge here. See Hawaiian Elec. Co., 723 F.2d at 1441 (holding that a challenge to another provision of 40 CFR § 52.21 was untimely under 42 U.S.C. § 7607(b)(1)).

Regardless, the regulations in question apply to

any State implementation plan which has been disapproved with respect to prevention of significant deterioration of air quality in any portion of any State where the existing air quality is better than the national ambient air quality standards.

40 CFR § 52.21(a) (emphasis added); see 42 U.S.C. § 7478. The list of states changes as implementation plans are approved and disapproved; as of July 1, 1988, for example, provisions of § 52.21 were applicable to numerous states other than Puerto Rico. See, e.g., 40 CFR §§ 52.144 (Arizona); 52.270 (California); 52.382 (Connecticut); 52.499 (District of Columbia); 52.632 (Hawaii); 52.738 (Illinois); 52.793 (Indiana); 52.931 (Kentucky); 52.1116 (Maryland); 52.1165 (Massachusetts); 52.1180 (Michigan); 52.1234 (Minnesota); 52.1280 (Mississippi); 52.1529 (New Hampshire); 52.1603 (New Jersey); 52.1689 (New York); 52.1884 (Ohio); 52.2178 (South Dakota); 52.2303 (Texas); 52.2451 (Virginia); 52.2497 (Washington); 52.2581 (Wisconsin); 52.2676 (Guam); 52.2779 (Virgin Islands); 52.2827 (American Samoa). These facts, in our view, mean that the regulations are "nationally applicable" and also "based on a determination of nationwide scope or effect." 42 U.S.C. § 7607(b)(1). Hence, the statute requires the Company to challenge their lawfulness in the Court of Appeals for the District of Columbia; it cannot proceed in this court. Id. (challenges to nationally applicable regulations must be brought in the District of Columbia Circuit; challenges to regulations of only local or regional applicability may be brought in any appropriate circuit.)

IV. [43] Credit for "Contemporaneous" Decreases in Emissions

The regulations, as we have previously mentioned, measure any increase in emissions by, first, calculating the "actual" increase in emissions, and second, offsetting any "contemporaneous" decrease in emissions, due, say, to other changes the firm has made at the plant. See pp. 296-97, supra. The Company undertook a coal conversion project in 1982-1983, which led to a significant decrease in emissions. The EPA refused to credit the Company with this decrease because, it found, the decrease was not "contemporaneous" with the present proposed project. The Company now argues that the EPA is wrong.

The EPA's regulations, however, make clear that the coal project was not "contemporaneous." They say that a decrease is "contemporaneous" if it occurs between

the date five years before construction on the particular change commences[,] and . . . the date that the increase from the particular change occurs.

40 CFR § 52.21(b)(3)(iii). Since construction on the kiln modification has not yet "commence[d]", and since more than five years has passed since the coal conversion, the Company cannot bring itself within this "contemporaneous" window. The Company says that it filed its NAD application within five years of the time it converted to coal, but that fact is irrelevant; the regulation speaks of " construction on the [kiln] . . . change," not of an application to make the change. 40 CFR § 52.21(b)(3)(iii). And, the history of the regulation, referring to an alternative, shorter (three year) window measured with respect to "the date an application was complete," makes clear that reference to a construction date (along with the longer five year window) was intended. See 45 Fed.Reg. 6803 (1980) (soliciting comments on proposed regulations defining "contemporaneous" for purposes of offsetting emissions).

Since the regulation is clear, since it does not count the 1982-83 coal conversion project as "contemporaneous," since the Company made no request of the agency to waive the rule, and since it cannot challenge the lawfulness of this "nationally applicable" regulation in this court, see pp. 299-300, supra; 42 U.S.C. § 7607(b)(1), we must reject its claim.

For these reasons, the petition for review is denied and the order of the United States Environmental Protection Agency is affirmed.