The Eagle Foundation, Inc. v. Elizabeth Dole, Secretary of Transportation

EASTERBROOK, Circuit Judge. For more than 20 years state and federal officials have been planning a four-lane, limited access highway running east-west through central Illinois. The plans call for a highway that will connect Decatur, Springfield, and Jacksonville, Illinois, with Hannibal, Missouri. Parts of this highway have been built. One that has not is the segment spanning the Illinois River, which runs roughly north-south. The holdup is caused by the topography near the place the highway must cross the river. The western bank of the Illinois River is a steep granite bluff between 100 and 150 feet high, covered by loess soils that support trees and other vegetation. Erosion has produced ravines at places along the bluff, and these natural breaks in the rock are the best places through which to build highways. A broad, shallow valley is best of all. Napoleon Hollow is such a place. But ravines also are attractive to wildlife, wildlife is attractive to people, and the 862-acre Pike County Conservation Area (PCCA), which includes Napoleon Hollow, was established to preserve wildlife, some to be watched and some to be hunted. The PCCA’s status as a wildlife refuge brings into play Section 4(f) of the Department of Transportation Act, 49 U.S.C. § 303(c), which provides that the Secretary of Transportation may approve a highway “requiring the use of publicly owned land of a public park, recreation area, or wildlife and waterfowl refuge of national, State, or local significance, or land of an historic site of national, State, or local significance ... 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 protected area. 1 Before these restrictions became effective in 1968, state and federal officials had tentatively selected Napoleon Hollow as the place to cross the Illinois River. The new statute — and especially Citizens to Preserve Overton Park, Inc. v. Volpe, 401 U.S. 402, 91 S.Ct. 814, 28 L.Ed.2d 136 (1971), which held that § 4(f) forbids the use of park land except in rare circumstances — led the planners to conduct extensive studies of the effect of a highway through Napoleon Hollow and its alternatives. Between 1969 and 1972 state officials prepared and federal officials approved a lengthy “design location study” that addressed technological and environmental implications of building a 52-mile stretch of the new highway (between Jacksonville and Barry, Illinois) in different locations. This study included the inquiries and findings required by § 4(f) and by the National Environmental Policy Act (NEPA), 42 U.S.C. §§ 4321-47. The Secretary approved the study in 1974 after receiving extensive comments. No one sued, and some of the construction began. On July 21, 1978, the Keeper of the National Register of Historic Places complicated the picture. He concluded that the Wade Farm, 190 acres immediately to the west of the PCCA, was eligible for inclusion in the Register. The Wade Farm contains a limestone house built in the late 1840s or early 1850s. The Keeper concluded that the Wade Farm is significant as one *801 of the first settlements in Pike County, Illinois, and because it contains a fine early 19th century stone farmhouse. Section 4(f) applies to the Wade Farm in the same way it applies to the PCCA, and the owners of the Wade Farm (Sam Wade and his sister Juliet Wade) did not approve of the highway. Joined by the Eagle Foundation, Inc., a conservation group with a leasehold interest in part of the Wade Farm, the Wades filed suit in 1980, seeking to block the construction of the highway on the ground that the Environmental Impact Statement in the location study was inadequate and that § 4(f) prohibited the construction as a substantive matter. After some skirmishing about intervention, Wade v. Goldschmidt, 673 F.2d 182 (7th Cir.1982), the district court rejected the challenge to the Environmental Impact Statement. Wade v. Lewis, 561 F.Supp. 913, 945-48 (N.D.Ill.1983). The court concluded, however, that the Secretary had not adequately demonstrated that a route bypassing the PCCA and the Wade Farm is not “prudent”. Id. at 948-53. The court thought that alternatives to the use of Napoleon Hollow had not been explored with sufficient care. The court also held that independent of the § 4(f) obstacle no federal funds were available to build the bridge. Id. at 934-45. It enjoined further work. While the case was on appeal, Congress amended the funding legislation, specifically approving the building of this particular bridge across the Illinois River. Section 9 of Pub.L. 98-229, 98 Stat. 55. Because the defendants had completed a new study of alternatives, we concluded that the dispute about funding was moot and dismissed the appeal as moot to the extent it dealt with § 4(f). Wade v. Baise, 767 F.2d 925 (7th Cir.1985) (unpublished order). Sam Wade had died in the interim, but Juliet Wade and the Eagle Foundation promptly filed a new suit challenging the fresh decision (after the new study) to build through Napoleon Hollow and the Wade Farm. Juliet Wade died before the district court could make a new decision, and her successors did not take over the litigation. That left the Eagle Foundation as the sole plaintiff. The district court concluded that the Foundation has standing to challenge the routing through both the Wade Farm and the PCCA. Wade v. Dole, 631 F.Supp. 1100, 1105-07 (N.D.Ill.1986). The court then held that the Secretary did not act arbitrarily or capriciously in concluding that the new study establishes that no other placement of the bridge across the Illinois River is “feasible and prudent”. Id. at 1107-17. It also held that the current plans satisfy the requirement of § 4(f)(2) that the Secretary “minimize the harm” to the protected property. 631 F.Supp. at 1117-21. The new plan will consume 12.5 acres of the Wade Farm and 31 acres of the PCCA; the district judge concluded that the Secretary rightly minimized the net effects on the two properties considered as a unit and did not have to spare the Wade Farm at the expense of doing more damage to the PCCA. Id. at 1117-20. The defendants concede that a highway through Napoleon Hollow and the Wade Farm will cause several of the kinds of injury with which Congress was concerned when it enacted § 4(f). The site of the highway is rural, and the construction will not mar the only green area within miles, as the building of a superhighway through a park in downtown Memphis would have done. See Overton Park, 401 U.S. at 406-07, 91 S.Ct. at 818-19. Nonetheless, the Farm and PCCA contain not only historic buildings but also sites of archaeological interest, several important enough to justify their own listing in the National Register of Historic Places. See Wade v. Lewis, 561 F.Supp. at 922-24. Both the Farm and Napoleon Hollow contain plants and animals that have restricted habitats and precarious grips on existence. (The Foundation nonetheless does not raise any claim under the Endangered Species Act.) The bald eagle, an endangered species, maintains winter roosts along the Illinois River. It does not breed in Napoleon Hollow, but a few eagles forage during winter from trees in the Hollow and other ravines. A substantial percentage of all bald eagles outside Alaska winters in central Illinois,’ and Napoleon Hollow (together with ravines that feed into it) contains 17 of the 63 *802 potential roosting trees in the PCCA. See also 561 F.Supp. at 930-34 (discussing in detail the potential effects of the highway on eagles). The extent of the effect of the highway on eagles is unclear, however. Only one eagle has been seen in Napoleon Hollow since 1980, and eagles in other places have tolerated highways. (Pesticides rather than noise have been the principal threat to eagles.) Some less-known animals also inhabit the Hollow. A few Indiana bats live there. 2 Illinois chorus frogs (see Figure 1) entertain passers by. The chorus frog is named for its habit of calling in unison. Although we have been unable to learn whether the frogs sing in four-part harmony, they are not repulsed by automobiles. “[D]uring the breeding season, choruses can be heard continually along the highways in the area____ The call of the male is a series of short, loud, birdlike whistles.” Smith, Amphibians and Revtiles of Illinois, 28 Illinois Natural History Survey Bulletin 1, 85 (1961). See also Natural Land Institute, Endangered and Threatened Vertebrate Animals and Vascular Plants of Illinois 29 (1981). The Illinois chorus frog, a subspecies (Pseudacris streckeri illinoensis) of a frog that is abundant in central Texas and Oklahoma, has been found in only seven counties of Illinois and is “considered vulnerable because its restriction to sand areas could subject it to habitat degradation. Much of the original sand prairie is being encroached upon by cultivated fields. In addition, many of the remaining open sandy areas are now being planted to pines for the Christmas tree market.” Ibid. As for plants: there is a beautiful (and unusual) double hedgerow of Osage orange trees more than 100 years old in the Wade Farm, and Napoleon Hollow contains jeweled shooting stars, American ginseng, and golden seal, all unusual in Illinois. 3 [[Image here]] The Department of Transportation took all of this and more into account in reassessing the route of the highway. The district court summarized the two-volume study, 631 F.Supp. at 1108-17, which is backed up by two volumes of technical *803 appendices, sixteen additional technical reports, a new survey of the effect on eagles, and three volumes containing correspondence and the record of a public hearing. All of this culminated in a formal decision by the director of the Office of Planning & Program Development of Region 5 of the Federal Highway Administration, the Secretary’s delegate. Director E.V. Heath-cock concluded that so long as a highway is to be built (and the Foundation does not contest the wisdom of building this highway), all technically feasible alternatives to Napoleon Hollow are imprudent — some would be unsafe because of sharp turns, all would be longer and more expensive to build, and most other alternatives entail blasting through the bluff of the Illinois River, which would create an ugly “concrete canyon” for the highway and lead to the erosion of the loess soil remaining on the bluff nearby. Some of the other alignments would endanger roosting sites for eagles and expose other wildlife to injury. Director Heathcock believed that the cumulative drawbacks of other routes made them imprudent compared with the Napoleon Hollow route. The district court concluded that this assessment did not overstep the leeway allowed under Overton Park. See 631 F.Supp. at 1107-17. Overton Park calls for “a thorough, probing, in-depth review” (401 U.S. at 415, 91 S.Ct. at 823), of a decision to build a highway through land covered by § 4(f). The Court also said that “[although this inquiry into the facts is to be searching and careful, the ultimate standard of review is a narrow one. The court is not empowered to substitute its judgment for that of the agency.” Id. at 416, 91 S.Ct. at 824. This calls for a fine distinction, one that is hard to make in practice, between the standard of inquiry (“thorough”) and the standard of review (“narrow”). The “probing” inquiry ensures that the court learns what is going on and does not decide on the basis of superficial beliefs and assumptions; the deferential review ensures that once the court is satisfied that the Secretary took a close look at the things that matter and made the hard decisions, those decisions stick. The court’s role is to find out whether the Secretary considered what she had to consider, put out of mind what she was forbidden to consider, and dealt rationally with the competing relevant issues. As with cases under other statutes dealing with the environment, “we must see to it that the agency took a ‘hard look’ at the environmental factors implicated and based its decision on a rational consideration of relevant factors”, Van Abbema v. Fornell, 807 F.2d 633, 636 (7th Cir.1986). Because § 4(f) has a substantive as well as a procedural component, we must also ensure that the Secretary acted with her thumb on the scale, conscious of the importance of protecting the lands listed in the statute. Searching but deferential review, an in-depth probe to find a hard look but then to accept the results, is a delicate task. Such a thorough inquiry runs the risk that the judge, having learned enough to make up his own mind about the wiser course, will deem any other decision to be arbitrary. The judge must combine dogged pursuit of the evidence with humility about his own role, always keeping in mind the possibility that competent people of good will may reach different conclusions on the basis of the same record. The district judge in this case has reviewed two records dealing with the decision to build the highway through Napoleon Hollow. Each time he approached the record and the Secretary’s decision from the appropriate perspective, determined to learn whether the Department has dealt with important matters yet willing to accept decisions honestly made. Judge Will’s thorough presentations of the evidence and the conflicting inferences, 561 F.Supp. at 919-34, 945-54, and 631 F.Supp. at 1108-21, are models of the judicial role well conceived and carried out. We doubt that we could improve on them, and we would not try to if we could; twicé is enough. We shall limit our discussion to the Eagle Foundation’s contentions that Judge Will made mistakes of law on the way to his conclusions. 1. The district court looked for evidence that the Secretary’s delegate reached a prudent accommodation of competing interests. The court satisfied itself that the *804 Department had started the inquiry with a strong presumption against using land covered by § 4(f) and had not taken into account the sorts of things that according to Overton Park, 401 U.S. at 411-13, 91 S.Ct. at 821-22, § 4(f) excludes from consideration, such as the convenience of going through land from which no one need be evicted. (It is a lot easier to oust an eagle than an elderly grandmother determined to live out her life on the farm where she was born.) Having determined that the Secretary included what was pertinent and excluded what she must, the district court asked whether a responsible, well-informed public official could think the decision to build through Napoleon Hollow “prudent”. This is the standard widely employed in § 4(f) cases. Arizona Past and Future Foundation, Inc. v. Lewis, 722 F.2d 1423, 1425 (9th Cir.1983); Louisiana Environmental Society, Inc. v. Dole, 707 F.2d 116, 122 (5th Cir.1983); Coalition for Responsible Regional Development v. Coleman, 555 F.2d 398, 399-400 (4th Cir.1977). It is the right standard. The Secretary must decide what is “prudent”. Such an inquiry calls for judgment, for balancing, for the practical settlement of disputes on which reasonable people will disagree. The statutory standard makes deferential review inevitable. The Eagle Foundation replies that the approach disregards sterner language in Overton Park. The Court stated that § 4(f) “is a plain and explicit bar to the use of federal funds for construction of highways through parks — only the most unusual situations are exempted.” 401 U.S. at 411, 91 S.Ct. at 821. “If the statutes are to have any meaning, the Secretary cannot approve the destruction of parkland unless he finds that alternative routes present unique problems.” Id. at 413, 91 S.Ct. at 822. “[T]he reviewing court must be able to find that the Secretary could have reasonably believed that ... there are no feasible alternatives or that alternatives do involve unique problems.” Id. at 416, 91 S.Ct. at 823. Yet the Court also concluded that the Secretary’s decision may be set aside only if “arbitrary or capricious”, id. at 414-16, 91 S.Ct. at 822-24, one of the most deferential standards of administrative law. The Foundation confuses what the Secretary must try to do, which is to decide whether there is a powerful reason to build through the § 4(f) property, with what the court must try to do, which is to determine whether the Secretary made a reasoned judgment after a hard look at the facts. In saying that the Secretary must have a strong reason, we do not overlook the Supreme Court’s use of “unique”. If the reasons not to use the § 4(f) property must be “unique”, then the Secretary’s decision here cannot stand. “Unique” is a word without degrees; a situation is unique (nonpareil, one of a kind) or it is not. There is nothing “unique” about the reasons given for believing that Napoleon Hollow is the best location for the bridge. The other routes are longer and more expensive, will take more time to complete, disrupt farm lands, produce erosion and ugly concrete canyons, kill the local wildlife, and may contain steep grades and sharp bends that endanger travelers (especially in icy conditions). None of these is “unique”. America’s highways present examples of delay, cost overruns, disfigured landscape, deceased wildlife, severed farms, and dangerous grades in abundance. Yet we cannot believe that the Supreme Court meant that if a risk or cost has been accepted, or an obstacle overcome, for any highway in the United States, then it always must be accepted or overcome in preference to the use of any § 4(f) lands, however trifling the effects of using the § 4(f) lands. For some statutes cost is no object. E.g., Union Electric Co. v. EPA, 427 U.S. 246, 96 S.Ct. 2518, 49 L.Ed.2d 474 (1976). Nothing in the language or legislative history of § 4(f) suggests such an extreme position, and the statutory use of “prudent” cuts the other way. Cf. Natural Resources Defense Counsel, Inc. v. EPA, 804 F.2d 710 (D.C.Cir.1986) (a court should read a statute as excluding consideration of costs and related difficulties only when the language and structure are unmistakable). Overton Park was being emphatic, not substituting “unique” for “prudent” in the text of *805 § 4(f). The Court made it pellucid that the reasons for using the protected land have to be good ones, pressing ones, well thought out; the Secretary must start with a strong presumption against turning chlorophyll cloverleafs in the parks into concrete ones; but even after starting from this presumption it may be prudent to use the § 4(f) land. If the Secretary makes that hard decision, it must be respected. 2. The two-volume study and the decision of the Secretary’s delegate examine the drawbacks of routing the highway around Napoleon Hollow. They conclude that all of these drawbacks together make use of the Hollow prudent, even though no one of them would have. The Eagle Foundation’s brief appeared to contest this cumulation of effects, which the district court approved. The brief maintained that each drawback must be examined separately, and unless one whale of a problem is enough by itself to justify use of the § 4(f) property, the highway must be rerouted. At oral argument the Foundation withdrew this contention, portraying it as a version of its belief that only “unique” difficulties (that is, a lollapalooza of a problem) justify use of § 4(f) lands. The Foundation’s modification of its position was, in the language of § 4(f), prudent. A prudent judgment by an agency is one that takes into account everything important that matters. A cumulation of small problems may add up to a sufficient reason to use § 4(f) lands. It would be imprudent to build around the park if the Secretary were convinced that the aggregate injuries caused by doing so exceeded those caused by reducing the size of the park. Even a featherweight drawback may play some role. No feather weighs very much, but a ton of feathers still weighs as much as a 2,000 pound block of lead. See Town of Fenton v. Dole, 636 F.Supp. 557, 567 (N.D.N.Y.), aff'd, 792 F.2d 44 (2d Cir.1986) (holding that the Secretary may take cumulative effects into account); Citizens to Preserve Wilderness Park v. Adams, 543 F.Supp. 21, 35 (D.Neb.1981) (same), aff'd mem., 685 F.2d 438 (8th Cir.1982). 3. The Department examined more than ten routes as alternatives to Napoleon Hollow. First the Department selected two end points for the study, one on the east of the river representing the existing terminus of the four-lane highway, and one on the west of the Wade Farm representing a plausible place from which to continue the highway. The two end points are ten miles apart; the Wade Farm and PCCA together run about two miles east to west. Next engineers selected places in which the Illinois River could be bridged within about three miles, north or south, of Napoleon Hollow. The nature of the bluff placed constraints on this choice. They chose seven potential sites. Then they examined possible routes to and from these bridges that would depart from the east end point and arrive at the west end point. Many routes to and from each bridge site were possible. The ones to which they gave the greatest attention are depicted on a map (Figure 2). The Department analyzed each of these alternatives and gave reasons, which the district court thought sufficient, for believing it prudent to prefer Napoleon Hollow over each of the alternatives. *806 [[Image here]] *807 The Eagle Foundation’s criticism of this methodology can be summed up in a word: more. The Foundation wants the Department to use more control points. Why not, it asks, put a new control point one-half mile to the north at the west end and see whether this affects things? Why not look at additional .routes that do not turn south or north, after passing the Wade Farm, as quickly as all except Alignments 1C and 7 do? The Foundation does not seriously claim that the study chooses terminal points too close to the protected property, thereby ensuring that the only feasible route runs through § 4(f) land. Ours is not a case like San Antonio Conservation Society v. Texas Highway Department, 446 F.2d 1013 (5th Cir.1971), cert. denied, 406 U.S. 933, 92 S.Ct. 1775, 32 L.Ed.2d 136 (1972), in which the Department built roads up to either side of a park and then proposed to let the state build the only feasible connecting route. The ten by six mile area of this study is sufficient to contain many realistic alternatives to the use of the § 4(f) lands. The Secretary’s obligation is to look at enough alternatives to make possible an informed judgment about whether one is likely to be feasible and prudent. See Monroe County Conservation Council, Inc. v. Adams, 566 F.2d 419, 425 (2d Cir.1977), ce rt. denied, 435 U.S. 1006, 98 S.Ct. 1876, 56 L.Ed.2d 388 (1978); Life of the Land v. Brinegar, 485 F.2d 460, 470-72 (9th Cir.1973), cert. denied, 416 U.S. 961, 94 S.Ct. 1979, 40 L.Ed.2d 312 (1974). The choice of terminal points here left ample room for the generation of alternatives. The 60 square miles within the area of study could contain millions of different alignments for this highway. Ideally we would like to know which of these is the best. Yet analyzing each one would take forever and thereby deny the premise of the project — that it is better to have a highway, with the attendant costs, than to have none. The selection of one route from among many alternatives is a familiar problem in linear programming, a branch of mathematics. Suppose a salesman with 75 customers wants to know the shortest route that will enable him to call on each, or an airline that serves 75 cities wants to plan its schedules to take maximum advantage of each plane. Starting from any point, there are 74 places to go; from each of these second points one must choose among 73 remaining alternatives, and so on. The total number of routes for a plane or salesman is 74! (which means 1x2x3 ... X 74). Seventy-four factorial is 3.31 X 10107, which is larger than the number of elementary particles in the universe. Finding ways to winnow the task to one that can be accomplished (approximately) within a finite amount of time has proved to be a challenging task. All of the existing computer programs stop looking when further computation is unlikely to produce a significant improvement. 4 The task of picking a route for a highway is very similar. At any point the highway may go north, south, or straight ahead; this choice must be made yard-by-yard, mile-by-mile, until the number of alternatives is staggering. The answer to the question “Why not look at a few more?” is that there is no limit to the number of additional options, and the addition of even a few vastly complicates the analysis. The proper inquiry is not whether more options remain to be exam ined — for that will be true always — but whether enough have been examined to permit a sound judgment that the study of additional variations is not worthwhile. The Department studied alternatives based on bridges up and down the Illinois River for a stretch of six miles. All of the routes derived from these bridges ran into difficulties that were spelled out in detail. Two of the routes ran well to the west of the Wade Farm before rejoining the proposed path of the longer highway. The Eagle Foundation does not offer a reason to believe that any other route within the area of the study would do better. Perhaps, if the Eagle Foundation had designed a different route and demonstrated the advantages of this path, the Department *808 would have been required to analyze it. Cf. River Road Alliance, Inc. v. Corps of Engineers, 764 F.2d 445, 452-53 (7th Cir.1985), ce rt. denied, — U.S. -, 106 S.Ct. 1283, 89 L.Ed.2d 590 (1986); Van Abbema, 807 F.2d at 643. But the Foundation wanted the Department to shoulder the entire burden. The Secretary has discharged the only burden properly placed on her. 4. The Department took into account the costs of avoiding Napoleon Hollow. Each alternative would be more expensive than the use of the Hollow. Some of the excess is attributable to the additional length of the other alignments, some to the need to blast through the bluff, some to the need to duplicate (for the new routes) engineering studies and plans that have been completed for the route through the Hollow. The route through Napoleon Hollow is expected to cost about $80 million, or $8 million less than the next-least-expensive alternative. (Some routes are substantially more expensive still.) Of the $8 million difference, about $2.7 million is attributable to engineering work done before the injunction in 1983 stopped further preparations; these sunk costs need not be incurred a second time to go through the Hollow, but would be borne anew on any other route. The Foundation raises two objections to the Secretary’s use of costs: that $8 million is a small fraction of the total cost of the highway, and that the $2.7 million in sunk costs should be disregarded. To decide whether an alternative to Napoleon Hollow was prudent, the Secretary needed to know the incremental cost of the alternatives. The answer that came back is $8 million and up. The Secretary then could ask intelligently whether it is worth $8 million to build around the Hollow, in light of the other benefits and drawbacks of each course of action. The Foundation wants to express these costs in percentages. Eight million dollars is 10% of the cost of the 10-mile segment under study; $8 million is a smaller percentage of the whole highway’s cost from Decatur, Illinois, to Hannibal, Missouri; $8 million is a smaller percentage still of the cost of the whole highway system of which this will be a part. Eight million dollars is also a small fraction of the cost of an aircraft carrier, of the federal budget, and of the gross national product. Any percentage can be shrunk by increasing the denominator while holding the numerator constant. Increasing the denominator does not help to answer the pertinent question, however. The absolute comparison — Is it worth $8 million to avoid Napoleon Hollow? — is what matters. In any analysis of costs and benefits, the incremental costs and benefits of the decision under study are the sole concern. Other costs that will be incurred no matter what, such as the expense of building the rest of the highway, are distractions. They may also give the wrong impression. To express the cost of building around the Hollow as a percentage of the cost of the whole road is to assume that there will not be any problems anywhere else. For all we know, a highway encounters some § 4(f) property every 10 miles. In that event, the study at issue is being repeated for every other segment of this highway. Looking segment-by-segment will reveal the full costs and benefits of each decision. It is more useful to express these costs as absolute dollar figures than as percentages of other sums. 5 The Secretary’s consideration of sunk costs is more troubling. Suppose 90% of the cost of building a highway lay in the engineering studies, and that the Department incurred these costs after learning of a potential § 4(f) problem. Then the Secretary might say: “It costs $100 million to build the road through the park, and we have incurred $90 million of that; it would cost only $50 million to build the road elsewhere, but we haven’t spent any of that; so only the road through the park is now a *809 prudent choice.” It is possible to incur costs as a strategic matter in order to influence the outcome. Once the costs have been sunk, the incremental cost of completing the job is smaller than the full costs of an alternative route. It would be very troubling if this sort of thing had happened here, because it could stack the deck against preservation of § 4(f) property. Yet consideration of incremental costs is the prudent method over the long run. If you want to build a house, and the lot contains a foundation, you will compare the costs of completing the house on the existing foundation with the costs of building an entire house, including the foundation, on another part of the land; only a fool would disregard the fact that some work had been done. Perhaps the right accommodation is to allow the Secretary to take account of sunk costs unless the expenses were incurred as part of an effort to undermine a fair comparison of costs. We need not offer a complete analysis, however, because as the district court held there is no reason to think that the $2.7 million in sunk costs made the difference to the decision. 631 F.Supp. at 1111-12. Even disregarding these costs the Napoleon Hollow alignment is least expensive; anyway, costs were not the dominant consideration in the Department's choice. Too, much of the $2.7 million was incurred after the decision of 1974, which followed a full study, and before the filing of suit in 1980. A court cannot fault the Department for proceeding apace on a project that had been studied and was apparently lawful. 5. Section 4(f) is not the only statute telling highway planners to protect certain lands. The Farmland Protection Policy Act of 1981, 7 U.S.C. §§ 4201-09, instructs all agencies to “minimize the extent to which Federal programs contribute to the unnecessary and irreversible conversion of farmland to nonagricultural uses”. 7 U.S.C. § 4201(b). The Secretary considered two effects on farms of moving the highway out of Napoleon Hollow. One is that farms will be split into two, causing farmers to travel substantial extra distances to bridges over the highway in the course of working the fields; the other is that a diagonal cut of a highway across a farm presents complications for the farm’s mechanical operation. The diagonal cut skews the rows, and at triangular intersections machines operate less effectively or break down more often. See 631 F.Supp. at 1113 & n. 5, reproducing part of the discussion in the Department’s analysis. Some of these effects may be alleviated by selling land so that each farm lies on one side of the road (the farmer whose “south 40” now lies on the other side of the highway may swap land with the farmer whose “north 40” got cut off). But the creation of triangular rather than square comers cannot be dealt with by changing the orientation of the furrows. The Foundation objects to any consideration of agricultural effects on the ground that § 4(f), “more specific” because it applies only to highways, overrides § 4201(b), which applies to agricultural lands in general. This brings into play a battle of maxims, and the Secretary predictably responds that § 4201(b) is the more specific because it applies only to farmland, while § 4(f) more generally deals with highways in every part of the country; anyway, the Secretary observes, § 4201(b) is the more recent. See also the battle of canons among the Justices in Radzanower v. Touche Ross & Co., 426 U.S. 148, 153-54, 96 S.Ct. 1989, 1992-93, 48 L.Ed.2d 540 (1976), and id. at 158-60, 96 S.Ct. at 1995-96 (Stevens, J., dissenting). Still another maxim is that statutes should be harmonized when possible, so that each can play some role. Sections 4(f) and 4201(b) are not antagonists. Section 4(f) says to steer clear of the parks when prudent to do so; § 4201(b) identifies something for the prudent administrator to consider. Neither must yield to the other. The Secretary was entitled to consider effects on farmland when deciding where to build the highway. 6. What remains is the minimization of harms. The statute decrees that minimal harm be done to any § 4(f) land that must be employed. The Secretary has pledged to build in the Hollow only during the summer (when eagles are elsewhere), to *810 split the highway so that two lanes run along each wall of the canyon (minimizing visual clutter and disruption of animals’ travel, and facilitating natural drainage), to build restraining walls that minimize erosion, to build bridges (at a cost of $6 million) across the side ravines that feed into Napoleon Hollow (to allow wildlife to cross the road), to put the highway eight feet underground near the Wade Farm’s buildings (in order to reduce noise and visual distraction), and to plant several rows of 10-15 year old pine trees between the Wade buildings and the highway to serve as a screen. (All of this is included in the $80 million cost.) The district court remarked: “The minimization plan goes well beyond ... anything we have seen in prior reported cases.” 631 F.Supp. at 1120. The Eagle Foundation wants the Secretary to redo the whole plan in order to minimize the harm to the Wade Farm taken as an entity — even if that means greater harm to the PCCA. Anything less, the Foundation insists, denigrates the status of the Farm as a separate § 4(f) site. This is silly, as the district court held. 631 F.Supp. at 1117-18. The statute tells the Secretary to minimize harm to § 4(f) lands. The national interest lies in all § 4(f) lands as an aggregate, and the Secretary is serving the national interest rather than the private interests of the Foundation or the Wades. Everyone would like someone else to bear the costs. There is no benefit to the nation in doing more total damage to protected lands to avoid taking 12.5 acres of the Wade Farm. What is more, the Secretary concluded that alignment 5A is not “feasible” because it contains a steep grade leading to a sharp turn. In icy conditions, the combination is lethal. Still, the Foundation insists, the Secretary is estopped to use any of the Farm because in the course of the first suit a lawyer for the Secretary represented that the Department would avoid the Farm. The Foundation invokes the doctrine of judicial estoppel, which precludes parties from abandoning positions taken in earlier litigation. Himel v. Continental Illinois National Bank, 596 F.2d 205, 210 (7th Cir.1979). The principle is that if you prevail in Suit # 1 by representing that A is true, you are stuck with A in all later litigation growing out of the same events. That does not assist the Eagle Foundation. The Secretary did not prevail in Suit # 1; the district judge enjoined the construction notwithstanding the representation. The Department then restudied everything from scratch. One of the discoveries is that the route that goes through Napoleon Hollow but avoids the Wade Farm will kill people. That is an adequate reason to change position. See also 631 F.Supp. at 1118-19. A group concerned about haliaeetus leucocephalus should show similar concern for homo sapiens. Affirmed Footnotes [1] A similar requirement appears in 23 U.S.C. § 138, and the parties agree that this does not add anything to § 4(f). We therefore discuss § 4(f) alone. [2] Indiana bats rarely live in Illinois (or Indiana, for that matter). They prefer caves to ravines. See S.R. Humphrey, A.R. Richter & J.B. Cope, Summer Habitat and Ecology of the Endangered Indiana Bat, Myotis Sodalis, 58 J. Mammalogy 334 (1977); U.S. Fish & Wildlife Service, Recovery Plan for the Indiana Bat (1983). But there are some nesting in trees in central Illinois, and one lactating female was found in Napoleon Hollow. [3] For example, golden seal, which may be found in the eastern United States and adjacent Canada, is “threatened” in Illinois. "A species of mesic and wet-mesic upland forests, Hydrastis canadensis [golden seal] ranges across Illinois and was originally a common species in most of the forested areas of the State. It is becoming rare, however, since it has been extensively dug for its roots, which are reported to have medical properties. Together with habitat destruction, this exploitation has greatly depleted the species’ populations in Illinois.” Natural Land Institute at 173. [4] See Symposium, Practical Developments in Mathematical Programming, 36 J. Operations Research Society 345-550 (1985); Simplified Simplex, 252 Scientific American 54 (Jan. 1985). [5] It might be better still to express things in terms of rates of return. Given that the United States has $8 million to spend, can it do better by spending that money to save Napoleon Hollow or by providing better shoulders and shallower turns (to reduce the number of accidents) or by doing more basic research in cancer or nuclear fusion? A rate of return computation permits a more general comparison. But the Eagle Foundation did not call for such a computation, which is itself a complex problem.