Appellants Coalition on Sensible Transportation ("COST") and other groups1 bring a number of challenges concerning a road construction project in Montgomery County, Maryland. The project will widen approximately sixteen miles of Interstate 270 ("I-270") and modify five interchanges along the way; it is expected to take five years and cost more than $113 million. COST claims that various federal, state, and local officials2 violated Sec. 4(f) of the Department of Transportation Act, 49 U.S.C. Sec. 303(c) (1982), the procedural requirements of Sec. 102(2)(C) of the National Environmental Policy Act, 42 U.S.C. Sec. 4332(2)(C) (1982), and the hearings requirement of Sec. 128 of the Federal-Aid Highway Act, 23 U.S.C. Sec. 128 (1982). It also contends that the district court erred in refusing to consider certain evidence and in denying a discovery request. The court below granted summary judgment for the defendants. Coalition on Sensible Transportation v. Dole, 642 F. Supp. 573 (D.D.C.1986). The district court granted a short stay pending appeal, which this court extended on February 11, 1987; after oral argument the court vacated the stay. For the reasons given below, we now affirm the district court's decision.

The stretch of I-270 at issue runs north from the Spur connecting I-270 to I-495 (the Washington Beltway). It is a heavily travelled route for traffic entering and leaving the District of Columbia and also for local traffic between and within the various nearby towns. Traffic conditions in this area are congested, and unchallenged administrative findings indicate that the population growth anticipated in the coming years is likely to add to the congestion. See Joint Appendix ("J.A.") vol. 3, 3-4, 6-7. Among the amenities of the area are its parks, several of which are presently bordered or crossed by I-270. These parks will be affected by the proposed widening.

Under Sec. 4(f) of the Department of Transportation ("DOT") Act, the Secretary may approve a use of significant public park land3 only if "(1) there is no prudent and feasible alternative to using that land; and (2) the program or project includes all possible planning to minimize harm to the park ... resulting from the use." COST brings essentially three challenges to the district court's decision. It contests (1) the court's findings that the I-270 project's effect on parks did not amount to a statutory "use"; (2) the court's acceptance of DOT's conclusions that there was no prudent and feasible alternative within the meaning of Sec. 4(f)(1); and (3) its acceptance of DOT's conclusion that harm to the parks could not have been minimized by adoption of a rejected alternative. We agree with plaintiffs' view of what constitutes a park use, but reject their claim that DOT neglected its Sec. 4(f) duties.

Before analyzing the adequacy of the administrative Sec. 4(f) analysis, the district court found that the effects of the I-270 project on adjacent land were too insubstantial to amount to a "use" within the meaning of Sec. 4(f). 642 F. Supp. at 595-96. We find that the court construed the term too narrowly and that the project triggered the statutory requirements.

At issue are four parks--Cabin John Regional Park, Middlebrook Hill Neighborhood Conservation Area, Metropolitan Grove Road Park, and Seneca Creek State Park. In none of the four are there popular facilities in the immediate vicinity of the highway, and the expansion project seems unlikely to change this state of affairs. In each park the expansion plan calls for "temporary construction easements," narrow strips of parkland adjacent to the roadway. The topography will be graded to facilitate use in the construction project, leaving slopes along the edge of the highway; at the conclusion of the project, these strips, landscaped and revegetated, will be returned to their governmental owners. The district court characterized the uses as temporary and sought to apply "constructive use" cases that made the finding of a Sec. 4(f) use contingent on "substantial" harm. Sierra Club v. Dep't of Transportation, 753 F.2d 120, 130 (D.C.Cir.1985); Adler v. Lewis, 675 F.2d 1085, 1092 (9th Cir.1982). Finding the harms not substantial, 642 F. Supp. at 596, the court found no Sec. 4(f) use.

In the circumstances of this case, it hardly seems necessary to explore the nuances of "constructive uses." The I-270 project is expected to last five years and to change about ten acres of parkland permanently. This is enough. The encroachment during the construction period might be a sufficiently substantial harm to trigger Sec. 4(f) despite its temporary nature, but we need not decide this in view of the permanent changes. These take two forms. First, the project will remove 50-year-old oaks that will take two generations to replace; as man measures time, this is permanent. Second, the parks' topography will be altered, presumably forever. The appellees seek to undercut the permanence of the changes by pointing to the promise of mitigation measures at the end of the project. While these bear on the satisfaction of Sec. 4(f)(2)'s requirement that harms be minimized, they do not help in determining whether a use exists.

This is not to say that every change within park boundaries constitutes a use. For example, in Sierra Club v. Dep't of Transportation, the court considered whether permitting limited commercial jet landings in an airport in Grand Teton National Park amounted to a use triggering Sec. 4(f). The agency had determined that the change would result in a cumulative noise increase that was not significant. 753 F.2d at 128. This court characterized the increase in flights as an "insignificant adjustment[ ]" and a "relatively minor change[ ] in the operational characteristics of an established transportation facility" because the land had been used as an airport for 43 years and had had some jet traffic for years. Id. at 130. The court then held that such adjustments did not fall within Sec. 4(f). We do not think that Sierra Club signals a new approach to the use issue, but rather embraces what has been characterized in other administrative contexts as a de minimis exception to the statute. See Monsanto Co. v. Kennedy, 613 F.2d 947, 955-56 (D.C.Cir.1979). At some point a constructive use becomes too constructive, and the Sierra Club court discerned such a point. The holding does not speak to the problem of direct, physical uses. The project here will use parkland.

Despite its viewpoint on the use question, the district court proceeded to a careful review of the administrative Sec. 4(f) findings and found no violation of the statute. We agree and in the two sections below address the claims based on Sec. 4(f)(1) and Sec. 4(f)(2).

Before the Secretary approves a park use, he must under Sec. 4(f)(1) find that there is no feasible and prudent alternative to use of protected lands. Here at least one alternative is concededly feasible, but DOT contends that none is prudent. In Citizens to Preserve Overton Park v. Volpe, 401 U.S. 402, 413, 91 S. Ct. 814, 822, 28 L. Ed. 2d 136 (1971) (emphasis added), the Supreme Court stated that the Department may reject alternatives as imprudent only if there are "truly unusual factors present in a particular case or the cost or community disruption resulting from alternative routes reach[es] extraordinary magnitudes." See also id. at 416, 91 S.Ct. at 823; 23 C.F.R. Sec. 771.135(a)(2) (1986).

DOT and the other appellees concede that the use of construction easements for the I-270 project could have been avoided entirely through the use of retaining walls. Appellees do not argue that the cost of these retaining walls would have reached "extraordinary magnitudes," though our calculations indicate they are by no means cheap: $7,901,000. Nor do appellees argue that the retaining walls would cause any community disruption. Overton Park, 401 U.S. at 413, 91 S.Ct. at 822. Rather, appellees argue that there are other "truly unusual factors" that justify rejection of the retaining walls alternative as imprudent.4 Brief for Appellees at 21-23.

Unlike the usual park-or-concrete choice posed by a Sec. 4(f) decision, the judgment here was essentially aesthetic. The DOT determination noted that the retaining walls of the alternative would be "long and high." J.A. vol. 2, N, 2. At oral argument it was asserted without contradiction that some would have reached 50 feet in height. At such points, a park user would look 50 feet down onto the highway. Highway users would have faced high walls for considerable stretches; so would park users, in those areas where the park surface was below the highway. DOT could well have regarded the results as grotesque. In the selected alternative, by contrast, regrading, reseeding and reforestation were possible and would produce "a similar edge of the park as exists today." Id. See also J.A. vol. 4, VII-4 to 7. In essence, the officials charged with the task concluded that the "use" alternative would do less harm to park values than would nominal "non-use." Bearing in mind the concern for preservation of parklands at the core of Sec. 4(f), we think the facts here qualify as "truly unusual."5

COST argues that there is inadequate record support for the key aesthetic judgment. Reply Brief for Appellants at 4. We disagree. Aesthetic determinations under Sec. 4(f) do not require a ponderous record. Resolution of such issues requires little or no quantitative data. Of course we must assure ourselves that the aesthetic determination was made seriously and not merely as a pretext to circumvent Sec. 4(f). Here the record indicates a serious consideration of the appropriate factors; there is no allegation of bad faith or impropriety, nor, so far as we can find, any clue to such behavior. In such circumstances "the court is not empowered to substitute its judgment for that of the agency." Overton Park, 401 U.S. at 416, 91 S.Ct. at 824.6 We therefore accept the administrative determination.

COST points out that the Department of Interior supported retaining walls. See J.A. vol. 4, IX-6. While this is relevant, it cannot be determinative: Congress vested the Sec. 4(f) decision in the Secretary of Transportation. Moreover, we note that local park officials, presumably no less sensitive than Interior and much more immediately concerned, supported the construction easement choice. J.A. vol. 4, IX-12, -27, -31, -42 to 43. Finally, we reject COST's suggestion that DOT erred in considering the comparative visual impacts on motorists. While highway users' interests in speed and convenience of transportation are not within the ambit of Sec. 4(f) protection, their aesthetic interests surely are. In short, we see no indication of an abuse of discretion.

Besides lacking feasible and prudent alternatives, a transportation project using protected lands must under Sec. 4(f)(2) include "all possible planning to minimize harm." The chosen version of the I-270 project provides for mitigation in the form of revegetation and landscaping. J.A. vol. 4, VII-9. The final plans also included construction of small retaining walls in Cabin John Regional Park and Seneca Creek State Park. See J.A. vol. 2, N, 3. COST argues, however, that the Sec. 4(f)(2) determination was inadequate, because the immediate predecessor to the final plan would have inflicted less harm on the parklands.

In response to community objections, the responsible officials adopted the prevailing plan from its predecessor simply by shifting part of the roadbed 24 feet to the west. The realignment reduced impacts on certain residences, completely avoided an anticipated encroachment onto Tilden Park, and slightly increased the encroachment onto Cabin John Regional Park. J.A. vol. 2, N, 2; vol. 4, VII-4 to 5. The district court found that the penultimate and ultimate alignments would cause substantially equal damage and that the responsible officials were therefore free to choose between them. 642 F. Supp. at 600.

Section 4(f)'s requirement that harm be minimized does not constrain the Secretary's choice between two plans causing substantially equal damage to Sec. 4(f) values. See Druid Hills Civic Ass'n v. Federal Highway Administration, 772 F.2d 700, 716 (11th Cir.1985) ("The Secretary is free to choose among alternatives which cause substantially equal damage to parks or historic sites."); Maryland Wildlife Federation, 747 F.2d 229, 236 (4th Cir.1984) (accord); Louisiana Environmental Society, Inc. v. Coleman, 537 F.2d 79, 86 (5th Cir.1976) (accord). The only issue, then, is whether the harm to parklands was substantially equal under both alternatives.

So far as park values are concerned, the two plans vary trivially. The selected alternative will use an additional half acre of Cabin John Regional Park, but the resulting increment of harm is negligible. See J.A. vol. 4, VII-2 (closest park facility to I-270 is a trail that comes within 300 feet of the highway); Plate 10 (illustrating location of construction easements). The predecessor would have entailed a small (.1 acre) encroachment in Tilden Park, which under the selected alternative is untouched. The differences in the harms from the two competing alternatives are so small that it was reasonable to consider them substantially equivalent.7

COST again contends that there is inadequate support for the determination. It is true that there is no formal determination of the equivalence of the harms at issue, but formal findings are not required in a Sec. 4(f) determination. Overton Park, 401 U.S. at 417, 91 S.Ct. at 824. But see Louisiana Environmental Society, 537 F.2d at 86 (suggesting that the Secretary erred in "never articulat[ing] any ... balancing of the relative harms"). The reviewing court must carefully review the record to insure that there was "consideration of the relevant factors and [no] clear error of judgment," 401 U.S. at 416, 91 S. Ct. at 824. Those criteria are satisfied. The record demonstrates consideration of the relevant factors (e.g., amount of parkland used, the location of current and planned park facilities, the availability of mitigation measures). See J.A. vol. 2, N, 2; id. vol. 4, VII-1 to 5. There is no clear error of judgment; indeed, COST does not even argue that the harm to Cabin John Regional Park under the adopted alternative was greater than the harm to Tilden Park under the other alternative. Under such circumstances, a reviewing court need not and should not "fly speck" Sec. 4(f) determinations for technical deficiencies. Louisiana Environmental Society, Inc. v. Dole, 707 F.2d 116, 122-23 (5th Cir.1983).

COST also contends that the responsible officials were obliged to prepare an environmental impact statement ("EIS") under the terms of Sec. 102 of the National Environmental Policy Act ("NEPA"), 42 U.S.C. Sec. 4332 (1982). This provision applies to "major Federal actions significantly affecting the quality of the human environment" and requires "a detailed statement" on

(i) the environmental impact of the proposed action,

(ii) any adverse environmental effects which cannot be avoided should the proposal be implemented,

(iii) alternatives to the proposed action,

(iv) the relationship between local short-term uses of man's environment and the maintenance and enhancement of long-term productivity, and

(v) any irreversible and irretrievable commitments of resources which would be involved in the proposed action should it be implemented.

An exception to this requirement applies when a less comprehensive environmental review, or environmental assessment ("EA"), provides a basis for a finding of no significant impact ("FONSI"). See 40 C.F.R. Sec. 1501.4(b), (c), & (e) (1986) (describing process); id. Sec. 1508.9 (defining EA); id. Sec. 1508.13 (defining FONSI); Sierra Club, 753 F.2d at 126. An EA and a FONSI, but no EIS, were prepared for the I-270 project.

The NEPA process involves an almost endless series of judgment calls. Here we consider ones relating to the detail in which specific items should be discussed and the agencies' treatment of the project's relation to other government activities. It is of course always possible to explore a subject more deeply and to discuss it more thoroughly. The line-drawing decisions necessitated by this fact of life are vested in the agencies, not the courts. The latters' "role ... is simply to ensure that the agency has adequately considered and disclosed the environmental impact of its actions and that its decision is not arbitrary or capricious." Baltimore Gas & Electric Co. v. Natural Resources Defense Council, Inc., 462 U.S. 87, 97-98, 103 S. Ct. 2246, 2252, 76 L. Ed. 2d 437 (1983).

In examining decisions not to prepare an EIS, we consider four factors:

First, the agency must have accurately identified the relevant environmental concern. Second, once the agency has identified the problem it must have taken a "hard look" at the problem in preparing the EA. Third, if a finding of no significant impact is made, the agency must be able to make a convincing case for its finding. Last, if the agency does find an impact of true significance, preparation of an EIS can be avoided only if the agency finds that changes or safeguards in the project sufficiently reduce the impact to a minimum.

Sierra Club, 753 F.2d at 127. Accord Natural Resources Defense Council, Inc. v. Herrington, 768 F.2d 1355, 1430 (D.C.Cir.1985); Sierra Club v. Peterson, 717 F.2d 1409, 1413 (D.C.Cir.1983).

COST relies on these standards in making three general challenges. First, it alleges that the appellees failed to take a hard look at or make the requisite findings regarding impacts on traffic, development, wildlife habitat, and wetlands. Second, it charges that the project was impermissibly segmented. Third, it challenges the district court's findings concerning cumulative significance. We consider each of these challenges in the subsections below.

A. Impacts on Traffic, Development, Wildlife Habitat and Wetlands

According to the FONSI, traffic conditions on I-270 are presently extremely congested, and development is expected to continue in the area. J.A. vol. 4, III-1 to 3, Plate 8. COST alleges that defendants never gave the I-270 project's impact on traffic patterns in the region the requisite hard look or found them to be insignificant.

We disagree. It is evident from the record that, just as one would expect where a road project was at issue, the traffic aspect of the plan received a hard look. Projections of traffic volume for the "design year" of 2010 were derived from development projections from various sources. See J.A. vol. 2, T, 2; vol. 3, 3-4; vol. 4, VIII-72. The EA and FONSI demonstrate careful attention to the present traffic conditions at interchanges and connecting roads and the effects of the project on these roads. Id. vol. 2, T, 3-4; vol. 3, 3-4, 31-45; vol. 4, Plate 8. Certain improvements, including ones to the Capital Beltway, were assumed as part of the study, and with these improvements the connecting roads were anticipated to perform adequately in the design year. Id. vol. 2, T, 2-4, V, 2-3; vol. 4, VIII-22, -24, -30 to 31.

COST's contentions to the contrary amount at most to technical disputes of the sort most fit for expert administrative resolution and judicial deference. See Baltimore Gas & Electric Co., 462 U.S. at 103, 103 S.Ct. at 2255. In such disputes, "[a]bsent a showing of arbitrary action, we must assume that the agencies have exercised [their] discretion appropriately." Kleppe v. Sierra Club, 427 U.S. 390, 412, 96 S. Ct. 2718, 2731, 49 L. Ed. 2d 576 (1976). See also Environmental Defense Fund, Inc. v. Higginson, 655 F.2d 1244, 1247 (D.C.Cir.1981) (noting that burden is on plaintiffs to show that policy on comprehensive EIS is arbitrary or capricious). COST's traffic experts disagree with the agencies' experts, but we find nothing in this disagreement suggestive of arbitrariness.

More interesting are COST's suggestions that the review documents overlooked the tendency of the project itself to generate traffic increases, both by attracting development to the area and by inducing some persons to make journeys that otherwise would not be made at all. See Brief for Appellants at 24. The EA and FONSI both predict that more cars will use I-270 itself under the "Build" than under the "No-Build" alternative, see J.A. vol. 3, 51; vol. 4, Plate 8, so there can be no claim that the appellees were blind to the prospect of the impact at issue. The dispute is about the cause of the increase. The responsible officials concluded that the project's partial cure of congestion on I-270 would lure additional travelers away from other roads, until a new equilibrium (with less aggregate congestion) was established. J.A. vol. 3, iv, 31-32. Responding specifically to comments at a public hearing that the improvement would "facilitate excessive levels of development," the FONSI replied that the region's development levels were constrained by "local and secondary roads," id. vol. 4, VIII-72, i.e., not primarily by the condition of I-270. It also noted that its traffic projections were based on development levels estimated by local and county agencies. Id.

Plaintiffs contest these findings solely with generalities. They note that it is "well established in traffic engineering that travel demand is a function of the adequacy of the road system," pointing to decisions to reduce shopping trips, for example, on account of traffic. J.A. vol. 1, G, 9. Similarly they observe that it is "commonly recognized that ... increased highway capacity will induce further development." Id., H, 5-6.

The appellees' analysis is not arbitrary and capricious. COST fights it with truisms, but makes no effort to establish that this project itself will generate major development or net traffic increments (i.e., increments exceeding relief elsewhere). Given the lengthy process of consultation with the localities affected, the documentation by the responsible officials of the effects in general, and the officials' plausible thesis as to the project's role in regional development, we find that the officials made a sufficiently convincing case for dispensing with an EIS in which to explore in detail the likelihood and scope of these effects.

Our conclusion is quite consistent with Appalachian Mountain Club v. Brinegar, 394 F. Supp. 105, 115 (D.N.H.1975). There the plaintiffs introduced evidence of traffic increases of up to 50 percent on a contiguous segment of the highway, while defendants maintained that the project would have only negligible traffic impacts. The district court noted that assessment of traffic data was the responsibility of the agency, not the courts, but found that the agency had entirely failed to conduct such evaluation. Here, by contrast, COST has offered only generalizations and the agency's consideration of traffic impacts is apparent.

Nor did the EA slight the project's impacts on wildlife habitat and wetlands along the right of way. The EA characterized the areas within the existing right-of-way fence as being of "marginal use as wildlife habitat," a statement of obvious and literal truth. J.A. vol. 3, 56. It noted that even areas beyond the fence had less habitat value than they would in undeveloped areas. Id. This conclusion is supported by a 15-page study of these impacts. J.A. vol. 2, G. In its reply brief at 12-13, COST for the first time takes issue with the second statement, characterizing it as one that would invariably excuse failure to analyze destruction near an existing highway. But plaintiffs point neither to any defects in the underlying study nor to any other fact drawing the final determination into question. We do not find it arbitrary or capricious.

COST asserts that the scope of the I-270 project was improperly defined for NEPA purposes. Agencies may not evade their responsibilities under NEPA by artificially dividing a major federal action into smaller components, each without "significant" impact. Taxpayers Watchdog, Inc. v. Stanley, 819 F.2d 294, 298 (D.C.Cir.1987); National Wildlife Federation, 677 F.2d at 890; Piedmont Heights Civic Club, 637 F.2d at 439. In considering the proper scope of the I-270 project, the district court quite properly referred to Federal Highway Administration regulations requiring EIS or FONSI evaluation of any project which

(1) Connects logical termini and is of sufficient length to address environmental matters on a broad scope;

(2) Has independent utility or independent significance, i.e., is useable and a reasonable expenditure even if no additional transportation improvements in the area are accomplished; and

(3) Will not restrict consideration of alternatives for other reasonably foreseeable transportation improvements.

23 C.F.R. Sec. 771.111(f) (1987). A number of courts have relied on the same or closely similar factors in their segmentation determinations. See, e.g., Taxpayers Watchdog, 819 F.2d at 298; Piedmont Heights Civic Club, 637 F.2d at 439; Swain v. Brinegar, 542 F.2d 364, 369 (7th Cir.1976) (en banc ).

In COST's view, the scope of the EA and FONSI should have been extended to include the Spur connecting I-270 to the Capital Beltway and certain interchange projects along the route. We address these in turn.

COST maintains that the Beltway was the only logical terminus of the project. First, however, we note that in the context of a highway within a single metropolitan area--as opposed to projects joining major cities--the "logical terminus" criterion is unusually elusive. Thus we join the court in Piedmont Heights Civic Club, 637 F.2d at 440, in assigning it modest weight and focussing more on "independent utility." Second, COST's critique appears to rest on a perception of the I-270 project as essentially a commuter road to and from the Beltway. The facts belie this Beltway-centric vision. Fully 45 percent of the traffic now using the road neither originates nor terminates at the Beltway. Thus the Beltway is no more logical as a terminus than the Spur. Third, COST does not deny that the scope chosen by appellees was enough "to address environmental matters on a broad scope" or that the I-270 project possesses independent utility. Finally, while COST argues that the scope chosen enables appellees "to avoid addressing the potentially significant impacts the Project will have" as it discharges vehicles into the Spur, the appellees point out that previously approved Beltway improvements, including improvements to the Spur, are premised on higher traffic levels broughts by I-270. J.A. vol. 2, V, 2-3. The officials' choice does not appear to us arbitrary or capricious. See National Wildlife Federation, 677 F.2d at 888-89.

COST suggests that because the interchange projects at MD 189, MD 124/117, and I-370 will facilitate traffic movement along the I-270 corridor, they have little independent utility or independent significance apart from the I-270 project, and should therefore have been examined in the same EA. Brief for Appellants at 33-34. But it is inherent in the very concept of a highway network that each segment will facilitate movement in many others; if such mutual benefits compelled aggregation, no project could be said to enjoy independent utility. The proper question is whether one project will serve a significant purpose even if a second related project is not built. See Taxpayers Watchdog, Inc., 819 F.2d at 299 (finding independent utility in four-mile section of mass transit project originally planned as 18.6 miles); Piedmont Heights Civic Club, 637 F.2d at 440-41 (urban highway projects, although related to overall transportation plan, individually contribute to improving traffic conditions); College Gardens Civic Ass'n, Inc. v. Department of Transportation, 522 F. Supp. 377, 386 (D.Md.1981); see also Citizens for Balanced Environment and Transportation, Inc. v. Volpe, 376 F. Supp. 806, 813-14 (D.Conn.1974) (Newman, J.), aff'd, 503 F.2d 601 (2d Cir.1974), cert. denied, 423 U.S. 870, 96 S. Ct. 135, 46 L. Ed. 2d 100 (1975). The record clearly indicates that the highway and interchange projects here would serve such purposes in the absence of the I-270 expansion, and thus are sufficiently independent. They are expected to result in less congestion at interchanges, facilitate local traffic, and provide access to mass transit. J.A. vol. 2, R, 2-5; Supplemental Administrative Record ("Supp. A.R.") G, 5-7 (EA for MD 189 interchange); Supp. A.R. AA, II-1 to 2 (EA for MD 124 interchange project); Supp. A.R. D, i-ii (EA for I-370 project). Conversely, the widening of I-270 will reduce congestion regardless of the fate of the interchanges.

COST further claims that the I-270 widening will restrict consideration of alternatives for related projects. Like the mutual-benefit argument considered above, the statement is always true to a degree; future road projects will have to take into account the traffic facilitated by any given project. This is not enough by itself to render a FONSI inadequate. Only when a given project effectively commits decisionmakers to a future course of action will this form of linkage argue strongly for joint environmental evaluation. See National Wildlife Federation, 677 F.2d at 890-91; Indian Lookout Alliance v. Volpe, 484 F.2d 11, 14 (8th Cir.1973).

Such a commitment could occur for either of two reasons. First, a project might appear to be an unreasonable expenditure of funds unless some related project were also completed. See 23 C.F.R. Sec. 771.111(f)(2). But once the project was itself completed, its costs would be sunk, and presumably highway officials would not proceed with the second merely to vindicate their decision on the first. Second, completion of the first project may cause the benefit/cost ratio on the second to rise sharply. Thus if a two-lane road between two cities is widened to four lanes for half its length, the return from investment in completion of the project may be usually favorable and its approval by rational decisionmakers foreordained. Here, however, we see no evidence that completion of I-270 would compel adoption of any of the interchange projects.8

A separate but related claim is that the defendants failed to analyze the cumulative impact of the project together with those of various related interchange and Spur improvements. Of course one might view this argument as simply a replay of the improper-segmentation argument already considered. Council on Environmental Quality ("CEQ") regulations, however, provide a distinct meaning to the concept. They define cumulative impact as "the incremental impact of the action [at issue] when added to other past, present, and reasonably foreseeable future actions...." 40 C.F.R. Sec. 1508.7. See also id Sec. 1508.27(b)(7) (cumulative significance discussed as one of ten factors to be considered in gauging intensity for purposes of determining significance). We believe the defendants followed the approach indicated by the regulations.9

There are nine existing or planned interchanges within the affected stretch of I-270. Five interchange projects (at Montrose Road, MD 28, Middlebrook Road, MD 118, and MD 121) were considered part of the I-270 project and evaluated together with the road widening in the same EA and FONSI. Interchanges at four other locations (MD 189, Shady Grove Road, I-370, and MD-124-117) were planned before the I-270 project, and although discussed in the EA and FONSI, received detailed treatment only in other EA's, FONSI's and EIS's. See J.A. vol. 2, R, 2-5. Beyond the I-270 project's limits, previously approved improvements to the Spur which links I-270 and the Capital Beltway were also considered as part of the background in the EA and FONSI. Id. T, 2-4, V, 2-3. COST contends appellees should have assessed the cumulative impacts of the MD 189, I-370, MD 124-117 and Spur projects in greater detail. See Reply Brief for Appellants at 14 n. 21.

We think this misconceives both the CEQ regulation and the purpose of NEPA. It makes sense to consider the "incremental impact" of a project for possible cumulative effects by incorporating the effects of other projects into the background "data base" of the project at issue, rather than by restating the results of the prior studies. Piedmont Heights Civic Club, Inc. v. Moreland, 637 F.2d 430, 441-42 (5th Cir. Unit B Feb. 1981). Cf. National Wildlife Federation v. Appalachian Regional Commission, 677 F.2d 883, 889-90 (D.C.Cir.1981) ("when a NEPA challenge is leveled against some subsequent phase of a continuing federal action, the EIS obligation attaching at this latter point is realistically qualified by the elements of the program already in place"); 40 C.F.R. Sec. 1501.7(a)(3) (1986) (part of the "scoping process" is to "[i]dentify and eliminate from detailed study the issues ... which have been covered by prior environmental review"). The record satisfactorily demonstrates that appellees followed this approach. See J.A. vol. 2, T, 2-4, V, 2-3; vol. 3, 39; vol. 4, II-10, III-1 to 3, Plate 8 (reflecting other projects as part of background premises here). Further analysis in the present EA or FONSI would be redundant and in no material way serve the purposes of NEPA. The EA and FONSI were sufficient to alert interested members of the public to any arguable cumulative impacts involving these other projects.

COST quotes the following language of Kleppe v. Sierra Club, 427 U.S. 390, 410, 96 S. Ct. 2718, 2730, 49 L. Ed. 2d 576 (1976): "[W]hen several proposals ... that will have cumulative or synergistic environmental impact upon a region are pending concurrently before an agency, their environmental consequences must be considered together."10 Insofar as the Kleppe language refers to the issue of allegedly improper segmentation of related projects, we have addressed the issue above. Insofar as it may bear on an agency's duty to consider impacts in a context that realistically includes other pending projects, the appellees fully complied by planning on the basis of what might be regarded as the "worst-case scenario," i.e., it assumed ultimate completion of the related projects. This reading of the quoted Kleppe statement is supported by the Court's observation that "an agency could approve one pending project that is fully covered by an impact statement, then take into consideration the environmental effects of that existing project when preparing the comprehensive statement on the cumulative impact of the remaining proposals." Id. at 415 n. 26, 96 S. Ct. at 2732 n. 26. The approach taken by the officials here fits that suggestion and avoids unnecessary duplication.

Under Sec. 128 of the Federal-Aid Highways Act, 23 U.S.C. Sec. 128 (1982), public hearings must be held prior to approval of certain highway projects.11 COST raises two different challenges in connection with this statute. Brief for Appellants at 36-37. First, it argues that the public hearing held on February 15, 1984 was inadequate because of NEPA violations. Second, it claims that a shift in alignment made in response to issues raised at this hearing required another hearing. As we have explained, we find no NEPA violations, and thus we see no merit in the first of these arguments.

As for the second, we agree with the district court that project changes made after a Sec. 128 hearing require a new hearing only if those changes are substantial or significant. 642 F. Supp. at 606. See Louisiana Environmental Society, 537 F.2d at 89; Lathan v. Brinegar, 506 F.2d 677, 690 (9th Cir.1974); D.C. Federation of Civic Ass'ns, Inc. v. Volpe, 316 F. Supp. 754, 779 (D.D.C.1970), rev'd on other grounds, 459 F.2d 1231, 1243 (D.C.Cir.1971), cert. denied, 405 U.S. 1030, 92 S. Ct. 1290, 31 L. Ed. 2d 489 (1972). As the district court noted, the shift moved the highway nearer to a nursing home, some churches and an industrial area, but away from some residential areas. It concluded from the record that the noise levels would be affected only "slightly" in the obvious directions, 642 F. Supp. at 606, and COST does not argue to the contrary, much less offer proof. As noted in connection with Sec. 4(f), the move spared Tilden Park, but took an extra half-acre from Cabin John. The Cabin John shift was in an area 300 feet from the nearest intensive use (a trail). The district court's insubstantiality finding is clearly correct.

COST's final argument12 is that the district court improperly denied it opportunity for discovery. Although COST was vague as to what its discovery would entail, it indicated that it would seek evidence from the defendants tending to show that they failed to consider significant impacts. Tr., April 18, 1986, at 8-9. On appeal, it argues that Fed.R.Civ.P. 26(b)(1), which governs the usual scope of discovery, requires that it be permitted to investigate such matters.

As the district court observed, ordinarily "the focal point for judicial review should be the administrative record already in existence, not some new record made initially in the reviewing court." Camp v. Pitts, 411 U.S. 138, 142, 93 S. Ct. 1241, 1244, 36 L. Ed. 2d 106 (1973), accord Citizens to Preserve Overton Park, 401 U.S. at 420, 91 S.Ct. at 825; Environmental Defense Fund, Inc. v. Costle, 657 F.2d 275, 284 (D.C.Cir.1981). Although plaintiffs here were entitled to introduce evidence tending to show significant impacts or realistic alternatives that the responsible officials ignored, see Izaak Walton League of America v. Marsh, 655 F.2d 346, 369 n. 56 (D.C.Cir.), cert. denied, 454 U.S. 1092, 102 S. Ct. 657, 70 L. Ed. 2d 630 (1981), they were barred from inquiring into "the mental processes of administrative decisionmakers," absent "a strong showing of bad faith or improper behavior." Overton Park, 401 U.S. at 420, 91 S.Ct. at 825. See United States v. Morgan, 313 U.S. 409, 421-22, 61 S. Ct. 999, 1004, 85 L. Ed. 1429 (1941). In view of the vagueness of COST's discovery request, it seems reasonable to construe it as an effort to inquire into administrative mental processes. There was no serious, nonconclusory allegation of bad faith that might justify an exception to the general rule against such inquiries.

COST relies on Committee for Nuclear Responsibility, Inc. v. Seaborg, 463 F.2d 783 (D.C.Cir.1971), to support its proposal for unlimited discovery, but we do not read the case so broadly. In Seaborg, plaintiffs seeking to enjoin an underground nuclear test appealed an adverse summary judgment. Plaintiffs had alleged that responsible scientific opinion indicated adverse environmental consequences and that reports by federal agencies of such consequences had been suppressed. The court indicated that discovery was proper with regard to such matters and held that summary judgment prematurely foreclosed the plaintiffs from an opportunity to support their allegations. It did not indicate that free-wheeling discovery was appropriate in every NEPA challenge. We think the allegation of bad faith and apparent inability of the plaintiffs otherwise to secure relevant scientific data sufficiently distinguish Seaborg from this case.13

Evaluating environmental effects of major roadbuilding projects and adequately explaining the various choices made is a long and arduous process, as is reviewing such determinations. It is unlikely that any such undertaking will ever be absolutely flawless; perfection is elusive. We do not think the process was flawless here. However, after carefully examining each objection to the various administrative findings and explanations, we are persuaded that the appellees satisfied all statutory requirements. Accordingly, the judgment of the district court is

Affirmed.