JOHN R. BROWN, Chief Judge:

The Just Compensation Clause of the Fifth Amendment, which provides that private property shall not be taken without just compensation, is one of the most basic constitutional guarantees that the rights of the individual will not be unfairly sacrificed in the pursuit of the common weal. In this eminent domain case, appellant landowners claim that this constitutional guarantee has been breached, that because of numerous substantive and procedural errors in the proceedings below they have not been adequately and justly compensated for the loss of their properties condemned by the Federal Government.

The public project in this case is Everglades National Park, which covers approximately 2000 square miles of the southernmost portion of the Florida peninsula. The compensation trials below involved 63 tracts (totalling approximately 788 acres) taken for inclusion within the Park. The owners of 52 of the tracts have appealed. Among the issues they present is the proper application of the "scope of the project" rule and the principle of "highest and best use." We vacate because of procedural deficiencies, but we also enunciate the legal standards of just compensation that are to guide retrial of the 52 tracts appealed, as well as literally hundreds of other properties condemned for Everglades National Park and the neighboring Big Cypress National Preserve.

Because the history of Everglades National Park is so important to a proper understanding and just resolution of this case, our opinion begins with a brief review of that history. Then, in Part I, we recount the proceedings in the District Court, including the procedural error which invalidated those proceedings almost before they began. After an overview of the basic principles of just compensation law in Part II, we turn to the important substantive issues raised by this appeal. With respect to the scope of the project rule, we first discuss the rule generally (Part IIIA), including its underlying precepts and how it has been applied in other major federal court decisions. We then review several errors in its application by the District Court to the facts of these Everglades condemnation proceedings (Part IIIB), followed by a discussion of how the rule should be applied upon remand (Part IIIC). We conclude our consideration of the rule with a discussion of the differing responsibilities it imposes upon the judge and the fact-finder in just compensation trials (Part IIID). In Part IV, we review an evidentiary ruling concerning the admissibility of certain "comparable sales" in the trials below. Important issues concerning "highest and best use" and compensation for "illegal uses" are discussed in Parts V and VI respectively. Part VII is devoted to several miscellaneous evidentiary issues that arose during the District Court trials, including the admissibility of just compensation statements sent landowners prior to the institution of eminent domain proceedings. Finally, in Part VIII, we remand the issue of just compensation in these cases for trial by a commission in lieu of trial by a jury.

Everglades National Park And The Northwest Extension

The origins of this case trace back to 1929 when the Federal Government first began to investigate the prospects of establishing a National Park in the Everglades. In 1930, the Secretary of the Interior recommended the establishment of a national park encompassing about 2000 square miles (1,280,000 acres) of the southern tip of the Florida peninsula, and in 1934 Congress took the first tentative steps toward establishment of the Park. However, largely because Congress refused to authorize purchase of parklands with public money, Everglades National Park did not become a reality until 1947, when it was established and dedicated with a nucleus of 454,000 acres granted by the State of Florida. In addition to the land donations, the State of Florida made available to the Federal Government $2,000,000 with which the Department of the Interior began to acquire privately owned lands within the designated Park boundaries.

After certain changes in the park boundaries declared by the Secretary of the Interior, Congress, in the Act of July 2, 1958, officially redefined the park boundaries, since when the dimensions of the Park have not been changed. The 1958 Act included within the Park for the first time an area known as the Northwest Extension, in which the properties involved in this case are located. Overall, the 1958 Act increased the size of the Park to 1,390,000 acres, of which all but 197,000 acres were at that time owned by the United States. Of these 197,000 acres, some were contemplated to remain in private ownership, some were to be donated to the Federal Government, and 81,000 acres — among them the 788 acres at issue in this case — were to be acquired "as appropriations become available for the purpose."

The 1958 Act, however, severely limited the amount that could be appropriated for purchasing the remaining privately held lands, by putting a ceiling upon future appropriations of $2,000,000 (approximately $24 per acre). Moreover, not even a fraction of the privately held lands could be purchased immediately as Congress did not appropriate any money for that purpose for some time. When anxious landowners inquired of the Department of the Interior concerning the status of their lands and the possibility of imminent condemnation, they were assured, in a series of letters between 1958 and 1962, that the Department would not begin its acquisition program until Congress appropriated the necessary funds, that the Department did not know when Congress would in fact appropriate the funds, but that in the meantime the landowners were completely free to sell, use, or improve their within-Park properties as they were able to and as they saw fit.

As things developed, the Department of the Interior did not begin to purchase privately held lands with federal funds until 1966, and when it did, the $2,000,000 authorized back in 1958 was quickly exhausted. As Congress subsequently realized, the 1958 authorization had been made without the benefit of any detailed appraisals and the errors made in the estimates were compounded by "the rapid escalation of land prices" so that the $2,000,000 was insufficient for the acquisition of at least 74,000 acres of the remaining privately owned land. It was not until 1970 that Congress fully remedied its initial underestimation of the money necessary to complete its acquisition program, when it withdrew the $2,000,000 cap and increased the authorized appropriation elevenfold to $22,000,000. Congress actually appropriated most of this sum shortly thereafter and, sometime in 1971 or 1972, the Department of the Interior began the process of acquiring the remaining privately held parklands, including those involved in this case.

I. THE PROCEEDINGS BELOW

The 52 tracts of land involved in this appeal were among thousands of Everglades parcels destined for acquisition once Congress loosened the pursestrings. The acquisition of many tracts was accomplished through settlement negotiations, but hundreds more went to trial in the Southern District of Florida. Processing these hundreds of eminent domain cases, most of which involved parcels of land an acre or two in size, involved peculiar problems of judicial administration and, in turn, required somewhat unusual solutions. For most of the landowners, the costs of litigating their cases individually were prohibitive. Consequently, the "trial" in many cases — those in which the owner had neither retained counsel nor requested a jury trial — consisted of a Government appraiser taking the stand and reading his estimate of fair market value for a number of tracts, followed by the entry of judgment in the amount of the appraised sum for each tract on the list. Other landowners, however, would pool their resources and retain one attorney to represent their cases jointly. The District Judges, then, had the unenviable task of presiding over a process that was inherently chaotic — sending notice to hundreds of landowners, trying in summary fashion those parcels not represented by an attorney, holding pre-trial conferences and hearings for those tracts represented by a handful of attorneys, and then finally conducting full-blown compensation trials involving upwards of thirty parcels each.

To achieve justice in a situation like this, where the condemned properties are distinguishable only by a docket number and a tract number, is no easy task. Few if any Trial Judges — no matter how conscientious — could be expected to maintain order without the cooperation of the one other party that was an ongoing participant throughout the entire proceedings — namely, the United States Government. Unfortunately, the Government attorneys, at least in this case, did not shoulder their responsibility for facilitating justice in these unusual circumstances, but instead assumed the distinctly adversarial role more appropriate to the conventional one-on-one litigation, the cumulative result of which was that the landowners before us on appeal were not afforded the due process guaranteed them by the Fifth Amendment. What follows is an abbreviated story of three trials, which, largely because of certain inexcusable procedural irregularities during the pre-trial proceedings, must go for naught.

* * * * * *

Notices of condemnation were sent to the owners of the various properties in this case between July 28, 1975 and January 14, 1976. All of the owners eventually retained Mr. William Earle, a lawyer who specializes in eminent domain litigation, to try their cases. The various individual cases were consolidated into three groups. In the first group were 4 large tracts, ranging in size from 40 to 320 acres. The second and third groups consisted of smaller properties — 29 properties averaging 1.91 acres in group 2, and 30 properties averaging 2.83 acres in group 3. Each of the groups was tried separately, before three different juries, in successive trials presided over by Senior District Judge Mehrtens during the week of May 10-14, 1976. By stipulation and with the approval of the Court, all motions, objections, proffers, and rulings applied to all three trials.

The trials themselves were bitterly waged contests, object lessons that the adversary process can be carried to intolerable extremes. No participant comported himself in a manner conductive to a fair determination of the truth — in this case, just compensation. Instead, the trials were plagued with the repeated badgering of witnesses, numerous picayune objections and motions, and generally disrespectful and derisive behavior. The obstreperous conduct of the attorneys undoubtedly stemmed in large part from two crucial pre-trial rulings which the landowners perceived to place them at a serious disadvantage and therefore repeatedly attempted to circumvent or have overturned, and which the Government perceived to give it a decisive advantage which it was anxious to preserve and exploit.

The first of these two rulings was that the landowners, in attempting to show the fair market value of their properties at the date of taking, could not introduce any evidence of sales of either theirs or comparable properties that had occurred within the boundaries of Everglades National Park after July 2, 1958. This ruling was designed to give effect to a principle of condemnation law that has come to be known as the "scope of the project" rule. The underlying precept is that just compensation does not include any value attributable either to a unique governmental demand (as opposed to private demand) for the property or to the fact that the Government is exercising its powers of eminent domain on behalf of the public. Thus, if condemned property was "probably within the scope" of a public project from the time the Government became committed to that project no enhancement in value attributable to the public project may be considered in awarding compensation.

The second of the crucial pre-trial rulings concerned the "highest and best use" of the properties. Another principle of eminent domain law is that a landowner is not limited to showing the value of his land as presently used, but may introduced evidence of potential uses to which his property may readily and reasonably be converted, since demand for potential use affects market value. In this case, Judge Mehrtens entered a pre-trial ruling precluding the landowners from introducing any evidence that their properties were suitable for cabin sites or other construction purposes.

In the rather unusual circumstances of these Everglades condemnation proceedings, these two rulings combined to place extremely restrictive limits, which Judge Mehrtens rigorously enforced, on the evidence the landowners could admit to show the fair market value of their properties. Other rulings made during trial further restricted the landowners' presentation of their cases. The restrictive impact of some of these rulings is discussed in the following paragraphs.

The Scope Of The Project Ruling

On April 21, 1976, a few weeks prior to the three trials, the Government moved the Court to prohibit the landowners from making reference to any prior sales of their properties or of other lands within the Everglades National Park which took place after July 2, 1958, "on the grounds such lands were within the scope of the project after that date." By his order of April 27, 1976, Judge Mehrtens granted the Government's motion, ruling that the landowners "shall not offer evidence as to any sales which occurred within the boundaries of the Everglades National Park after July 2, 1958." Despite repeated objections during the three trials, Judge Mehrtens stood by this ruling throughout.

A full appreciation of the restrictiveness of this ruling requires some knowledge of Florida geography and of the peculiar characteristics of the condemned properties. As already mentioned, the 63 tracts involved in the three condemnation trials are all located in an area of the Park known as the Northwest Extension. More specifically, the properties are located on the southwest coast of Florida along a chain of marked, deep-draft intercoastal waterways just a few miles from the Gulf of Mexico. The land in this area is cut with numerous creeks, streams, and bays, and many of the landowners' properties lie along navigable waterways with access to the Gulf of Mexico; one of the larger tracts is actually located on the Gulf.

Portions of some of the properties are submerged except at low tide; other portions have an elevation ranging from several inches to about one foot above mean high tide. Much of the soil in the area is muck, especially during the wet season. The water is brackish — a mixture of fresh water flowing from inland regions and salt water coming in with the tides from the Gulf of Mexico. Because of this mixture, it is an extremely fertile area for a wide range of marine life. It is also populated by deer, wild pigs, raccoons, bobcats, wild turkeys, squirrels, and the like. As one Government witness testified, "[i]t is biologically one of the most productive areas that you will find anywhere in the world."

The characteristic that is most distinctive of this area, however, is its predominant vegetation — the mangrove tree. The mangrove is a "jungly" hardwood tree of about 20 to 30 feet in height (occasionally as high as 80 feet), with a very dense, tangled above-ground root system. It will grow even where its roots are covered by the tide for a portion of the day. In places, the mangroves are so thick that transportation, either by boat or over ground, is virtually impossible.

At one time, before the development of certain areas like Miami Beach, virtually all the coastal lands of South Florida were mangrove lands. Thus, the mangrove lands roughly form a horseshoe around South Florida. To the interior are sawgrass plains and the cypress zone — areas that in many respects are just as much a wilderness with as varied an animal life as the Everglades mangrove lands, but in many other respects are quite dissimilar to coastal mangrove lands.

Indeed, the landowners took the position that mangrove lands are not truly comparable to any other lands in all of Florida. Thus, as evidence of fair market value of their lands at the date of taking (which in these cases was the date of trial), they desired to introduce appraisals based on recent sales of other South Florida mangrove lands (the only truly "comparable" sales in their opinion). The Court's "scope of the project" ruling, however, precluded them from introducing post-1958 sales of land located within Everglades National Park. Sales of "comparable" land eighteen years or older of course have only marginal probative value with respect to the fair market value of land in 1976, especially in an area which has experienced an escalating demand for recreational land in the interim. Thus, the landowners were forced to look outside the Everglades National Park for "comparable" mangrove sales.

But now Florida geography presented an obstacle. To begin with, most extant mangrove lands in South Florida are encompassed by the Park (which, undoubtedly, is one reason the Park was established — to preserve the primitive state and the rich vegetation and animal life of the coastal mangrove areas of South Florida). To complicate matters, the closest privately held mangrove lands outside the Park, adjoining the northwestern edge of the Park, were owned by a large land company, and had not been sold for many years. This meant that the only recent sales of non-Park mangrove lands were isolated sales at least 10 miles to the north and west of the Park, or along the eastern edge of the Park (about 60 miles from the landowners' properties).

Judge Mehrtens, however, ruled that these isolated, recent non-Park mangrove sales were inadmissible as not sufficiently comparable, apparently because of their distance from the condemned properties and because of their proximity to more developed, urban areas. This noncomparability ruling (excluding those few recent mangrove sales outside the Park which were discovered by the landowners' expert appraisers) coupled with the scope of the project ruling (excluding post-1958 in-Park sales) precluded the landowners from referring to any recent mangrove sales in all of Florida.

In their effort to provide some admissible, nonspeculative evidence of the fair market value of their properties, the landowners were therefore forced to refer to recent sales of sawgrass lands in an area of Big Cypress 7-25 miles from their properties. The Government's expert witness relied on sales of the same or similar sawgrass properties in making his appraisal of the fair market value of the lands. The Government's expert admitted that recent sales within the Everglades National Park would be comparable sales, typical of the marketplace. But in his opinion the sales of sawgrass land he relied upon, located just outside the Park boundaries, were also comparable sales, since the sawgrass lands were subject to the same regulations for building and development, were flooded every year for about six months, and had similarly limited accessibility. The landowners hotly contested the comparability of the sawgrass lands to their mangrove properties, however. They pointed to the marked differences in vegetation (mangrove forests versus sawgrass plains), the differences in animal life and sporting opportunities (saltwater fishing for snook, tarpon, etc. in the coastal mangrove lands; game hunting, such as deer and duck, and limited freshwater fishing in the Big Cypress sawgrass lands), differences in location (owners' mangrove lands were generally waterfront properties directly accessible by navigable waterways to the Gulf of Mexico, while sawgrass lands were interior marsh or swamp); differences in accessibility to specific tracts (owners' tracts could be reached by normal power boats; access to many sawgrass tracts, if legal access was available at all, had to be by overland swamp buggy or air boat), and alleged difficulties in locating and identifying specific tracts of sawgrass land (which often were part of a vast and undifferentiated plain, whereas waterfront mangrove properties supposedly could be readily identified by their land contours).

A comparison of the mangrove sales tendered by the landowners and the sawgrass sales actually introduced into evidence tends to support the testimony of the landowners and their witnesses that their coastal mangrove properties were not comparable to inland sawgrass lands. For example, the owners of the 59 small tracts involved in Trials II and III generally paid from $1,000 to $3,500 per acre for their parcels. (Many of them purchased their tracts between 1967 and 1971.) In addition, the average sale price of the few non-Park mangrove sales proffered, but excluded as noncomparable, exceeded $3,100 per acre. In contrast, the average sale price of the sawgrass sales actually introduced into evidence by the Government, and relied upon by the Government in making its appraisals of the fair market value of the landowners' properties, was about $325 per acre. This disparity is a strong indication that the landowners' coastal mangrove properties do indeed possess a significantly higher fair market value than do inland sawgrass properties. To be sure, the difference in value may be due, at least in part, to the location of the landowners' properties within the Everglades Park — an element of value which, under the "scope of the project" rule, might be noncompensable. For our present purposes, however, it is sufficient to observe that the Court's rulings — excluding post-1958 in-Park sales as well as all non-Park mangrove sales uncovered by the landowners' expert witnesses on grounds of noncomparability — prevented the landowners from introducing any actual sales figures to rebut the Government's argument that their properties were worth no more than inland sawgrass properties.

The Highest And Best Use Ruling

The Court's second crucial ruling — on "highest and best use" — also was made pursuant to the Government's motion of April 21, 1976. In his order granting that motion, Judge Mehrtens ruled that the landowners "shall make no reference at trial to evidence that the subject properties are suitable for cabin sites or other construction purposes."

During the three trials the landowners repeatedly objected to this ruling. They conceded the validity of the underlying principle — that while a condemnee is entitled to compensation reflecting the highest and most profitable use for which his property, at the time of taking, is adaptable and needed, or is likely to be needed in the reasonably near future, he is not entitled to compensation for potential uses that are not reasonably probable either because the land is not reasonably adaptable to such a use or because there is no likely demand for that use. But they challenged this particular application of the principle, basically on two grounds.

First, the landowners contended that cabins were a reasonably probable potential use, for which there was a market and a demand. To support this contention, they introduced by way of a proffer evidence that there were at least 18 cabins in the area, as well as several permanent houseboats or barges on which permanent camps had been erected. Indeed, four of the 63 tracts involved in the three trials had been improved with cabins or weekend homes. The landowners also tendered evidence that there was a demand for recreational cabins, and many of them testified that they had bought their properties with an intent to erect fishing camps or cabins on them. And although on-the-ground cabins could not be built on many of the properties, in large part because of their low elevation and muckish soil, the landowners proffered expert testimony that stilt houses and elevated boardwalks could be constructed without damaging the mangrove forest and in compliance with all governmental regulations.

Second, the owners argued that the issue of highest and best use was for the jury to decide, with proper instructions of course that speculative uses are neither to be considered nor compensated. The owners contended that Judge Mehrtens' ruling, precluding them from introducing any evidence that their properties were suitable for cabins, boardwalks, boat docks, etc., was tantamount to instructing the jury that their properties had no use whatsoever, and therefore erroneously trenched upon the province of the jury.

Judge Mehrtens, however, stuck by his original ruling that use of the properties for recreational dwellings was speculative and conjectural and had not been shown to be reasonably probable. He indicated that this ruling was based in part upon the failure of the landowners to prove that elevated cabins and boardwalks were practicable, and in part upon various regulatory restrictions that legally precluded any and all construction on the landowners' properties, whether practicable or not. Evidence had been introduced at a pre-trial hearing to the effect that coastal mangrove properties within the Northwest Extension were subject to the jurisdiction of, inter alia, the Monroe County Building and Zoning Department, the Monroe County Department of Environmental Health, and the United States Army Corps of Engineers. Although there was no indication that applicable zoning laws prohibited cabins on coastal mangrove properties, evidence was introduced at that hearing that permits for a septic tank, pit privy, or any other permanent sewage disposal facility would not be granted for properties in the area, and further that the Army Corps of Engineers would not grant permits for dredging, filling, or proposed structures for any property in the area under its jurisdiction. Apparently relying on this evidence, Judge Mehrtens disregarded the trial testimony of a landowners' expert that permits could be obtained for stilt houses or cabins and that such structures could be built in compliance with all applicable regulations, and ruled instead that any structures, even if practicable, could not legally be built.

With respect to whether the issue of highest and best use was a judge or jury question, Judge Mehrtens vacillated. His final position was that it is a jury question, and he so instructed the juries, but by excluding all evidence that the properties were suitable for buildings or other improvements, the Judge effectively decided the issue himself. Thus the consequence of his pre-trial ruling, whether correct or erroneous, was to relegate the landowners to valuing their properties as mere land, muck, and water — suitable perhaps for hunting, fishing, or communing with nature, but for no other recreational or residential uses.

The Ruling On Illegal Uses

As mentioned, four of the properties had already been improved with cabins or weekend homes. The owner of a 40-acre parcel included in Trial I had built a cabin with an attached porch on his property in 1958, and had further improved his property with approximately 24 feet of dock on creosote pilings. A 2-acre waterfront tract in Trial II was owned by a club of sportsmen, who regularly used the property on weekends for recreational purposes and to entertain business customers. Among the improvements on the property was a cabin which one landowner's appraiser valued at $30,000. Two other properties in Trial II had been improved with cabins erected on high and dry ground.

The juries, however, were not permitted to hear evidence of any of these existing improvements. At trial, Judge Mehrtens made explicit what was implicit in his pretrial ruling on "highest and best use" excluding reference to "cabin sites or other construction purposes" — that the ruling applied to existing as well as potential structures. Judge Mehrtens unequivocally based this extension of his "highest and best use" ruling on the principle that landowners can not be compensated for an illegal use. He interpreted this principle to hold that before landowners can be compensated for an existing structure, they must prove that the structure was legally erected under a proper permit. The four landowners were unable to satisfy this burden of proof, and consequently their properties, along with the other 59 properties, were valued as if they were both unimproved and not reasonably suitable for any improvements.

The Verdicts

In the end, the juries returned just compensation awards much closer to the Government's appraisals (as undeveloped and undevelopable mangrove land, based on purportedly comparable sales of Florida sawgrass land) than to the landowners' valuations. The disparities among the Government's appraisals, the landowners' valuations, and the juries' awards are presented by the table in footnote. To summarize

the table, the landowners were awarded an average of $251/acre for land which the Government valued at an average of $207/acre, which the landowners valued at an average of $1,298/acre, and which the landowners had purchased for upwards to $3,500/acre.

The Fatal Procedural Error

This concludes our story of the trials themselves. We report no more because the trials, as tense and bitterly contested as they were, were doomed to futility when the Government prevailed upon Judge Mehrtens to adhere to his crucial pre-trial rulings. And the problem with those rulings was not so much substantive — although in the circumstances of this case one was erroneous and the correctness of the other is not free from doubt. Instead, the fatal flaw is that the two crucial pre-trial rulings, which (along with the trial rulings on comparable sales and illegal uses) channeled the admissible evidence of fair market value within narrow bounds indeed, were in effect entered ex parte.

The unhappy story of how this came to pass is chronicled in detail in the footnote. The short of the matter is that

Judge Mehrtens apparently thought that these 63 cases had been before him at an earlier pre-trial hearing in which he heard evidence and arguments concerning the two crucial issues of "scope of the project" and "highest and best use." In reality, however, that hearing involved none of these 63 cases nor were any of the landowners or their attorney Mr. Earle present at that hearing. Judge Mehrtens' mistake is somewhat understandable midst the welter of Everglades condemnation cases pending before him, especially since the Government, in requesting the pre-trial rulings, represented to the Court that a hearing involving these 63 cases already had been conducted on the two crucial issues. But the Court's misplaced reliance on the Government's misrepresentation, although understandable, cannot be sustained. To bind litigants to rulings on the basis of evidence and arguments they have had no opportunity to contest is the most elementary sort of due process violation. We therefore are compelled to vacate the 52 judgments of just compensation which have been appealed and remand those cases for another trial.

* * * * * *

Our labors have just begun, however. In addition to asserting procedural error, the landowners have denominated 13 substantive or evidentiary rulings to which they assign error. The record, which contains numerous proffers of evidence the landowners would have presented the juries but for the restrictive rulings, is without question sufficiently developed for us to consider the merits of the landowners' arguments on these points. We decline, however, to address those alleged errors which are unlikely to recur on retrial. Other contested issues will inevitably, or at least quite probably, arise again on remand. Of particular importance are the application of the "scope of the project" rule and the principle of "highest and best use" to the unusual circumstances presented by the condemnation of private lands — largely wilderness, arguably dissimilar to all surrounding lands, and strictly regulated — within the boundaries of a pre-established National Park. To ignore these issues might well be to condemn any retrials to the futility of the original trials. But before turning to their consideration, we pause to review the basic precepts and principles of just compensation law, lest we lose sight of the forest for the trees.

II. BASIC PRINCIPLES

The Fifth Amendment itself, specifically the Just Compensation Clause of that Amendment, contains no mention of the "scope of the project" rule or the principle of "highest and best use." It merely provides that no "private property [shall] be taken for public use, without just compensation." Thus, the only standard of compensation mandated by the Constitution is that it be "just," which in turn "evokes ideas of `fairness' and `equity,'" see United States v. Commodities Trading Corp., 1950, 339 U.S. 121, 124, 70 S.Ct. 547, 549, 94 L.Ed. 707.

But what is just, fair, and equitable? To begin with, the Supreme Court has made clear that in the condemnation context, as in all areas of the law, justice is not measured from the jaundiced perspective of either the individual alone or the collectivity alone: "Just compensation means a compensation that would be just in regard to the public, as well as in regard to the individual * * *." Bauman v. Ross, 1897, 167 U.S. 548, 570, 17 S.Ct. 966, 975, 42 L.Ed. 270. Giving yet more content to the concept of just compensation, the Court has explained that the underlying principle is that the dispossessed owner "is entitled to be put in as good a position pecuniarily as if his property had not been taken. He must be made whole but is not entitled to more." Olson v. United States, 1934, 292 U.S. 246, 255, 54 S.Ct. 704, 708, 78 L.Ed. 1236. But these obvious (although sometimes overlooked) precepts do not, by themselves, decide the multitudinous condemnation cases with their almost limitless range of fact complexes. Supplementing these precepts are a number of "working rules" and "practical standards" developed by the courts in their endeavor to do substantial justice in eminent domain proceedings. United States v. Cors, 1949, 337 U.S. 325, 332, 69 S.Ct. 1086, 93 L.Ed. 1392.

Among these working rules and practical standards are the following. The measure of compensation is to be the value of the property at the date of taking. E.g., United States v. Miller, 1943, 317 U.S. 369, 374, 63 S.Ct. 276, 87 L.Ed. 336. In most cases, the "value" of the property can be "justly" determined by its monetary "market value." United States v. Toronto, Hamilton Buffalo Nav. Co., 1949, 338 U.S. 396, 402, 70 S.Ct. 217, 94 L.Ed. 195. "Market value" has been defined as "what it fairly may be believed that a purchaser in fair market conditions would have given," New York v. Sage, 1915, 239 U.S. 57, 61, 36 S.Ct. 25, 60 L.Ed. 143, or "what a willing buyer would pay in cash to a willing seller," United States v. Miller, supra, 317 U.S. at 374, 63 S.Ct. at 280. And since a hypothetical, "reasonable man" buyer will purchase land with an eye to not only its existing use but to other potential uses as well, fair market value takes into consideration "[t]he highest and most profitable use for which the property is adaptable and needed or likely to be needed in the reasonably near future * * * to the full extent that the prospect of demand for such use affects the market value while the property is privately held." Olson v. United States, supra, 292 U.S. at 255, 54 S.Ct. at 708. Thus, "just compensation" is not limited to the value of the property as presently used, but includes any additional market value it may command because of the prospects for developing it to the "highest and best use" for which it is suitable.

These rules and standards are not absolute and invariable, however. Some cases may involve peculiar circumstances in which it is impossible to determine a "market value." There may be other cases in which strict adherence to the concept of market value would involve inclusion of elements which, because they would result in manifest injustice to the owner or the public, must be eliminated from the compensation calculus. United States v. Miller, supra, 317 U.S. at 375, 63 S.Ct. at 280; United States v. Commodities Trading Corp., supra, 339 U.S. at 123, 70 S.Ct. at 549. For cases such as these courts have fashioned and applied other rules and standards, including the "scope of the project" rule discussed in the next Part. But neither are these secondary rules and standards absolute. The only absolute standard is that provided by the Constitution — "just compensation" — and the only sure guide in a difficult condemnation case is the consideration "What compensation is `just' both to [the] owner whose property is taken and to the public that must pay the bill?" Id.

With this consideration foremost in mind, we turn now to the questions raised by this appeal.

III. THE SCOPE OF THE PROJECT RULE [51] IIIA. The Substance Of The Rule

The "scope of the project" rule can be stated easily enough: If the condemned land was probably within the scope of the governmental project for which it is being condemned at the time the Government became committed to that project, then the owner is not entitled to any increment in value occasioned by the Government's undertaking the project. But, as this case so aptly illustrates, it is not so easily understood or applied.

It is most profitably understood as one of the secondary rules refining the concept of market value as the basic measurement of compensation so that injustice does not result from a naive or mechanical determination of market value. As it so happens, the "scope of the project" rule [hereinafter SOP rule] is primarily concerned with awards that are unjust from the perspective of the public footing the bill.

As such, it shares a kinship with other secondary rules or principles that have been developed to ensure that the Government, in pursuing public benefits through the power of eminent domain, is not forced to overcompensate private propertyholders. One such principle is that special value to the taker, or value created solely by the taker's demand for the property, is not compensable. E.g., United States v. Cors, supra, 337 U.S. at 332-34, 69 S.Ct. at 1090-1091; United States v. Weyerhauser Co., 9 Cir., 1976, 538 F.2d 1363, cert. denied, 1976, 429 U.S. 929, 97 S.Ct. 336, 50 L.Ed.2d 300; United States v. Whitehurst, 4 Cir., 1964, 337 F.2d 765, 771-72; United States v. Catlin, 7 Cir., 1953, 204 F.2d 661, 665; In re Valuation Proceedings under Regional Rail Reorganization Act, Spec. Ct. R.R.R.A., 1977, 445 F. Supp. 994, 1013-16; cf. United States v. Michoud Industrial Facilities, 5 Cir., 1963, 322 F.2d 698, 708-09, cert. denied, 1964, 377 U.S. 916, 84 S.Ct. 1180, 12 L.Ed.2d 185. "Since the owner is to receive no more than indemnity for his loss, * * [the property's] special value to the condemnor as distinguished from others who may or may not possess the power to condemn, must be excluded as an element of market value." United States v. Miller, supra, 317 U.S. at 375, 63 S.Ct. at 280.

At least three considerations support this principle excluding from compensation "value resulting from the government's special or extraordinary demand for the property" — value that is caused solely by the Government's actions as taker of the condemned property. First, this value is not true "market value" as determined by what a willing buyer would pay a willing seller under "fair market conditions." The Government has entered the "market" as a "purchaser" with a unique and pressing demand, and in so doing has distorted the market; absent the Government's activity as "purchaser" or condemnor, there would be no market reflecting this unique demand. That element of value created solely by the Government's activity as purchaser or condemnor is more "hold-up value" than "fair market value." Second, to force the Government to pay, because of a special public need for property, a premium over that which the property would bring on the open market absent the Government's demand, obviously would increase the cost of public projects and perhaps frustrate some public objectives. Third, to permit recovery of value that is not created by fair, open market conditions would be to award a few private propertyholders windfall gains solely because of public needs and exigencies. See generally United States v. Cors, supra, 337 U.S. at 332-34, 69 S.Ct. at 1090-91; Note, Valuation of Conrail Under the Fifth Amendment, 90 Harv.L.Rev. 596, 597-606 (1977). The underlying notion of the "no value attributable to Government demand" principle, then, is that the Government, when pursuing public benefits through its condemnation power, should not have to spend more for property than would a reasonable and willing private purchaser solely because it is exercising its condemnation power on behalf of the public; instead, the Government is to be equated to a private purchaser buying the property for its "highest and best" nongovernmental use in an open market.

The Supreme Court has also announced another, related refinement of "market value" compensation, excluding from compensation an element of value in order to ensure that condemnation awards are just, fair, and equitable to the public as well as to the dispossessed propertyholder. In a limited number of cases, the Court has held, in effect, that the condemnee is not to be compensated for the value of benefits conferred on his property by virtue of its proximity to Government property.

The leading decision announcing, and giving effect to, this principle is United States v. Fuller, 1973, 409 U.S. 488, 93 S.Ct. 801, 35 L.Ed.2d 16. In Fuller, the Government condemned privately-owned grazing lands that were adjacent to Government-owned grazing lands. The owner had been granted a revocable permit to use the Government lands in conjunction with his own lands, and contended that he was entitled to the value by which his lands were enhanced because of their actual or potential use in conjunction with the permit lands. Although the Court conceded that this element of value would be considered by a potential buyer on the open market, it held that the Government need not compensate the owner for this "Government-conferred" value. In so doing, it referred to "the general principle that the Government as condemnor may not be required to compensate a condemnee for elements of value that the Government has created, or that it might have destroyed under the exercise of governmental authority other than the power of eminent domain." 409 U.S. at 492, 93 S.Ct. at 804.

This "general principle," as applied in Fuller, is distinct from the "no value attributable to Government demand" principle. In the cases applying the latter principle, the value excluded is a direct product of the Government's actions as condemnor, and is, by definition, a value that could not be obtained from private purchasers; it is an artificial value not to be found in the private market. The value excluded in Fuller, on the other hand, could have been obtained in the private market; it was an element of real market value. As the Court conceded, a private purchaser of grazing lands would pay a premium for lands that adjoin Government grazing lands (especially when Government grazing permits are granted almost as a matter of course). Moreover, the value was not created by the Government's demand for the condemned property and its activity as condemnor, but rather by a special relationship between the condemned property and adjoining Government property that happened to benefit the private property.

Unlike the "no value attributable to Government demand" principle, the "no value of Government-conferred benefits" principle announced in Fuller is far from being a "general principle," much less a hard and fast rule. The only cases (other than "no value attributable to Government demand" cases) cited by the Supreme Court as examples of its application are the line of Supreme Court decisions holding that compensation for fast lands taken pursuant to the Government's navigational servitude with respect to navigable waters does not include as an element of value any benefits conferred upon the fast lands by their access to the navigable waters and potential use as portsites or hydro-electric sites. E.g., United States v. Rands, 1967, 389 U.S. 121, 88 S.Ct. 265, 19 L.Ed.2d 329; United States v. Twin City Power Co., 1956, 350 U.S. 222, 76 S.Ct. 259, 100 L.Ed. 240. Moreover, the Fuller Court admitted that other decisions defined limits, and in effect major exceptions, to this "no Government conferred value" principle. Nonetheless, Fuller does stand for the proposition that in some circumstances, a condemnee may not recover that element of fair market value which his property enjoys at the grace of the Government in part because of its special relationship to Government property.

We have considered in some detail these two sorts of excludable values — value attributable to Government demand and (in some circumstances) value of Government conferred benefits — because both values are relevant to the "scope of the project" rule. Furthermore, as we will see, both values arguably are at issue in the Everglades condemnation proceedings before us.

The origins of the SOP rule are found in two nineteenth-century decisions of the Supreme Court: Kerr v. South Park Commissioners, 1886, 117 U.S. 379, 6 S.Ct. 801, 29 L.Ed. 924, and Shoemaker v. United States, 1893, 147 U.S. 282, 13 S.Ct. 361, 37 L.Ed. 170. Both cases involved lands taken for an urban public park (in Chicago and the District of Columbia, respectively). In both cases there was evidence that subsequent to the legislative authorization of the park, lands in the area of the park-to-be increased appreciably in value. And in both cases, the Supreme Court approved lower court instructions that in effect informed the fact-finding body (a jury in Kerr and a commission in Shoemaker) that it was not to consider any such value attributable to the authorization of the park. The condemned properties could not be valued as parklands, for the simple reason that they had no use as parklands to their private owners. Nor could they be valued the same as neighboring tracts that were not condemned and ended up fronting upon the park; since the condemned lands were within the proposed park, they (unlike their neighbors) received no special benefits, and possessed no special value, resulting from the establishment of a park in the neighborhood. See Kerr, 117 U.S. at 385, 6 S.Ct. at 804. Finally, and in sum, the condemned properties were to be valued without "reference to * * * any supposed or speculative value given to [them] by reason of the act * * * creating the park project." Shoemaker, 147 U.S. at 305, 13 S.Ct. at 392-393. The values the landowners unsuccessfully sought to have considered were values created solely by the Government's undertaking the park projects for which their lands were condemned. In approving exclusion of these values, Kerr and Shoemaker can be seen as rather straightforward applications of the "no value attributable to Government demand" principle.

The SOP rule, which was as yet somewhat inchoate in Kerr and Shoemaker, was given full articulation in United States v. Miller, 1943, 317 U.S. 369, 63 S.Ct. 276, 87 L.Ed. 336. In Miller, the United States had condemned the subject lands for the relocation of railroad right-of-ways, made necessary by a Government dam-and-reservoir project, which, when completed, would inundate the old right-of-way. The project had been under consideration for a number of years; construction of it had been approved by the President in 1935; funds were appropriated for it by Congress in both 1936 and 1937; and in August 1937 Congress gave final authorization for the project. Although the need for new railroad right-of-ways had been anticipated early in the planning stages, and alternative routes had been surveyed and staked out in 1936, the actual location of the new railroad right-of-ways was not made known until the United States filed its declaration of taking in December 1938. Prior to 1937, the condemned lands at issue in Miller were largely uncleared brush lands. Sometime in 1937 they were plotted for subdivisions. By the date of taking in December 1938, they were part of a town that had been built up for business and residential purposes.

The issue in Miller, as stated by the Supreme Court, was whether the owners should "have the benefit of any increment of value added to the property taken by the action of the public authority in previously condemning adjacent lands," and, "[i]f so, were the lands in question so situate as to entitle respondents to the benefit of this increment?" 317 U.S. at 375, 63 S.Ct. at 281.

To decide this question, the Court referred to the following "working rules," which remain the definitive statement of the "scope of the project" rule:

If a distinct tract is condemned, in whole or in part, other lands in the neighborhood may increase in market value due to the proximity of the public improvement erected on the land taken. Should the Government, at a later date, determine to take these other lands, it must pay their market value as enhanced by this factor of proximity. If, however, the public project from the beginning included the taking of certain tracts but only one of them is taken in the first instance, the owner of the other tracts should not be allowed an increased value for his lands which are ultimately to be taken any more than the owner of the tract first condemned is entitled to be allowed an increased market value because adjacent lands not immediately taken increased in value due to the projected improvement.

The question then is whether the respondents' lands were probably within the scope of the project from the time the Government was committed to it. If they were not, but were merely adjacent lands, the subsequent enlargement of the project to include them ought not to deprive the respondents of the value added in the meantime by the proximity of the improvement. If, on the other hand, they were, the Government ought not to pay any increase in value arising from the known fact that the lands probably would be condemned. The owners ought not to gain by speculating on probable increase in value due to the Government's activities.

Id. at 376-77, 63 S.Ct. at 281.

Applying this test to the facts of the case, the Court held that the owners were not entitled to any increment in value due to the Government's project:

The project, from the date of its final and definite authorization in August 1937, included the relocation of the railroad right-of-way, and one probable route was marked out over the respondents' lands. This being so, it was proper to tell the jury that the respondents were entitled to no increase in value arising after August 1937 because of the likelihood of the taking of their property. If their lands were probably to be taken for public use, in order to complete the project in its entirety, any increase in value due to that fact could only arise from speculation by them, or by possible purchasers from them, as to what the Government would be compelled to pay as compensation.

Id. at 377, 63 S.Ct. at 282.

The two reported opinions in Miller — the Supreme Court decision and the circuit court decision which it reversed — contain strong indications that the owners had purchased and subdivided previously undeveloped land, gambling that it would increase considerably in value once the adjoining dam-and-reservoir project became a reality. And it is apparent from the Supreme Court opinion in Miller that the Court's foremost concern was to ensure that the owners would not extract, under the compulsion of the Just Compensation Clause, any speculative profits from the Government (and the public) because of the Government's activities as condemnor in completing the dam-and-reservoir project. Thus Miller as well is essentially yet another application of the "no value attributable to Government demand" principle.

The most significant aspect of Miller, however, is its statement of the "scope of the project" rule. For in that statement the Court categorizes compensable and noncompensable values due to Government projects: if, on the one hand, the property in question can be viewed as part of the project "from the beginning," any increment in value attributable to the project is not compensable; if, on the other hand, the property in question is taken pursuant to a subsequent decision to enlarge the project, the owner is entitled to compensation for "the value added in the meantime by the proximity of the improvement." This categorization was reaffirmed and explained further in United States v. Reynolds, 1970, 397 U.S. 14, 90 S.Ct. 803, 25 L.Ed.2d 12, the Supreme Court's only other major SOP decision:

The Court early recognized that the "market value" of property condemned can be affected, adversely or favorably, by the imminence of the very public project that makes the condemnation necessary. And it was perceived that to permit compensation to be either reduced or increased because of an alteration in market value attributable to the project itself would not lead to the "just compensation" that the Constitution requires. On the other hand, the development of a public project may also lead to enhancement in the market value of neighboring land that is not covered by the project itself. And if that land is later condemned, whether for an extension of the existing project or for some other public purpose, the general rule of just compensation requires that such enhancement in value be wholly taken into account, since fair market value is generally to be determined with due consideration of all available economic uses of the property at the time of the taking.

397 U.S. at 16-17, 90 S.Ct. at 805 (emphasis added and footnotes omitted).

This categorization, it should be observed, parallels the distinction developed earlier between "value attributable to Government demand" and the value of "Government-conferred benefits." In essence, the SOP rule excludes compensation for the former but commands compensation for the latter. By holding that the Government, in condemning property x as part of project A, need not pay compensation for value attributable to project A if property x was included in the project "from the beginning," the SOP rule ensures that the just compensation award will not reflect the Government's special demand for property x for inclusion in public project A. But the SOP rule also provides that if project A did not initially include property x, the Government must pay compensation for any additional value property x enjoys because of its proximity to project A as originally undertaken. This value, while in a sense created by the Government, is not attributable to the fact that the Government has a unique demand for property x and has committed itself to taking property x for inclusion in project A, but rather is the value of beneficial uses for which there is a private market demand, and, as such, is a true element of fair market value.

Nonetheless, Miller did not explicitly formulate the test for distinguishing between noncompensable and compensable value in terms of value attributable to the Government's special demand and its actions as condemnor (noncompensable) and the fair market value of benefits derived from the property's proximity to the Government project as originally undertaken (compensable). Instead, the shorthand test announced in Miller is whether "[the] lands were probably within the scope of the project from the time the Government was committed to it." 317 U.S. at 377, 63 S.Ct. at 281; United States v. Reynolds, supra, 397 U.S. at 21, 90 S.Ct. at 807. If so, the condemnee is not entitled to enhancement attributable to the project; if not, the Government must compensate the condemnee for any such enhancement.

As subsequently explained in Reynolds,

[A]pplication [of this test] to any particular set of facts requires discriminating judgment. The rule does not require a showing that the land ultimately taken was actually specified in the original plans for the project. It need only be shown that during the course of the planning or original construction it became evident that land so situated would probably be needed for the public use.

Id.

In Miller itself, the "discriminating judgment" called for by the test dictated that the enhanced value be excluded from the compensation award. Two facts in particular were relied upon by the Court in reaching this conclusion: first, from the date of its final and definite authorization in August 1937, the project had included the relocation of the railroad right-of-way, and second, one alternative route for the relocation had been marked out over the owners' lands in 1936. 317 U.S. at 377, 63 S.Ct. at 281. Thus, the owners undoubtedly had been placed on notice that the subject lands might be taken in conjunction with the dam-and-reservoir project.

Subsequent cases in the lower federal courts illustrate that application of the test and categorization of compensable and noncompensable values is often a more difficult task, requiring even more discriminating judgment, than in Miller. A case in point is United States v. Crance, 8 Cir., 1965, 341 F.2d 161, cert. denied, 1965, 382 U.S. 815, 86 S.Ct. 36, 15 L.Ed.2d 63, which involved the condemnation of 35 acres of land for public recreational facilities abutting a dam-and-reservoir project. In 1958 the Government had purchased about 7 other acres from the same landowners for inundation by the reservoir itself. From its inception, however, the project had contemplated the establishment of 20-40 acre recreational areas as well, located approximately every five miles around the reservoir. But neither a 1956 preliminary design memorandum nor a 1960 proposal of sites approved by the Chief of Engineers designated any of defendants' land for use as one of these recreational areas. However, after a 1960 public hearing announcing the proposal and after the public response to that proposal, the Government determined that another recreational site in the area of defendants' land was desirable. After further study, it decided to acquire 35 additional acres of defendants' property for this recreational site. On these facts (even though defendants' 35 acres were not among those originally designated for acquisition and the probability was slight that their property would ever be acquired under a recreational-area-every-five-miles principle), the court found that the additional acreage was within the scope of the original dam-and-reservoir project and therefore subject to condemnation without regard to enhancement.

A similar case in this Circuit involving land condemned as a recreational site adjoining a reservoir project is Louisiana, Through the Sabine River Auth. v. Lindsey, 5 Cir., 1975, 524 F.2d 934, cert. denied, 1976, 426 U.S. 948, 96 S.Ct. 3166, 49 L.Ed.2d 1184. In that case we held that the owners of the condemned recreational site were not entitled to enhancement when that site had been included in 1961 in the very first preliminary map of possible recreational sites, had been submitted as a proposed site to the Federal Power Commission in 1964, and had been approved as a recreational site by the FPC in 1967 more than four months before the condemnees signed a purchase agreement to buy the to-be-condemned land from its previous owner. Although the facts in Lindsey do not present as close a question as those in Crance, we observed that the "project scope is not to be narrowly interpreted." Id. at 942.

Change the facts ever so slightly, however, and the results are just the opposite: the landowner, instead of being compensated for his property without regard to the Government's project, receives that element of value conferred on his property because of its proximity to the Government's project. A striking example, particularly because of the close similarity of its facts to Crance, is United States v. 172.80 Acres of Land [ Mercer County], 3 Cir., 1965, 350 F.2d 957. This case, like Crance, involved a dam-and-reservoir project. Plans for the project, which included provision for small recreational areas in the immediate vicinity of the proposed dam, were approved in 1957, and in 1959 Congress appropriated funds for the project. The defendants in the case owned a 100-acre riparian tract some 2,000 feet upstream from the proposed dam on the side where no recreational facilities were planned. After negotiations with the owners, the Government purchased from them a fee interest in about 20 acres of their land that would be affected by flooding, as well as easements over about 14 more acres. During these negotiations, the Government representative assured the owners that the proposed acquisitions were the only parts of their land that would be needed for the project. Shortly thereafter, however, President Kennedy assumed office and there was a change in Government policy toward greater public development and use of reservoir areas. Pursuant to this change in policy, a decision was made in 1961 to expand the project and to acquire in fee the remaining 80 acres of the owners' lands for recreational and other public use. The owners were notified of this decision in October 1961, and in January 1962 the Government condemned the remainder of their land. The Third Circuit conceded that at the time the Government committed itself to the project in 1959, a perceptive landowner or prospective purchaser might have been aware of the possibility that the project would be expanded across the river to include this tract less than a half mile upstream from the dam. Nonetheless, Government plans at that time disavowed any such inclusion and, moreover, the owners had received specific assurance that their remaining 80 acres would not be taken. These factors, according to the Court, distinguished the case from Crance, and tipped the balance in favor of permitting the owners compensation for the project-enhanced value of their land as lakefront property, instead of its unenhanced value as farmland.

In sum, a purchaser contemplating the acquisition and private development of the Totten property after the 1969 initiation of the reservoir project could reasonably anticipate that he would be able to devote the land to its highest economic use, enjoying advantages inherent in the proximity of the nearby government development without serious apprehension that his land would soon be condemned. Thus, enhancement of value during this period was properly included in the subsequent condemnation award.

350 F.2d at 959. See also United States v. 959.68 Acres of Land, 3 Cir., 1969, 415 F.2d 401, cert. denied, 1970, 397 U.S. 964, 90 S.Ct. 991, 25 L.Ed.2d 256.

The Third Circuit's decision in Mercer County is closely paralleled by this Circuit's decision in United States v. 2,353.28 Acres of Land [ Brevard County], 5 Cir., 1969, 414 F.2d 965. The land in that case, owned by defendant Colton, had been condemned in 1963 for an expansion of Government facilities at Cape Canaveral, Florida (now Cape Kennedy). The project had had its origins in August 1961 when the Government, responding to President Kennedy's challenge to place a man on the moon by the end of the decade, decided to build a new launch site at Cape Canaveral and, toward that end, filed a complaint in condemnation describing a 72,644-acre tract bounded on the north by Haulover Canal. Colton's land, however, lay north of the canal and was not included in the original condemnation. In September 1961, a NASA official appearing at public hearings before a Senate committee to seek appropriations for the acquisition of the 72,644-acre tract, exhibited a map of that tract and testified that additional acquisitions were not anticipated. In June 1962, however, a NASA official revealed to the Senate committee that it needed additional funds to acquire 14,800 acres north of the canal, including the Colton property. Subsequently, in December 1963, the Government condemned Colton's land.

The Government of course argued that Colton was not entitled to any enhancement in the value of his land caused by the original project action in August 1961. We found, however, that the only evidence as of that date that the Colton land might be condemned was contained in two confidential NASA reports, and that these classified materials "did not serve as public notice of the project's scope." 414 F.2d at 970. We acknowledged the importance to the Government of flexibility during the planning stages of a project, but found this argument unpersuasive in view of the fact that NASA officials had conveyed to the public the impression that no land in addition to the 72,644 acres originally condemned would be needed. We concluded that the earliest disclosure by a project official that the Colton land probably would be taken did not occur until June 1962, and that prior to that date "a buyer in the real estate market reasonably could expect * * that the Colton land would not become a part of the lunar landing project, but would enjoy a position proximate to the project." Id. at 971. We therefore held that Colton was entitled to the value by which his land was enhanced between August 1961 and June 1962 because of its proximity to the Government project for which it ultimately was taken.

We have since derived from Brevard County the proposition that if enhancement due to the project for which property is taken is to be excluded under the SOP rule, "[i]t must * * * be evident to the public that [the property] might be taken for the project." Louisiana, Through the Sabine River Auth. v. Lindsey, supra, 524 F.2d at 942. In United States v. 62.17 Acres of Land [ Jasper County], 5 Cir., 1976, 538 F.2d 670, we identified three factors relevant to this determination of whether, from the time the Government committed itself to project A, it was evident to the public that property x might be taken for the project, id. at 680:

(i) The foreseeability that the original proposed dimensions of project A might have to be changed to include property x.

This factor, which figured prominently in the decisions in Crance and Lindsey, was also important in Jasper County. In that case, the Government had announced in 1960 that as part of a dam-and-reservoir project it would purchase in fee all lands in the project area below the 5-year flood line, which was calculated to lay on the contour 171 feet above mean sea level. After staking out the lines of acquisition drawn in a design memorandum the Government, in 1961, purchased from the defendant landowner's predecessor in title so much of the property as was included within the 171 foot line as then drawn and staked. In 1965 it became obvious that the Government had made some errors in demarking the 171 foot line. Because of congressional delay in appropriating additional funds, the perimeter area was not resurveyed and restaked until 1969. Jasper County involved the 1971 condemnation of 62 acres of land, bought by the defendant six months earlier as part of an 1800-acre tract, which were within the 171 foot line as resurveyed in 1969. In remanding the case, this Court observed that there was a reasonably foreseeable "likelihood of some mistakes in surveying and [consequently a] probability that this perimeter acreage would be subject to adjustment for such mistake." 538 F.2d at 681.

(ii) The length of time between commencement of project A and condemnation of property x.

In Jasper County we commented that the most disturbing factor in suggesting that the 1971 taking might be within the original scope of the project was the length of time that passed before the Government acted (public disclosure of initial 171 foot line in 1960, manifestation of mistake and need for adjustments in 1965, resurveying in 1969, condemnation in 1971). We noted, however, that the need for adjustments might not reveal itself until several years after impoundment began, and observed that "these projects are not of the stop watch variety and * * * gradualism in acquisition is oft times fact and not fiction." Id. at 680. The crucial question we left to the District Court on remand was whether "Government action or inaction after the need for such adjustment became clear made it no longer evident to the public that the taking here might be forthcoming." Id. Thus, the length of time before condemnation of property x does not, in and of itself, decide the scope question, but "time can be a factor in removing the mote of potential acquisition from the eyes of area landowners," id. at 681.

(iii) Government representations concerning the finality of project A as originally announced.

In Jasper County we found the record ambiguous concerning any Government representations that the original 171 foot line drawn in 1961 would be final, and we indicated that further factfinding with regard to Government disclosures of the project at its commencement might be dispositive of the scope question. This same factor, of course, was likely the crucial consideration in Mercer County (landowners had been assured upon initial acquisition that no more of their land would be needed) and Brevard County (project official represented publicly that initial acquisition would be sufficient to complete project).

This review of prior decisions that have struggled with whether lands were probably within the scope of the Government's original project or not illustrates that the "scope" test announced in Miller, while a convenient rule of thumb in many cases, does little more than restate the problem in difficult cases. In those cases, it is often necessary to confront the underlying problem head-on: would compensation for value attributable to the very project for which property is taken be just or unjust? Miller and Reynolds teach that compensation would not be just, and therefore is not required, if the additional value reflects the Government's special demand for the property and the fact that it is acquiring private property for a public project through eminent domain. But they also teach that compensation is required, and just, where the increment in value attributable to the Government project is instead an element of fair market value inherent in the property's proximity to the Government project.

Whether or not that increment in value is attributable more to the Government's special demand for the property or more to a private market demand for benefits conferred upon the property by its proximity to the Government project is largely a function of reasonable expectations. The crucial inquiry is whether, after commencement of project A but prior to condemnation of property x, the owner or a private purchaser contemplating acquisition and development of property x could reasonably anticipate that he would be able to devote that property to its highest economic use, enjoying the advantages inherent in its proximity to the nearby Government project, without serious apprehension that property x would soon be condemned.

The factors identified in Jasper County will often be relevant to this inquiry. For example, if the Government unequivocally represents to the landowner or the public that it will not need property x for project A, the Government in effect relinquishes property x as enhanced by its proximity to project A to private market forces. Although not as decisive, the length of time between commencement of project A and condemnation of property x may also serve to transform noncompensable value into compensable value: the longer the Government waits to condemn property x, the more it tends to abandon that property as benefitted by project A to the private market. Finally, the less foreseeable it is at the outset of project A that property x will not be needed, the more likely it is that the enhanced value is not due to the Government's special demand and its actions as condemnor, but instead is an element of fair market value that a reasonable private purchaser would pay for property x by virtue of its proximity to project A.

Regardless of how the inquiry is framed, however — whether in terms of the Miller test or in terms of reasonable expectations — the object is the same: to distinguish value attributable to Government demand from true fair market value of Government-conferred benefits, and to ensure that the landowner is not awarded a premium for the former but, at the same time, is justly compensated for the latter.

IIIB. Errors In The District Court's Application Of The Rule

With all this as a background, we are finally in a position to direct our attention specifically to these Everglades condemnation proceedings and consider how the SOP rule was applied below and how it should be applied on remand.

To recapitulate briefly, the pertinent facts are as follows: In 1947, after 18 years of discussion, Everglades National Park was established and dedicated with land donated by the State of Florida. In 1958, the boundaries of the Park were redrawn by Congress to include an area known as the Northwest Extension. At that time, Congress authorized the expenditure of $2,000,000 to purchase some 81,000 acres within the Park boundaries that were then privately owned, but Congress did not actually appropriate this sum until a number of years later, by which time it was inadequate to cover the purchase of at least 74,000 acres. In the meantime, the landowners in the area had been informed by the Department of the Interior that until Congress appropriated the necessary funds, they were free to sell, use, or improve their property as they saw fit. Congress authorized additional expenditures and appropriated the necessary funds early in the 1970's, at which time the Department of the Interior was able to begin in earnest the process of acquiring the remaining privately-held parklands, including the 52 tracts involved in this appeal.

In his pre-trial order of April 27, 1976, Judge Mehrtens ruled that under the SOP rule, the landowners could not offer evidence as to any sales that occurred within the boundaries of Everglades National Park after July 2, 1958 (the date upon which Congress redefined the boundaries of the Park to include the Northwest Extension). Because Everglades National Park encompassed almost all the coastal mangrove land in Florida, this ruling left the landowners with precious few post-1958 sales of coastal mangrove lands upon which to base their appraisals of the fair market value of their mangrove lands. And when the Court ruled that the few non-Park sales also were inadmissible on grounds of noncomparability, the landowners were put to the choice of basing their appraisals upon either in-Park sales of mangrove lands of at least 18 years vintage or sales of sawgrass lands, which they contended were not at all comparable to their properties.

In addition to the procedural errors involved in this pretrial ruling discussed earlier, the ruling also represented an erroneous application of the SOP rule to the facts of the case, for three reasons.

(1) Application Of The Rule To The Park In Its Entirety

All of the properties involved in this case were located within the Northwest Extension, which was added to the Park in 1958. The Northwest Extension, as the name itself signifies, was clearly an enlargement of the project. Miller and Reynolds unmistakably command that the owners of lands condemned pursuant to the enlargement of a Government project must be compensated for any enhancement resulting from the proximity of their lands to the project as originally defined. Miller, supra, 317 U.S. at 376-77, 63 S.Ct. at 281; Reynolds, supra, 397 U.S. at 16-17, 90 S.Ct. at 805; see pp. 785-788, supra. The 63 tracts in this case may have enjoyed such an enhancement by virtue of their proximity to the 1,000,000-plus acre Park as defined by the pre-1958 boundaries, but the pre-trial ruling entered below effectively excluded any evidence of such an enhancement.

All that the 1958 Act did was declare an enlargement of the Park and indicate that the Government would eventually condemn certain private lands within the Extension. Whether the Government thereby committed itself to acquiring the lands involved in this case is discussed below. But assuming for the moment that in passing the 1958 Act the Government did commit itself to the land acquisition program that eventually culminated in these condemnation proceedings, the Act nonetheless did nothing more than trigger the SOP rule to exclude compensation for value attributable to the land acquisition program itself and any special Governmental demand for properties within the Extension. The 1958 Act did not, however, suddenly terminate any benefits conferred upon lands within the Extension by the presence of the Park as defined prior to 1958. And, under Miller and Reynolds, the landowners are entitled to the fair market value of such benefits, if any, as of the date of taking ( i.e., May 1976).

To be sure, post-1958 sales of in-Park lands located in areas of the Park other than the Northwest Extension may themselves be tainted under the SOP rule. This problem of whether to admit into evidence sales arguably inflated by a Government project is discussed below. Our point at this juncture is that the District Court's ruling, by excluding post-1958 sales occurring elsewhere in the Park, erroneously excluded as an element of just compensation the fair market value, if any, the condemned properties enjoyed by virtue of their proximity to the Park as it existed prior to 1958.

(2) Application of the Rule as of 1958

While the SOP rule should not have been applied to exclude any value attributable to the Government's project unenlarged by the Northwest Extension, the rule is without question applicable to the Northwest Extension applicable to the Northwest Extension itself. But what is the "just" application of the rule with respect to the Northwest Extension? Was it "just" in the circumstances of this case to exclude from the compensation awards any and all value arguably attributable to the Northwest Extension after 1958?

In most "scope" cases, there has been a "second" round of takings, necessitated by an adjustment, redefinition, or expansion of the bounds of a project, and the critical question is whether property condemned pursuant to this second round of takings was "probably" within the scope of the project from the very beginning. In this case, the condemned properties were unquestionably within the scope of the project (the Northwest Extension) as originally announced and defined in 1958. Therefore, a mechanical application of the SOP rule, using the Miller test (Was the land probably within the scope of the project from the time the Government was committed to it?), would indicate that all value attributable to the Northwest Extension after 1958 must be excluded. But the rule is not to be divorced from its objective — compensation awards that are just to both the public and the dispossessed landowner. See Part II, supra. Viewed from this perspective, it becomes apparent that for three reasons it would not be just to the condemnees to extend the reach of the SOP rule back to 1958 in these condemnation proceedings.

First is the sheer length of time — here, 18 years. As we observed in Jasper County, the length of time between commencement of a project and condemnation of property for that project can be a "factor in removing the mote of potential acquisition from the eyes of area landowners." Of course, Jasper County involved the more typical "scope" case — the condemnation of property on the fringes of a project pursuant to an adjustment or redefinition of the project. This case, to repeat, is unusual as "scope" cases go in that there are no questions of probability — the lands were clearly within the scope of the project as originally defined and announced. But even in this situation, where the Government defines the bounds of a project and announces its intent to acquire eventually lands within those project bounds, there comes a point in the passage of time when it no longer would be just to apply the SOP rule in the Government's favor. Without any limits on the temporal reach of the SOP rule, the Government could encumber the free use and marketability of private property indefinitely simply by announcing a project and its intent to condemn property for that project some time in the future. Because of other factors present in this case in addition to the sheer length of time, we need not define here the "just" limits on the temporal reach of the SOP rule. We caution the Government, however, that 18 years may indeed exceed those limits.

A second reason why the SOP rule should not be applied in this case as of 1958 is that the Government's "commitment" to the project in 1958 is not very convincing. True, in passing the 1958 Act the Government did enlarge the Everglades National Park to include the Northwest Extension. But with respect to a land acquisition program, the 1958 Act merely authorized a niggardly $24/acre for the purchase of 81,000 acres located throughout the Everglades without actually appropriating any of this sum. An intent to acquire lands at some future date, unaccompanied by the appropriation of any funds for that purpose, hardly evinces a firm commitment, especially when the Government has permitted certain lands within numerous other National Parks to remain privately owned. Furthermore, it was not until 7 or 8 years later that the Government got around to appropriating even a paltry sum for the land acquisition program, and it was not until 1966 that the Department of the Interior first acquired private lands with federal funds. In these circumstances, a finding that the Government became sufficiently committed to acquiring private property within the Northwest Extension for purposes of the SOP rule cannot be sustained.

The assurances made by the Department of the Interior to area landowners constitute yet a third reason against extending the SOP rule back to 1958. In 4 letters sent between 1959 and 1962, the Department of the Interior informed the landowners that it could not begin its land acquisition program until Congress appropriated the necessary funds, that it did not know when Congress would in fact appropriate the funds, but that in the meantime the owners were perfectly free to use and sell their properties as if their lands were outside a Government project. These letters of course represent additional evidence that the 1958 "commitment" was more talk than substance. But they also indicate that, at least until 1962, there was a private market, and a fair market value, for the landowners' properties as benefited and enhanced by their location within the Northwest Extension of Everglades National Park. And, as discussed earlier, the landowners were entitled under Miller and Reynolds to compensation for the value of any such Government-conferred benefits.

Individually and together, these three factors — the sheer length of time between the 1958 Act and the takings, the lack of any firm commitment to a land acquisition program in 1958, and the Governmental assurances that the landowners could use and sell their properties freely — compel the conclusion that application of the SOP rule to exclude any and all value attributable to the Northwest Extension from 1958 to the date of taking was erroneous.

(3) Implementing The Rule By Means Of An Evidentiary Exclusion

If in any condemnation case it is appropriate and just to exclude from compensation under the SOP rule any value attributable to the Government's special demand for the properties it is taking, there remains the question of how procedurally to implement the rule and exclude the prohibited element of value. In this case, by ruling inadmissible all evidence of sales within Everglades National Park the District Court chose to give effect to the SOP rule by means of a sweeping evidentiary exclusion. This evidentiary exclusion was neither required by previous "scope" decisions, nor essential to the substantive principle excluding value attributable to the Government's special demand. Moreover, an evidentiary exclusion might well be unjust and improper given the particular facts of this case.

Courts have consistently recognized that, in general, comparable sales constitute the best evidence of market value. We have defined comparable sales as "sales from a willing seller to a willing buyer of similar property in the vicinity at or about the same time [as the taking]." United States v. Trout, 5 Cir., 1967, 386 F.2d 216, 223. As the definition indicates, comparability is largely a function of three variables: characteristics of the properties, their geographic proximity to one another, and the time differential. For any particular condemned property, there may be an entire spectrum of comparable open market sales: from almost simultaneous sales of adjoining, virtually identical property to sales of such dissimilar and distant properties occurring so long ago that they are not in any sense of the term "comparable" sales. Generally, the more comparable a sale is, the more probative it will be of the fair market value of the condemned property. In most cases, of course, there are no open market sales "ideally" comparable ( i. e., virtually identical characteristics, immediate vicinity, and within a short time of the taking), but instead an assortment of sales that are only reasonably comparable in all or several respects. Sound and just trial practice is to admit as many of the "most comparable" sales available as is necessary to fairly permit each side to present its argument of fair market value for the jury's consideration. As with all evidentiary matters, the trial judge has considerable latitude in admitting or excluding tendered "comparable sales," but it is not an unbounded latitude and the trial judge's rulings are reversible for abuse of discretion.

In this case, the District Court, in attempting to ensure that the landowners would receive no value that is noncompensable under the SOP rule, may well have unnecessarily and erroneously excluded from evidence the most probative and comparable sales available — sales of other coastal mangrove properties in the same general vicinity and within a reasonable time of the taking. Instead, under the Court's ruling, the parties were relegated to relying upon two other sets of "comparable" sales: (1) sales within the Northwest Extension that occurred before July 2, 1958, and (2) sales of inland sawgrass properties, 7 to 25 miles distant, which the landowners persuasively argued were dissimilar to their coastal mangrove lands in many respects.

To be sure, the sale price of lands within the Northwest Extension sold after 1958 may have reflected the fact that the Government planned eventually to acquire all privately owned parklands. Thus, post 1958 sales of parklands may be tainted by the same element of value that is noncompensable under the SOP rule. This consideration does not, however, automatically dictate that such sales be excluded from evidence.

Faced with the analogous problem of whether it is necessary to exclude from evidence all proffered comparable sales that occurred subsequent to the taking of the property whose fair market value is being determined, Judge Lumbard of the Second Circuit observed.

There is no absolute rule which precludes consideration of subsequent sales. The general rule is that evidence of "similar sales in the vicinity made at or about the same time" is to be the basis for the valuation and evidence of all such sales should generally be admissible, * * * including subsequent sales. * * * The generality of this rule is limited, however, by the consideration that a condemnation itself may increase prices and the government should not have to pay for such artificially inflated values. See International Paper Co. v. United States, 5 Cir., 1955, 227 F.2d 201. But that possibility does not produce a hard and fast exclusionary rule. In every case it is a question of judgment as to the extent of this danger and, particularly where a judge is sitting without a jury, it would seem the better practice is to admit the evidence and then to weigh it having due regard for the danger of artificial inflation.

United States v. 63.04 Acres of Land, 2 Cir., 1957, 245 F.2d 140, 144 (reversing the lower court for excluding evidence of subsequent sales). Other Circuits have followed suit in rejecting a per se rule excluding evidence of subsequent sales. E.g., United States v. 1,129.75 Acres of Land, 8 Cir., 1973, 473 F.2d 996, 998-99 (holding that it was an abuse of discretion for trial judge, in a jury trial, to exclude all comparable sales made subsequent to the date of taking); United States v. 691.81 Acres of Land, 6 Cir., 1971, 443 F.2d 461, 462-63; accord United States v. An Easement and Right of Way, 6 Cir., 1969, 405 F.2d 305, 308 (eminent domain case tried before a commission).

The reasoning of these subsequent sales cases applies with equal force to SOP cases. Rather than exclude from evidence through an absolute, pre-trial ruling all sales that may have been artificially inflated by the Government's nearby condemnation activities, in many cases the preferable course and the only just course would be to adopt a liberal policy of admission, permit the possibly tainted sales to go to the fact-finder under proper instructions, and leave it to the fact-finder to weigh them "having due regard for the danger of artificial inflation." After all, the presence of artificial inflation merely indicates that the "comparable" sales are in fact dissimilar to that extent, and generally, dissimilarities in sales that are otherwise comparable go to the weight of their evidentiary value, rather than to their admissibility, e.g., United States v. 691.61 Acres of Land, supra, 443 F.2d at 463; United States v. 124.84 Acres of Land, supra, 387 F.2d at 915; see United States v. Certain Land in the City of Fort Worth, supra, 414 F.2d at 1031.

We have been unable to find any decisions in which an appellate court has squarely addressed the question and concluded that the SOP rule requires an evidentiary exclusion. In their discussions of the substantive exclusion (no compensation for value attributable to the Government's special demand and its actions as condemnor), the Supreme Court's two major SOP decisions — Miller and Reynolds — contain no hint that an evidentiary exclusion is either necessary or, for that matter, proper. And this Circuit's opinion in Jasper County (heretofore the most extensive consideration of the SOP rule in this Circuit) carefully avoids any suggestion that an evidentiary exclusion is proper, despite the fact that the Government had sought "to exclude evidence of any enhancement" at trial, and that this Court agreed that the SOP rule might be applicable to the facts of the case.

There are, however, decisions affirming application of the SOP rule in cases in which the trial court had excluded evidence of sales possibly tainted by the Government's condemnation activities. Kerr and Shoemaker, the two nineteenth-century progenitors of the SOP rule in the Supreme Court, are among those decisions. So too is one decision from this Circuit — Wardy v. United States, 5 Cir., 1968, 402 F.2d 762. These decisions, however, did not squarely address the propriety of the evidentiary exclusion; rather, they were concerned with the propriety of the substantive exclusion or, as in Wardy, with some other collateral matter. And, significantly, there is no indication in these decisions that the condemnees were seriously hampered in their efforts to prove fair market value by the exclusionary evidentiary rulings.

This last consideration is an extremely important factor in evaluating the propriety of an evidentiary exclusion. To repeat, sound trial practice is to admit a liberal number of the "most comparable" sales available, leaving it to the fact-finder to assess the ultimate probative worth of any and all sales admitted. To exclude some of these sales on the ground that they may reflect a noncompensable element of value is not unfair so long as a reasonable number of other "most comparable" sales remain. If, for example, the parties offer 20 sales, all relatively recent and involving substantially similar properties in the same general vicinity, and it is reasonably possible that the price in 10 of the sales was influenced by the Government's condemnation activities, exclusion of the 10 tainted sales would still leave 10 more or less equally comparable sales for consideration by the fact-finder. Neither party could complain that the evidentiary ruling excluded altogether the best evidence of fair market value available. Moreover, in these circumstances, an evidentiary exclusion might well be preferable, as a means for giving effect to the SOP rule, to admitting all 20 sales with cautionary instructions that the fact-finder, in considering the sales as evidence of fair market value, must discount them to the extent it finds them tainted by an element of noncompensable value. The fairness of an evidentiary exclusion may be entirely different, however, if the exclusion would eliminate all or nearly all of the "most comparable" sales. For then the parties would be forced to resort to sets of "less comparable" sales (sales much further removed in time or distance, or sales of properties with markedly different characteristics), and the very factors that diminish the comparability of these sales vis-a-vis the excluded sales may significantly favor one party and disadvantage the other. Nor would the exclusion necessarily serve to expedite the trial proceedings in these circumstances, for it is likely that more time and effort would be devoted to litigating whether and what sort of adjustments must be made to reflect the greater dissimilarities between the condemned property and the "less comparable" sales.

Thus, where an evidentiary exclusion would eliminate the "most comparable" sales available, the better course would be to admit the possibly tainted sales with appropriate instructions, permit the parties to present testimony pro and con regarding whether the sales are tainted and how much the taint distorts true market value, and then instruct the fact-finder as to what elements of value must be included or excluded from the compensation award. The scope of the project rule commands only that value attributable to the Government's special demand and its actions as condemnor be excluded from compensation (and, conversely, that the value of Government-conferred benefits be included). It does not command that evidence be excluded. In many cases, the substantive exclusion can be accomplished as effectively through the use of proper instructions as it can through an evidentiary exclusion. And in some cases it is possible that only by the use of instructions can the substantive exclusion be given effect justly and fairly.

We do not decide if this is such a case. That depends largely upon a new determination, according to the principles announced in this opinion and after a full and fair evidentiary hearing, of whether and as of when the substantive SOP exclusion applies to the 52 tracts left in this case. We hold only that an evidentiary exclusion is not required and that, in the circumstances of these condemnation proceedings as revealed by the record before us, it is possible that an evidentiary exclusion would not produce a compensation award that is just to the landowner as well as to the public.

IIIC. Proper Application Of The Rule In These Eminent Domain Proceedings

If, as we have already concluded, it would be improper and unjust to apply the SOP rule as of 1958 to exclude any and all value thereafter attributable to the Northwest Extension, then how should the rule be applied to these takings? Because the parties have not yet had a full and meaningful opportunity to present and include in the record evidence relevant to a proper determination of this question, its answer must await further proceedings in the District Court upon remand. At this juncture we can do no more than review some of the pertinent considerations and provide a framework of analysis to guide the decision below.

As an initial matter, it should be remembered that the SOP issue is, after all, but one aspect of the just compensation determination. There is a danger that our extended discussion of the SOP rule might exaggerate its importance and obscure the principal function of the trial on remand — to determine the fair market value of the landowners' properties at the date of taking (which was May 1976), see Part II, supra. The SOP rule refines the concept of fair market value only with respect to alterations in value attributable to the Northwest Extension. It has no bearing whatsoever upon alterations in value attributable to other events or market forces. If the value of these coastal mangrove lands has increased between 1958 and 1976 because of factors other than the creation of the Northwest Extension, the landowners are entitled to compensation for any such increments in value regardless of how the SOP rule is ultimately applied. Thus, to the extent that the value of these properties has increased due to a general escalation in market demand for properties suitable for such recreational activities as camping, fishing, and boating — no matter where in South Florida such properties are located — the compensation awards must reflect that increase. Similarly, the awards must also include any increase in dollar value attributable to inflation. At the risk of belaboring the point, the SOP problem involves only whether the compensation awards for the 52 tracts remaining in this case should reflect any alteration in value occasioned by the expansion of Everglades National Park to include the Northwest Extension.

The reason the SOP issue constitutes something of a problem is, of course, that the SOP rule is dual-pronged. As articulated in Miller and Reynolds and their progeny, the rule commands that any increment in value attributable to the Government's special demand for the properties and its actions as condemnor must be excluded from the compensation awards, but that the landowners are entitled to compensation for the value of any beneficial uses their properties enjoy because of their relationship to the Northwest Extension. One of the perplexing aspects of this case is that it is plausible, more so than in many SOP cases, that both of these values are present — value attributable to Government demand as well as value attributable to Government conferred benefits.

The Government, for example, has argued that the landowners have derived unique and valuable benefits from the Wilderness Waterway — a north-south channel through the meandering waters of the Everglades, including the Northwest Extension, which has been marked and maintained by the National Park Service. This Waterway, according to the Government, serves much as a marine path through an aquatic jungle, enabling owners to locate their parcels of land more readily, and reducing the chances of becoming hopelessly lost while fishing and boating in the labyrinthine waters of the area. The Government also suggests that the landowners have been benefited by the presence of Park rangers in the area and the fact that the governmental regulations covering the Park preserve the unique characteristics, fish, and wildlife (and the concomitant recreational opportunities) of these ecologically sensitive coastal mangrove lands.

The Government contends that under the SOP rule the landowners are not entitled to the value of these benefits. Absent (1) the length of time between the announcement of the Northwest Extension and any significant land acquisitions in the area, and (2) the governmental representations back in the late '50s and early '60s that area landowners were free to use and sell their properties as they desired, this contention would have some merit. But by waiting 18 years to condemn these properties, and in particular by encouraging landowners in the interim to enjoy any and all attributes of ownership (including sales, exchanges, and resales), the Government in effect permitted a private market for these lands to continue. Moreover, the Government permitted that market to reflect the value of whatever beneficial uses these lands enjoyed because of their location within the Northwest Extension. Thus, contrary to the Government's argument, the landowners are entitled to the value of any Government-conferred benefits — such as proximity to the Wilderness Waterway — that might reasonably be enjoyed by private parties interested in purchasing these lands on the open market.

But the Government also argues that these lands were enhanced in value by the knowledge that the Government would eventually acquire them to complete the transfer of the Everglades from private to public ownership. Although a decade of congressional inaction and the Department of Interior letters may have negatived the possibility of imminent acquisition until the late '60s or early '70s, area landowners and the public knew, from 1958 on, that someday the Government would acquire these properties. This knowledge may have given these properties an additional value as speculative investments — buy cheap and hope to exact a profit from the Government when it finally commences its acquisition program. Of equal if not greater importance than the possibility of hold-up, the Government argues, was simply the assurance that there would be a buyer. The 1958 Act, according to the Government, provided a floor on down-side risk, which, in light of the roller-coaster, cyclical boom-to-bust nature of the South Florida real estate market, constituted a significant element of value.

These arguments are not implausible. And it is clear that any value derived from the knowledge that the Government would eventually acquire the properties — whether hold-up value or simply the insurance value of a guaranteed buyer — is not compensable. Any such value is attributable solely to the fact that the Government has a unique demand for these properties and the privilege to satisfy that demand through eminent domain. To permit the landowners to recover this value from the Government would be to force the Government to pay a premium solely because it is exercising its condemnation power on behalf of the public. The SOP rule had its origins in the perception that any value attributable to Government demand and its actions as condemnor is not an element of just compensation, and the rule commands that any such value be excluded from the compensation awards in this case.

Compounding the problem of how to treat any alterations in value caused by the Northwest Extension is the possibility, suggested by the landowners, that the value of their properties was not enhanced, but rather depreciated by the threat of condemnation. This too is not an implausible possibility. Anyone interested in buying these coastal mangrove lands for recreational use might well be dissuaded by the knowledge that their privileges of ownership would probably be relatively short-lived, and that they might not be able to recoup their investment through the negotiations or compensation trials that would follow governmental appropriation. And, just as it is clear that it would be unjust for private landowners to profit from the Government's condemnation activities, it is equally clear that it would be unjust for them to suffer because the promise of condemnation has depressed the private market. The principle that no weight should be given to the prospect of governmental appropriation is an evenhanded principle. Once the Government is committed to a project, depreciations in value caused by a prospective taking must be disregarded.

How are these possible elements of value — positive value attributable to Government-conferred benefits, positive value attributable to Government demand, negative value attributable to the threat of condemnation — to be sorted out and handled so that the compensation award is just to both the Government and the landowners?

This problem is best approached by reconsidering how the SOP rule operates. If applicable, the rule dictates that any alterations in value (positive or negative) attributable to the project for which property is taken are to be disregarded in ascertaining fair market value and just compensation. Under the traditional formulation of the SOP test, the crucial date which triggers the rule is the date the Government became committed to the project. The SOP rule, then, operates much like a presumption: as of the crucial date of Government commitment, it is presumed that any positive alteration in value is attributable to Government demand (and therefore must be excluded from the compensation award), and that any negative alteration in value is similarly attributable to the Government's condemnation activities (and must also be disregarded). The key to a just determination of the SOP issue is to set the date as of which the rule is triggered with due regard for the compensation consequences, so that the presumption does not unfairly favor either the Government or the landowner.

In most cases where the condemned property has been found to have (probably) been within the scope of the project from its inception, a just "commitment" date triggering the rule will be the date the Government announced the project. In this case, however, we have seen that three factors militate against triggering the rule as of the date of the project announcement (July 2, 1958): first, the length of time between that announcement and the takings; second, the mere token commitment at the time of the project announcement to the land acquisition program necessary to complete the project; and third, the governmental blessing given area landowners to use and sell their properties freely. The question to be decided on remand, then, is, if not 1958, at what date would it be fair and just to trigger the SOP rule and its presumption that alterations in value thereafter attributable to the project must be disregarded?

As foreshadowed in our earlier discussion, the appropriate date is largely a function of reasonable expectations. It is the date as of which the landowners or prospective purchasers no longer could reasonably anticipate being able to devote these properties to their highest and best use in the context of the surrounding governmental project, without serious apprehension that the properties would soon be condemned. In other words, it is the date as of which the prospect of imminent condemnation becomes sufficiently definite that it would be a major factor in the decision of any reasonable person to buy or develop the property. From this date on, it may fairly be presumed that any depreciation in value is attributable to the threat of condemnation, and that any increase in value is attributable to the Government's special demand for the property and its actions as condemnor.

Whatever the date turns out to be after a proper consideration of relevant evidence, fully and fairly presented, the rule for trial will be that compensation must be awarded without regard to any increase or decrease in value attributable to the Northwest Extension after that date. And, under the rule, the question for the fact-finder will be: what would land having the characteristics of these properties (and the characteristics include whatever beneficial uses these properties derived from the Northwest Extension as of the crucial date) be worth in the private real estate market as of May 1976 (the date of taking)?

IIID. Division Of Responsibilities Between Judge And Jury (Or Commission)

Implicit in the foregoing is that the primary responsibility for deciding the scope of the project issue is committed to the trial judge. This has unequivocally been the law in this Circuit since our 1968 decision in United States v. Wardy, supra, which was expressly approved by the Supreme Court in United States v. Reynolds, supra, as the law in all federal courts. Nonetheless, as explained below, the jury (if there is one) shares to a limited extent responsibility for ensuring that the SOP rule is properly and justly reflected in its compensation award.

Because there is no constitutional right to a jury trial in eminent domain proceedings, the distribution of responsibilities between judge and jury is governed by F.R.Civ.P. 71A(h), which, along with the other sections of Rule 71A, defines the procedures for federal court condemnation proceedings. Rule 71A(h) limits the role of a jury in two important ways. First, it provides that in some eminent domain proceedings there need not be a jury at all: Congress may designate and constitute a special tribunal for trial of the issue of just compensation, or even where Congress has not so designated a special tribunal, the trial judge may, in certain broadly defined circumstances, exercise his discretion and appoint a three-person commission to try the issue of just compensation in lieu of a jury. Second, the Rule provides that even where a jury is used, it decides only the issue of just compensation and "[t]rial of all [other] issues shall * * * be by the court."

In Wardy, we held that the scope of the project issue — specifically, whether the condemned "lands were probably within the scope of the project from the time the Government was committed to it" — was among the "other issues" committed by Rule 71A(h) to the trial judge. As we explained, the judge, in deciding the SOP issue, "merely decide[s] a legal question which limit[s] the factors necessary to the determination of `just compensation.'" 402 F.2d at 763. The Supreme Court further elaborated upon this in Reynolds:

[W]e think the Rule's basic structure makes clear that a jury in federal condemnation proceedings is to be confined to the performance of a single narrow but important function — the determination of a compensation award within ground rules established by the trial judge. The Rule gives the trial court discretion to eliminate a jury entirely. And when a jury is afforded, the sweeping language of the final sentence of the Rule discloses a clear intent to give the district judge a role in condemnation proceedings much broader than he occupies in a conventional jury trial. It is for him to decide "all issues" other than the precise issue of the amount of compensation to be awarded. It follows that it is for the judge to tell the jury the criteria it must follow in determining what amount will constitute just compensation, and that in order to do so he must decide the "scope-of-the-project" issue as a preliminary matter.

397 U.S. at 20, 90 S.Ct. at 807.

While it is thus clear that the SOP issue is a "preliminary matter" whose determination is committed to the trial judge by Rule 71A(h), this is an oversimplification to the extent that it suggests that the jury (or the commission or special tribunal) has no responsibilities whatsoever for effectuating the SOP rule. What the trial judge decides, essentially, is whether — under the Fifth Amendment and subsidiary principles of just compensation — the jury is to consider any alterations in value attributable to the Government project for which the condemned property is taken. In other words, the judge decides whether any increase in value due to the project must be excluded from the compensation award (because it is value attributable to the Government's special demand and its condemnation activities) or instead must be included in the award (because it represents the fair market value of Government-conferred benefits). In addition, the judge decides whether any depreciation in value must be disregarded by the jury in determining compensation (because attributable to the threat of condemnation). Under the SOP rule, the judge makes these preliminary determinations of value that must be included or disregarded by deciding whether the condemned lands were probably within the scope of the project from the time the Government was committed to it, or (using the more precise formulation of the test), by deciding the date as of which the owner or a prospective purchaser no longer could have reasonably anticipated being able to devote the condemned lands to their highest and best use, enjoying the beneficial uses if any derived from a nearby or surrounding public project, without serious apprehension that the property would soon be condemned for that project. In sum, the judge decides (a) whether the evidence warrants application of the SOP rule, (b) how the rule should be applied — i.e., whether or not any change in value attributable to the project should be included or disregarded, and (c) where appropriate, the date after which any change in value should be disregarded.

But whether there in fact has been any increase or decrease in value due to the project, and by how much, is ultimately for the jury (or commission) to decide as part of its just compensation determination on the basis of the evidence presented it. Thus, although the landowners may produce sufficient evidence to raise the possibility that their properties depreciated in value because of the threat of condemnation, and by virtue of that possibility to necessitate appropriate instructions, the jury is free to disregard that evidence in favor of contrary evidence presented by the Government. Nor is the jury required to credit landowners' estimates of the fair market value of compensable beneficial uses derived from the Government project. And by the same token, the jury may decide that the evidence supports the landowners' estimates of fair market value despite Government evidence that those estimates include a noncompensable element of value attributable to the Government's special demand and its condemnation activities.

To echo the Supreme Court's language in Reynolds, the judge establishes the "ground rules" or "criteria" for the determination of just compensation by the jury. And he enforces those ground rules through a combination of evidentiary rulings (tempered by the concerns discussed in Part IIIB(3), supra) and instructions. But once he has instructed the jury to disregard any value attributable to the project after a certain date, his responsibility ends, and it becomes the jury's responsibility — as with any other legal instruction — to follow the instruction in its fact-finding mission. And so long as the jury's verdict is within the range of the evidence presented it, it must be assumed that the jury fulfilled its responsibilities faithfully.

IV. COMPARABLE SALES

Before turning to the other crucial pretrial ruling (relating to highest and best use), we pause to consider the landowners' contention that the District Court erred in excluding 5 sales of mangrove lands located outside Everglades National Park which the landowners had tendered as comparable sales. According to the landowners, this ruling, coming on top of the pre-trial ruling excluding all post-1958 sales within the Park, vitiated their case. The only "comparable" sales to reach the juries were inland sawgrass sales which averaged about $325/acre. The landowners identified a number of differences between their coastal mangrove properties and inland sawgrass properties, but having been prohibited from introducing the sales price of any coastal mangrove lands (which averaged about $3100/acre), they were unable to prove their position that because of these differences their lands commanded a much higher fair market value than did inland sawgrass properties.

In Part IIIB(3), supra, we held that the SOP rule does not require the exclusion of sales within the scope of the project which are offered as comparable sales. Moreover, we strongly indicated that an evidentiary exclusion might constitute reversible error where it would compel the parties to rely on distinctly "less comparable" sales to the prejudice of one of the parties. For the same reasons which led us to this conclusion concerning comparable sales within the project, we hold that the District Court erred in excluding the 5 non-Park comparable sales in the trials below.

We recognize that the decision upon whether to admit or exclude tendered comparable sales is committed to the sound discretion of the trial judge. And in this case, the Court's decision to exclude the 5 mangrove sales in question was not wholly arbitrary; Judge Mehrtens indicated that the 5 tendered sales were too far away from the landowners' properties and too close to developed, urban areas to be truly comparable. But the inland sawgrass sales that were admitted were equally dissimilar to the condemned properties, and perhaps more so, albeit in different respects. Although the sawgrass sales, like the condemned properties, were also fairly distant from urban areas, they possessed substantially different physical characteristics and were not suitable for the same recreational uses. From the standpoint of admissibility — namely, the degree of comparability — there was no reason to prefer the sawgrass sales over the 5 excluded mangrove sales. To be sure, the dissimilarities between the 5 mangrove sales and the condemned properties may not have affected market value to the same extent. Proximity to urban areas might have been a more significant element of value than the unique characteristics of coastal mangrove lands. But the extent to which dissimilarities are reflected in market value is a question of the evidentiary weight of the comparable sales, which is a question for the jury. By excluding the 5 mangrove sales, the Court in effect substituted itself for the jury, and there is a strong likelihood that by so doing the landowners were substantially prejudiced since they were left with no tangible evidence to support their contention that the differences between coastal mangrove properties and inland sawgrass properties translated into differences in fair market value. In these circumstances — where but one of two sets of sales possessing relatively equal degrees of comparability is admitted, coupled with a strong likelihood of substantial prejudice to one of the parties — it was an abuse of discretion to exclude the 5 tendered mangrove sales.

V. HIGHEST AND BEST USE

The second crucial pre-trial ruling which the landowners contest concerned the issue of the highest and best use of their properties. One of the basic principles of eminent domain law is that condemnees are entitled not just to the value of their properties as used at the date of taking, but rather to the value their properties would command in the open market in light of the highest and most profitable use to which they might reasonably be devoted in the near future. Perhaps the best summarization of the law with regard to highest and best use is found in the Supreme Court's opinion in Olson v. United States, 1934, 292 U.S. 246, 54 S.Ct. 704, 78 L.Ed.2d 1236:

Just compensation includes all elements of value that inhere in the property, but it does not exceed market value fairly determined. The sum required to be paid the owner does not depend upon the uses to which he has devoted his land but is to be arrived at upon just consideration of all the uses for which it is suitable. The highest and most profitable use for which the property is adaptable and needed or likely to be needed in the reasonably near future is to be considered, not necessarily as the measure of value, but to the full extent that the prospect of demand for such use affects the market value while the property is privately held. * * * But the value to be ascertained does not include, and the owner is not entitled to compensation for, any element resulting subsequently or because of the taking. Considerations that may not reasonably be held to affect market value are excluded. * * *

* * * [Market value is] the amount that in all probability would have been arrived at by fair negotiations between an owner willing to sell and a purchaser desiring to buy. In making that estimate there should be taken into account all considerations that fairly might be brought forward and reasonably be given substantial weight in such bargaining. * * * The determination is to be made in the light of all facts affecting the market value that are shown by the evidence taken in connection with those of such general notoriety as not to require proof. Elements affecting value that depend upon events or combinations of occurrences which, while within the realm of possibility, are not fairly shown to be reasonably probable should be excluded from consideration, for that would be to allow mere speculation and conjecture to become a guide for the ascertainment of value — a thing to be condemned in business transactions as well as in judicial ascertainment of truth.

With respect to the highest and best use of these coastal mangrove properties, Judge Mehrtens granted the Government's motion in limine and ruled prior to trial (and without a proper evidentiary hearing) that the landowners "shall make no reference at trial to evidence that the subject properties are suitable for cabin sites or other construction purposes." During the trials, he adhered to and enforced this ruling over the landowners' repeated objections that their undeveloped properties were suitable for cabins erected on stilts, for boat docks, and for elevated boardwalks. Although he never explicitly set forth his reasons for excluding evidence that the highest and best use of the properties was as sites for the sorts of permanent improvements proffered by the landowners, the record indicates that Judge Mehrtens based this ruling on two considerations. First, he evidently believed that stilt cabins, boat docks, etc., were impracticable, and hence speculative uses for which compensation may not be awarded as a matter of law. But second, even if practicable uses for which there was a market demand, Judge Mehrtens believed that various regulatory restrictions precluded these uses without a permit, and that absent proof that the necessary permits had been or could be obtained, the properties could not be valued as suitable for buildings or other improvements.

The landowners contend that this ruling, the result of which was that their lands were valued as unsuitable for any development purposes, was erroneous for three reasons. First, the owners contend that to the extent the ruling was based on any pre-trial evidentiary hearings to which they were not a party, it is void for failure to satisfy the requirements of due process. Their second argument is that the evidence they tendered during the trials on highest and best use was not speculative, and therefore should have been admitted for the jury's consideration in determining the fair market value of their properties. Finally, the landowners argue that the issue of highest and best use is a jury issue, and that the Court's ruling, tantamount as it was to an instruction that their properties had to be valued as undevelopable virgin land, erroneously trenched upon the province of the jury.

We have of course already addressed the landowners' first argument and ruled in their favor. Although at trial the landowners did proffer in summary form evidence pertaining to the suitability of their lands for permanent structures and some conclusory testimony pertaining to permits, it is apparent that Judge Mehrtens based his ruling primarily upon evidence presented elsewhere, presumably at the February 5th evidentiary hearing involving other tracts and other landowners. Thus, the landowners in the three May trials currently before us never had a meaningful opportunity to present fully their evidence on the feasibility of permanent structures or to rebut the Government's contention that their properties were within the jurisdiction of various regulatory authorities which would not grant the necessary permits for any permanent structures. Regardless of the merits of the landowners' second and third contentions, the ruling on highest and best use cannot stand because of this most basic failure of due process.

As the landowners point out, the only evidence on highest and best use introduced in proceedings to which they were a party consisted of their proffers during the trial, to the effect that stilt houses, boatdocks, and elevated boardwalks were feasible uses for which there was a market demand and for which there was a reasonable probability of obtaining permits. Consequently, the owners urge us to hold that stilt houses and the like are not speculative uses. We decline to do so. The record on this issue not only has been developed in a most awkward fashion, but is as yet quite sparse. Because the Government had no reason to call rebuttal witnesses or otherwise respond to the owners' trial proffers, the record before us is virtually one-sided with little more than conclusory assertions at that. Although we are reluctant to make any predictions, it seems to us that the issue is far too complex to be decided on the present record. Determination of the highest and best use issue, as with the scope of the project issue, must await further proceedings in the District Court in accordance with the principles announced in this opinion.

Because it will inevitably be of central importance to the proceedings on remand, we turn to the landowners' third argument — that the issue of highest and best use is committed to the fact-finder (the jury or a commission). We have already observed that F.R.Civ.P. 71A(h) commits only the ultimate issue of "just compensation" to the fact-finder, leaving determination of "all [other] issues" to the court. And we have held that pursuant to this division of responsibilities under Rule 71A(h), determination of the scope of the project issue is delegated to the judge to the extent that the judge decides whether the rule is applicable, how it should be applied, and the date after which all changes in value attributable to the project must be disregarded by the fact-finder. The question we consider here is the appropriate division of responsibilities under Rule 71A(h) with respect to highest and best use.

In light of the fact that highest and best use is an issue in almost every eminent domain case (whereas scope of the project is an issue in only a small percentage of condemnation cases), it is somewhat surprising that this question apparently has been considered only once before by an appellate court, and even in that case, which comes from this Circuit, the question was not resolved. In United States v. 158.24 Acres of Land [ Bee County], 5 Cir., 1975, 515 F.2d 230, the appellant landowner contended that the District Court had erroneously invaded the jury's domain (just compensation) in the guise of deciding "other" legal and factual issues when it ruled that the highest and best use of the condemned property was ranching (its existing use), and excluded the landowner's testimony that the highest and best use was subdivision into small rural tracts held for recreational and residential purposes. In affirming the District Court we noted that the judge's power under Rule 71A(h), as construed by the Supreme Court in United States v. Reynolds, supra, "may be broad indeed," 515 F.2d at 232 n. 3, but we declined to define the outer parameters of that power. Instead, we affirmed on the ground that the proffered evidence as to subdivision was "speculative," and therefore "could have been excluded by the trial court on this ground." Id. at 232.

The principle that just compensation cannot be predicated upon potential uses which are speculative and conjectural is of course firmly established. It is but the converse of the principle, as stated in Olson, that just compensation must take into account "[t]he highest and most profitable use for which the property is adaptable and needed or likely to be needed in the reasonably near future," 292 U.S. at 255, 54 S.Ct. at 708. Indeed, in Olson itself, the Court held that the condemned lands of the petitioners could not be valued as suitable for use as a reservoir where the landowners had failed to prove that there was any reasonable probability of their acquiring the numerous flowage easements necessary to adapting their lands for reservoir purposes. See id. at 260, 54 S.Ct. 704. To like effect is United States ex rel. TVA v. Powelson, 1943, 319 U.S. 266, 63 S.Ct. 1047, 87 L.Ed. 1390, in which the fact-finder valued the condemned property as suitable for a hydroelectric plant, but the Supreme Court reversed because there was insufficient evidence of a reasonable likelihood that the condemned lands could be incorporated as part of a power project within the reasonably near future. See also McGovern v. New York, 1913, 229 U.S. 363, 372, 33 S.Ct. 876, 57 L.Ed. 1228. In a similar case from this Circuit, United States v. Cooper, 5 Cir., 1960, 277 F.2d 857, we too overturned a just compensation award based on the value of the land as a potential hydroelectric site, on the ground that there had been a "complete failure of proof" that use as a dam site was a reasonably probable use, id. at 859. Relying on Olson, we held that the prerequisite to having land valued on the basis of a particular potential use was proof that the use was practicable and that there was a reasonable likelihood that the land would be so used in the reasonably near future. Id.

Absent such proof, there would be no legitimate reason for concluding that there is any market demand for the property as so used, and therefore no legitimate reason for concluding that the property has any additional market value because of its asserted suitability for that use. In short, absent such proof the alleged potential use remains a speculative use for which, as a matter of law, the landowner cannot be compensated. Consequently, there is no reason to allow the jury (or commission), in deciding the issue of just compensation, to even consider any use that is not reasonably probable. "Elements affecting value that depend upon events or combinations of occurrences which, while within the realm of possibility, are not fairly shown to be reasonably probable should be excluded from consideration * * *." Olson v. United States, supra, 292 U.S. at 257, 54 S.Ct. at 709. At the very least, then, the trial judge has the responsibility under Rule 71A(h) to screen the proffered potential uses and exclude from jury consideration those which have not been demonstrated to be practicable and reasonably probable uses.

But should the trial judge go further and actually decide, from among various practicable and reasonably probable uses, what one use (or combination of uses) actually represents the highest and best use of the property for purposes of the just compensation determination? A number of cases indicate that he should not, that if the landowner's evidence is sufficient to show that a prospective use is feasible and reasonably probable, the landowner is then entitled to have that evidence considered by the jury (or the commission).

Among these decisions is one from the Supreme Court — McCandless v. United States, 1936, 298 U.S. 342, 56 S.Ct. 764, 80 L.Ed. 1205. The condemnees in that case had tendered evidence that, by bringing in water from neighboring lands for irrigation purposes, their lands could be converted from their existing use as a cattle ranch to a more profitable use as a sugar plantation. The trial court did not permit the jury to hear this evidence. The Supreme Court reversed. It first reiterated the principle of Olson that the most profitable use to which the land can probably be put in the reasonably near future should be considered to the extent it affects market value. It then held, contrary to the conclusion of the trial court, that on the basis of the landowners' tendered evidence the possibility of obtaining water from outside sources was not so remote and speculative as to preclude the jury from considering as an element of value the potential use of the land for the cultivation of sugar cane. 298 U.S. at 345-46, 56 S.Ct. at 765.

A similar, and particularly instructive, decision is Hembree v. United States, 8 Cir., 1965, 347 F.2d 109. In that case, the Government had contended that the highest and best use of the condemned properties was limited to grazing purposes. The landowners, on the other hand, claimed that the highest and best use was either commercial quarrying for limestone or residential development. Although the landowners offered evidence in support of both these claims, the trial court ruled that evidence insufficient to warrant submission to the jury. The Eighth Circuit affirmed concerning the insufficiency of the evidence as to residential development, but held that it was error to exclude the landowners' evidence pertaining to potential use as a dolomite limestone quarry. After reviewing the landowners' proffer against the principles announced in Olson, the Eighth Circuit concluded that "there was * * * evidence from which a jury could permissibly find for appellants on this issue," id. at 113. Accordingly, the Court held, potential use as a limestone quarry should have been submitted to the jury:

We are convinced * * * that the [trial] court's decision was dictated by its evaluation of the weight of the evidence which turned in part at least on credibility. The court, in the role of the fact-finder, concluded that there was no evidentiary basis for the claim that the value of the condemned land was enhanced because of the limestone deposit. Perhaps a jury would reach the same conclusion but that was not for the trial court to determine and is not, of course, an issue before us. We are constrained to hold that the evidence provided a sufficient factual background to require submission of the limestone issue to the jury.

Id. at 113-14.

To like effect is United States v. 100 Acres of Land [ Marin County], 9 Cir., 1972, 468 F.2d 1261, cert. denied, 1973, 414 U.S. 864, 822, 94 S.Ct. 37, 119, 38 L.Ed.2d 84, 54, in which the Ninth Circuit had to decide whether the District Court had erred in permitting the jury to decide the highest and best use issue. At trial, the Government had urged that the highest and best use of the subject property (a small peninsula of sandy real estate about 45 miles north of San Francisco) was for recreational purposes only. The landowner, however, offered expert testimony that the highest and best use was a recreational and residential development, involving subdivision into beach lots. The trial judge permitted this testimony to reach the jury, after finding that there was a sufficient foundation for it. The Ninth Circuit affirmed:

It appears to be well settled that in determining market value of condemned property, evidence as to the reasonable probabilities of its use is admissible and may be considered. See McCandless v. United States, supra * * *. The weight to be given evidence of reasonable probability is a question for the trier of fact. Cf. Olson v. United States, supra. Evidence in regard to prospective uses is admissible and whether adaptability and probability affect market value is for the jury.

468 F.2d at 1267-68. See also United States v. 1,291.83 Acres of Land, 6 Cir., 1969, 411 F.2d 1081; Atlantic Coast Line R.R. v. United States, 5 Cir., 1943, 132 F.2d 959.

To be sure, all but one of these decisions antedate — and the other ( Marin County) does not mention — the Supreme Court's opinion in United States v. Reynolds, supra, in which the Court stated that under Rule 71A(h) "[i]t is for [the judge] to decide `all issues' other than the precise issue of the amount of compensation to be awarded," 397 U.S. at 20, 90 S.Ct. at 807; see p. 808, supra. Nonetheless, we align ourselves with these decisions and hold that once the landowner has produced credible evidence that a potential use is reasonably practicable and reasonably probable within the near future, it is for the jury (or commission) to decide whether the property's suitability for this use enhances its market value, and, if so, by how much. To that extent and in that sense, the jury decides the highest and best use issue.

Our reason for so holding is that adaptability or suitability for nonexisting uses is an inextricable factual element of market value. Consider, for example, the following hypothetical: The Government contends that the highest and best use of a farm on the outskirts of a rapidly expanding urban area is its existing use (farming). The landowner offers evidence showing that reasonably practicable and probable alternative uses in the reasonably near future are (i) a commercial airfield for small planes, (ii) quarrying for substrata granite deposits, (iii) subdivision into residential lots, or (iv) a combination of farming and residential subdivision. Were the judge to decide that any one of these alternatives, say residential subdivision, was the highest and best use — and so instruct the jury — he would in effect be deciding that residential subdivision is the most valuable use. No matter how one slices it, in order to reach the conclusion that residential subdivision is the highest and best use, the judge would have to engage in a rather detailed evaluation of the characteristics of the property and its immediate environs, the likely pattern of land use development in the area, and finally current and future market demand in the area — precisely the sort of factual analysis that enters into the determination of "fair market value." In so doing, he would be deciding more than just the "preliminary matters" which Reynolds and Rule 71A(h) tell him to decide. Instead, he would be deciding in part the very issue of just compensation which, under Reynolds and the Rule, is committed to the jury or commission. For, as Olson v. United States defines it, "just compensation includes all elements of value that inhere in the property"; it is "to be arrived at upon just consideration of all the uses for which it is suitable."

In other words, if suitability for a particular use might reasonably affect fair market value — in the sense that a just compensation award based in part upon the potential for that use would not be invalid as a matter of law — then evidence pertaining to that use must be admitted for consideration by the trier of fact. And it is for the trier of fact to decide — on the basis of all the evidence (including rebuttal evidence) and upon proper instructions — whether the property has any additional market value because of its suitability for this use. If, on the other hand, a proffered potential use is not reasonably practicable or probable, so that no reasonably minded trier of fact faithfully applying the law could find that it represents an element of fair market value, then of course the landowner is not entitled to have evidence concerning that use considered by the trier of fact. Indeed, as we have already held, under Rule 71A(h) the trial judge must exclude evidence of any such speculative uses.

Finally, it bears emphasizing that a prospective use may be speculative, and therefore subject to exclusion from consideration by the trier of fact, not just because there is insufficient evidence of either the property's adaptability or of any market demand for that use, but rather because the use simply is not permitted. So far we have ignored the legal framework of land use restrictions to which virtually all private real estate is subject. But regulatory restrictions may preclude an otherwise possible use even more decisively than the inherent physical characteristics of a property. And it is clear that just compensation must be determined in light of such regulatory restrictions. Thus, in screening proffered potential uses to ascertain whether there is sufficient evidence of their practicability and probability to warrant submission to the trier of fact, the judge must take into account any regulatory restrictions applicable to the property and the proposed use. And if the judge finds that the regulations preclude the proferred use, he must under Rule 71A(h) exclude evidence of that use from the just compensation determination.

Construing governmental regulations is of course a paradigmatic judicial function. With respect to this issue, there is no room for a division of fact-finding responsibilities between judge and jury. There are no "if's," "and's," "but's," or "maybe's" for the jury to consider. A legal restriction either applies to the proposed use or it does not — and that is a quintessential question of law for the judge to decide.

However, it is not at all uncommon within regulatory systems for permits or variances to be granted, or for the regulations themselves (especially zoning regulations) to be changed. And since the prospect of obtaining a permit or a change of zoning classification is a factor that might well be considered by a prospective purchaser, and thus a factor affecting the price a willing buyer would pay for the property, it will often represent an element of fair market value. Accordingly, it is well settled law that if the landowner can demonstrate a "reasonable possibility" that a permit would be issued or that rezoning will occur, thereby freeing the property for a use which otherwise would be precluded by regulatory restrictions, the owner is entitled to have that "reasonable possibility" considered by the jury, provided of course that the use is otherwise a practicable and reasonably probable one. See, e.g., H R Corp. v. District of Columbia, D.C. Cir., 1965, 122 U.S.App.D.C. 43, 45, 351 F.2d 740, 742; Wolff v. Commonwealth, 1 Cir., 1965, 341 F.2d 945, 946-47; United States v. Meadow Brook Club, 2 Cir., 1958, 259 F.2d 41, 45, cert. denied, 1958, 358 U.S. 921, 79 S.Ct. 290, 3 L.Ed.2d 239. See generally 4 Nichols, The Law of Eminent Domain § 12.322[1] (rev'd 3d ed.). We emphasize, however, that it is for the judge to determine whether this "reasonable possibility" exists. If and only if he finds that it does exist, then the jury must consider and decide what effect this reasonable possibility of a permit or zoning change has upon the fair market value of the condemned property.

In sum, we hold that it is the trial court's responsibility under Rule 71A(h) to screen all proffered potential use evidence and exclude from consideration by the trier of fact evidence of any potential uses upon which a just compensation award may not, as a matter of law, be based. This of course encompasses alleged potential uses which have not been demonstrated to be practicable and reasonably probable even absent regulatory restrictions as well as potential uses that are proscribed by legal restrictions from which there is no reasonable possibility of relief. If, however, the party offering the potential use evidence (almost invariably, the landowner) is able to meet this threshold burden, then evidence pertaining to that use must be submitted to the trier of fact under whatever instructions are appropriate, and the weight of that evidence as regards the fair market value of the property is for the trier of fact to decide.

With respect then to the 52 tracts in this case, the landowners on remand bear the burden of producing credible evidence that as of the time of taking, stilt houses, boat docks, elevated boardwalks, etc., were practicable uses and that there was a reasonable likelihood that their properties would be so used in the reasonably near future. United States v. Cooper, supra (p. 814, supra). And if the Government demonstrates that any of these uses are subject to local zoning restrictions or to the regulatory authority of other local or state agencies or federal agencies such as the U.S. Army Corps of Engineers, the landowners bear the additional burden of showing that there was, at the time of taking, a reasonable possibility that the use nonetheless would have been permitted. On the basis of the little evidence in the record before us, it appears likely that some of the owners will be unable to meet this threshold burden of proof, that for some of the properties these claimed potential uses are indeed too speculative and remote to serve as a basis for awarding compensation. On the other hand, some of the owners may well be able to produce "juryworthy" (or "commission worthy") evidence that their properties were, even given the existence of various regulatory restrictions, reasonably suitable for at least some of these uses. But all this is a matter for the trial court to decide in the first instance upon a much more complete evidentiary presentation than is currently available to us. And if an owner's evidence is sufficient to survive the preliminary judicial screening, it then falls to the trier of fact to decide upon all the evidence and under proper instructions whether, and by how much, his property is enhanced because of its suitability for any of these uses.

VI. ILLEGAL EXISTING USES

Of course, it would be impossible to hold that permanent structures such as cabins were not feasible, practicable uses for any of the properties. For 3 of the 52 tracts that remain before us on appeal, along with one parcel whose owner did not appeal, had already been improved with cabins at the time of taking. Yet at trial, the juries were not permitted to receive evidence of these existing improvements, and the four improved properties were valued along with all the other tracts as undeveloped, suitable only for hunting, fishing, and perhaps camping.

The reason given by Judge Mehrtens for excluding any evidence of these existing structures was that they were illegal uses. The record does not reveal which laws and regulations rendered them illegal, but we presume that they were the same zoning, sewage disposal, and environmental regulations (also never precisely identified) relied upon to exclude testimony that permanent structures were not potential uses for which the properties were suitable. Later, Judge Mehrtens explained that before the owners could introduce evidence of existing structures, they had to prove that the structures had been legally erected under a proper permit.

Our review of this illegal use ruling is severely handicapped by the wholly inadequate state of the record. Most of the evidence proffered at trial pertained to the improved parcel which has not been appealed, and even that evidence is sketchy: in 1958 the owner sunk pilings and built a frame cabin with a metal roof, insulated ceiling, and attached porch on a spot on his property above the high tide mark; at the time the cabin was built no construction permits were required and subsequently the owner was never questioned about a permit or the legality of the cabin; the National Park Service had been aware of the cabin for years and had never objected to its existence; during the 18 years after it was built, the owner used the cabin at least once a month. Aside from this and the bare fact that there were cabins on three other properties, the record contains virtually no other evidence pertaining to the existing structures and their supposed illegality. As already mentioned, there is no evidence of the laws, ordinances, or regulations under which these cabins were purportedly illegal uses. A fortiori, there is no evidence of when these laws and regulations went into effect. Nor does the record contain any indication of when the other 3 cabins were constructed. Finally, there is no evidence of how the unidentified regulatory authorities would treat nonconforming uses or whether they would grant after-the-fact permits.

Given that the current record is almost devoid of relevant evidence, we can provide only limited guidance as to how the three remaining improved properties should be valued upon remand. If, contrary to the Government's as-of-now unsupported contention, the cabins on these properties are perfectly legal and unobjectionable uses, the owners certainly are entitled to full compensation for them. Assuming that there was a market in 1976 (the date of taking) for similarly improved recreational properties, then the familiar "market value" approach, using comparable sales of similar properties, should be used to determine the compensation due these owners. If for some reason it is impossible or would be unfair to use a "market value" approach, the court will have to resort to whatever alternative valuation approach (for example, reproduction cost less depreciation) is fairest. See generally Orgel, Valuation Under the Law of Eminent Domain §§ 150-199 (2d ed. 1953).

If, however, the cabins were, at the time of taking, illegal uses, the issue of just compensation is much more complex than the rulings below would suggest. As a preliminary matter, it is the Government's burden to prove that an existing structure is unlawful and therefore should not, as a matter of public policy, be compensated. Only if the Government proves that a permit or variance was required at the time the structure was erected does it become incumbent upon the landowner to prove that the structure was built legally with a valid permit. But even failing this burden of rebuttal, it is still open to the landowner to prove that the relevant regulatory authority might nonetheless grant an after-the-fact permit or otherwise allow the structure to remain. Furthermore, compensation for an existing and otherwise valuable structure can be completely denied only if it was an illegal use at the time it was built. If it is only by virtue of supervening laws and regulations that the structure has become an illegal use, the owner does not forfeit his constitutional right to be justly compensated for his property. In this situation the appropriate compensation calculation may be difficult and somewhat unorthodox. It may be necessary to refer to whatever provision, if any, the supervening law makes for nonconforming uses, and it may be necessary to use some standard of valuation other than "market value," such as reproduction cost less depreciation. But the difficulties involved in ascertaining " just compensation" in these circumstances do not warrant denying compensation altogether. See generally Orgel, supra, §§ 32-34; Note, Determination of Just Compensation in Eminent Domain Proceedings for Land Subjected to Illegal Uses or Conditions, 38 Notre Dame Law. 196 (1963).

Of course, the compensation due the owners of the three remaining improved properties cannot properly be considered until all the relevant facts have been ascertained. On remand, the trial court should obviously concern itself with ensuring that the compensation awarded the owners of these three properties is "just" both to them and the public, but the trial court must also do that which it alone is able to do — develop a competent and complete factual record, so that at the very least all its other efforts do not go for naught.

VII. MISCELLANEOUS TAG-ENDS

In addition to all the above, the landowners have assigned error to a number of evidentiary rulings that occurred during the three trials and to the charges given the jury at the end of each trial. We see no need to address ourselves to the asserted deficiencies in the jury instructions or to several of the allegedly erroneous evidentiary rulings that are unlikely to recur on remand. However, one of the evidentiary disputes would likely be repeated on remand absent our review, and another of the appealed rulings was so manifestly erroneous that it warrants express disapproval.

A. Admissibility of Just Compensation Statements Sent Landowners Prior to Instituting Proceedings in Eminent Domain

One of the admissibility questions that may recur on remand concerns the statements of just compensation provided various landowners by the Department of the Interior in compliance with the Uniform Relocation Assistance and Real Property Acquisition Act of 1970 (the Act). Section 301 of the Act ( 42 U.S.C. § 4651) establishes certain uniform nationwide "guidelines" for the public taking of any lands by agencies of the federal government. Among these guidelines or policies is the exhortation that, prior to instituting proceedings in eminent domain or to entering into any negotiations with the owner of land sought for a federal project, a federal agency should appraise the property, establish an amount which it believes to represent just compensation, and then offer to acquire the property for no less than this appraised fair market value. In addition, the statute states that the owners of the property should be provided with "a written statement of, and summary of the basis for, the amount * * established as just compensation." See 42 U.S.C. § 4651(3).

Pursuant to § 4651, the Department of the Interior sent landowners Grantt and Ciccone statements that its "best estimate of `just compensation'" was $20,000 for the Grantt property and $80,000 for the Ciccone property. Early in Trial I, Earle unsuccessfully attempted to introduce into evidence, as an admission against interest, the $20,000 appraisal provided the Grantts. Later, when the Government presented its case, its expert appraiser testified that the fair market value of the Grantt property was $12,000 and that of the Ciccone property was $64,000. In order to impeach this testimony, Earle then sought — again without success — to introduce the higher figures contained in the statements of just compensation sent to the Grantts and Ciccones.

On appeal, the landowners renew their argument that these and like statements of just compensation are admissible, either as admissions or as impeachment by a prior inconsistent statement. The Government counters with an array of arguments in support of the District Court's ruling excluding the statements. First, it contends that the statements were properly excluded under F.R.Evid. 408 as offers of compromise made during settlement negotiations. Second, it points to the fact that the § 4651 guidelines and policies, including those concerning statements of just compensation, are not enforceable requirements. Finally, it relies upon a decision of the Fourth Circuit, Washington Metropolitan Area Transit Auth. v. One Parcel of Land, 4 Cir., 548 F.2d 1130, holding that a landowner who rejects a pre-condemnation offer made pursuant to § 4651 may not introduce that offer as proof of value in the subsequent compensation trial involving that property.

We do not find any of these arguments sufficiently persuasive to justify excluding § 4651 statements of just compensation which, if sent to the landowner, presumably represent the appraised fair market value of the condemned property — precisely the question at issue in a compensation trial. We fully recognize that under 42 U.S.C. § 4602(a) the relevant federal agency cannot be required to provide such statements. But § 4602(a) does not speak to the evidentiary competence of any statements that are prepared and sent to prospective condemnees.

Nor is F.R.Evid. 408 applicable. The objective of the § 4651 procedure — including the provision of statements of just compensation and any ensuing negotiations — is the acquisition by purchase of privately-owned lands short of condemnation. Technically, at the time the statements are provided, there is no disputed claim, and hence no settlement negotiations of a disputed claim, to serve as the predicate of Rule 408. But even if Rule 408 is viewed to apply to the negotiations envisioned by § 4651, there is no reason to extend it to the statements of just compensation. For, under § 4651, they are simply "statements" of the "amount * * * established as just compensation" after an approved appraisal of the fair market value of the property; they are not "offers," which under § 4651 may well be for more than the appraised fair market value. In other words, they are statements of the amount which the Government believes the landowner is constitutionally entitled to should negotiations fail and condemnation proceedings be initiated. As such, the policy reasons for excluding offers of compromise, as identified by the Advisory Committee which drafted Rule 408, are not applicable to these statements. First, unlike offers of compromise, there is no possibility that these statements are motivated by a desire to buy peace, irrespective of the merits. To repeat, the statements represent the bottom line — the sum which the Government believes the landowner is constitutionally entitled to. Second, there is no significant danger that admission of such statements will deter the negotiated acquisition of private lands. Insofar as the statements do represent the Government's appraised estimate of fair market value and just compensation, they represent what the Government believes it will have to pay in any event after incurring the costs of a condemnation proceeding. Governmental desire to avoid these costs should be sufficient incentive to attempt a negotiated and voluntary sale of the property, regardless of the admissibility of the statements. Finally, we point out that among the express purposes with which Congress enacted § 4651 were "to assure consistent treatment of owners" and "to promote public confidence in Federal land acquisition practices." Against this background, we do not believe that Congress even entertained the notion that the statements of just compensation would be anything other than what they profess to be — a good faith assessment of what the owner is constitutionally due, made by one authorized by this statute to make that every statement. As the law of admissions is really the law of agency, not evidence, one can hardly imagine a statement more fitting the definition than this — a statement made by one acting within his authority asserting a position contrary to the one asserted at trial. See Brown Root, Inc. v. American Home Assurance Co., 5 Cir., 1965, 353 F.2d 113, 116, cert. denied, 1966, 384 U.S. 943, 86 S.Ct. 1465, 16 L.Ed.2d 541; Compagnie De Navigation, etc. v. Mondial United Corp., 5 Cir., 1963, 316 F.2d 163, 171 n. 10.

For these reasons we hold that if § 4651 statements of just compensation are provided a prospective condemnee, they are admissible at a subsequent compensation trial as an admission, once it becomes known that at trial the Government is valuing the property at a lower figure. Of course, as generally true of all admissions, the § 4651 statement is not binding, and the Government is free to explain why it now believes its earlier appraisal to be inaccurate. But the Government is not completely free to play fast and loose with landowners — telling them one thing in the office and something else in the courtroom.

B. Exclusion of Landowners' Demonstrative Evidence

In each of the trials, the Government depicted the landowners' properties as muck and swamp, perpetually inundated by water. To support his description of the properties, the Government appraiser in Trial II introduced and published to the jury a series of large, color photographs taken in the general vicinity of some of the parcels. No land at all could be discerned in the photographs, but only dense stands of mangrove trees, their roots covered by brackish water.

On rebuttal, Earle tendered a movie made by one of the landowners, which showed an area of his property about 1600 square feet in size which had been cleared, was high and dry, and was being used by his family for camping. Earlier this owner had testified that he had been out to his property about 50 or 60 times, that he had never seen it under water, and that he had cleared a campsite on it which he used as a base for fishing expeditions in the area. In response to Earle's motion to show the movie to the jury, the Government volunteered to stipulate that this owner had cleared a portion of his property upon which he camped with his family. The Court then ruled that "[i]n view of the stipulation," the owner's movie was "cumulative" and the Court therefore excluded the movie from evidence.

This was a clear abuse of discretion. Government witnesses had testified that the landowners' properties were, for the most part, swamp, muck, and water. The Government was allowed to introduce pictorial evidence of lands in the area that tended to corroborate this testimony. Many of the landowners, on the other hand, had testified that at least portions of their properties were high and dry. But without any corroborating pictorial evidence, and in the face of the Government's photographs, the jury might well have dismissed their testimony as self-serving lies. To exclude the owners' only demonstrative evidence that their properties were high and dry and useable was in these circumstances tantamount to instructing the jury that the lands were muck and swamp. The error is obvious, as is its prejudicial effect.

VIII. REMAND FOR TRIAL BY COMMISSION

Although this concludes, at least for the time being, our trek through the labyrinthine ways of the Everglades and just compensation law, it does not mark the end for the Government and the owners of the 52 tracts presently before us. Determination of the just compensation constitutionally due these owners must begin anew in the District Court. This will be no easy task.

Much of the burden will of course fall to the Court. For example, with respect to the scope of the project rule, it will be the Court's responsibility to decide the proper and just date as of which any alterations in value attributable to the Northwest Extension must be disregarded in setting the compensation award. And, with respect to the issue of highest and best use, the Court will have to decide whether the landowners can satisfy the threshold burden of demonstrating that cabins, boat docks, elevated boardwalks, etc., are both reasonably practicable and, in light of any regulatory restrictions found applicable, reasonably probable uses. Further, with respect to any existing structures, the Court will have to determine whether the owners must be compensated for their taking, and, if so, the appropriate standard of compensation.

These and any other preliminary factual and legal issues are the Court's responsibility under Rule 71A(h), whether trial is had before a jury or a commission. But even after the Court has narrowed the issues by deciding these preliminary matters, the trial of the ultimate issue of just compensation will remain a formidable undertaking. In the first place, there are 52 tracts involved. And even if the Government is correct in contending that these properties are pretty much homogeneous, at the very least they differ substantially in size — ranging from 0.34 acres to 320 acres. But despite the Government's position that once you've seen one coastal mangrove tract, you've seen them all, the record indicates that there are significant differences other than size alone — among them the extent to which the properties are high and dry, the amount of waterfront footage, and the location of the properties relative to easily traversed waterways such as the Wilderness Waterway and the Gulf of Mexico. As observed earlier, these differences may mean that some of the properties are suitable (or at least sufficiently suitable that the landowners may introduce evidence on this point) for permanent structures, whereas others must be valued as undeveloped and unimprovable. But even if it develops that all the properties must be valued solely on the basis of their existing use, some will almost surely be better suited for camping and boating than others, thereby precluding valuation according to some mechanical formula pegged, for example, to size alone. To complicate matters further, the fact-finder will have to apply to each of these 52 parcels what may prove to be a rather sophisticated standard of just compensation — including some value due to the Northwest Extension, while disregarding any decrease or increase attributable to the project after the still-to-be-determined commitment date. And, of course, the standard may be even more complex for any properties upon which there are permanent structures which, although no longer legal, are nonetheless compensable.

All this suggests that the issue of just compensation should be decided on remand by a commission appointed by the District Court pursuant to F.R.Civ.P. 71A(h). As pointed out earlier, the right to a jury trial in federal eminent domain proceedings is governed not by the Constitution, but by Rule 71A(h). That Rule gives the District Court the discretion to appoint a commission in lieu of a jury "because of the character, location, or quantity of property to be condemned, or for other reasons in the interest of justice." Although other courts have held that this discretion is quite broad," in this Circuit we purport to adhere to the principle that a commission is to be used only for exceptional cases. Nonetheless, we have previously recognized that among such exceptional cases are those involving "large areas held by many small landowners, or property too distant for a jury to view the premises," United States v. Buhler, supra, 254 F.2d at 880, and we have upheld the use of a commission in a project involving only sixteen parcels of land, United States v. 2,477.79 Acres of Land, 5 Cir., 1958, 259 F.2d 23.

In light of the number of parcels involved in these condemnation proceedings, the remoteness of these parcels from any federal courthouse in which a jury trial might be conducted, and the complexity of the issues, this certainly is one of those exceptional cases in which trial by a commission is not only appropriate, but is required "in the interest of justice." We therefore direct under 28 U.S.C. § 2106 that on remand the issue of just compensation is to be determined by a commission pursuant to the procedures outlined in Rule 71A(h) and the principles announced in this opinion.

The judgments appealed from are VACATED and the cases REMANDED for further proceedings consistent with this opinion.