FRIENDLY, Circuit Judge:

For the fourth time this Court has before it an order of the Interstate Commerce Commission in the Penn-Central merger — this one relating to the terms for the inclusion of The New York, New Haven and Hartford Railroad Company (NH). The principal decision under review is a Second Supplemental Report on Further Hearing, hereafter "the Report," 331 I.C.C. 643 (1967); there is also a Third Supplemental Report on Reconsideration, I.C.C. (1968). In five consolidated actions holders of various N.H. bonds and trustees under N.H. bond indentures (all collectively referred to hereafter as "the bondholders") challenge the Commission's conclusion that a Purchase Agreement between the Pennsylvania and the Central on the one hand and the Trustees of N.H. on the other dated April 21, 1966, and an interim loan arrangement set out in the Report constitute "just and reasonable" and "equitable" terms within §§ 5(2)(b) and (2)(d) of the Interstate Commerce Act. Erie Lackawanna R.R. has intervened, under the broad provisions of 28 U.S.C. § 2323, to assert an unrelated claim — that the Commission erred in failing to set conditions to protect its interchange with N.H. at Maybrook, N.Y., similar to those prescribed for the benefit of the Central Railroad of New Jersey, the Reading and the Western Maryland. With some regret in view of the public urgency of effecting inclusion of N.H. in Penn-Central, we are constrained to vacate the order and remand for expeditious proceedings and further report by the Commission.

We begin by noting the somewhat curious posture in which, so far as concerns the bondholders' actions, the case comes to us. Since 1961 N.H. has been in reorganization under § 77 of the Bankruptcy Act in the District of Connecticut under the supervision of Judge Anderson. The Commission's report approved the sale of N.H. to Penn-Central not only under § 5 of the Interstate Commerce Act but as a means for the execution of a plan of reorganization of N.H. under §§ 77(b)(5) and (d). The latter determination must be certified to the District Court for Connecticut, which may approve it only if that court finds the plan "fair and equitable," § 77(e), with review of its decision lying in the Court of Appeals. When the bondholders' actions were filed, the United States moved for dismissal on the ground that the reorganization court had exclusive jurisdiction. We denied the motion in a memorandum, the salient portions of which are reproduced in the margin. We adhere to the view there expressed; unfortunate as the duplicitous system of review may be, we see no basis on which we can properly decline to exercise the jurisdiction conferred upon us by 28 U.S.C. § 1336(a) and 2321-25. Compare General Protective Committee for Holders of Option Warrants of United Corp. v. SEC, 346 U.S. 521, 74 S.Ct. 261, 98 L.Ed. 339 (1954); Phillips v. SEC, 388 F.2d 964, 969 (2 Cir. 1968).

I. The Commission's Approach to the Valuation of NH.

In determining fair and equitable terms for the sale of NH, the Commission was faced with a task far more difficult than the one it had just performed in setting terms for the acquisition of Erie-Lackawanna, D H and B M by the Norfolk Western, see Erie-Lackawanna R.R. v. United States, supra, 279 F. Supp. at 337-340, 342-345; 389 U.S. at 523-526, 88 S.Ct. 602. Those three roads had, or were claimed to have, earning power. Here, because of the large and increasing losses of NH, the Commission properly found, 658, that "Past experience * * * furnishes no basis for valuing N.H. from an earnings standpoint except to emphasize the chronic deficit character of the operation which is likely to persist and render the operation a burden rather than an income producer." A fair price for N.H. on the usual basis of capitalization of earnings would thus be negative or, at best, zero; on the other hand N.H. is conceded to have a very substantial liquidation value.

The Commission discussed the valuation problem at several places. It began by noting, 656-58, that in the negotiations leading to the agreement between the acquiring railroads and the N.H. trustees "asset value was the primary determinant rather than earning power," and that the evidence before it "was directed largely toward showing asset values upon an assumed liquidation." After saying that "This approach had the appearance of satisfying" the provision in § 77(b)(5) that the means for execution of a plan or reorganization might include "the sale of all or any part of the property of the debtor either subject to or free from lien at not less than a fair upset price," it continued that "We do not look upon liquidation value as necessarily being the equivalent of `a fair upset price'" since "liquidation contemplates a dismantling of the railroad and, therefore, an abandonment of the railroad's operations — eventualities hardly likely, in view of our findings in the prior reports herein that the services of N.H. are essential." It thought that "a fair upset price should be geared to the requirement that very large segments of the railroad, though probably not all, would have to be maintained in operation" through purchase by public agencies. The Report states that in such event "the pricing would be lower than a complete liquidation could produce" — without, however, elucidating on what basis N.H. could be required to sell property to public authorities for less than liquidation value. The Commission concluded this portion of the discussion by saying that "if the consideration to be paid is the equivalent of the liquidation value, it will exceed the fair upset price contemplated by section 77(b)(5)."

After extensive discussion and findings as to the liquidation value of NH, the value to N.H. of the consideration under the Purchase Agreement, and the cost of the acquisition to Penn-Central, the Commission stated its summary and conclusions as follows, 697-98:

As detailed earlier, we have found the liquidation value of N.H. to be $125 million. We have also found that the value to N.H. of the consideration to be received is $125 million. It is evident, therefore, that the consideration is at least equivalent to the liquidation value and [it] will, therefore, be equitable to all parties having an interest in the N.H. estate to approve the transaction.

From the point of view of Penn-Central, we have found that the total cost to them of acquiring the assets is $159 million. In these circumstances, it would be unfair to Penn-Central to require them to pay a price greater than that agreed upon in the purchase agreement.

Penn-Central must assume a loss-operation which not merely lacks promise of foreseeable recovery, but in fact is plagued by an unrelenting erosion. Thus, they are faced with constantly diminishing going concern values. The cold, hard fact is that N.H. has been losing money for years and that unfortunate situation continues unarrested. As thorough as may be the analyses forecasting profitable operations for N.H. as a freight-only railroad, they are subject to the fulfillment of a number of conditions and the achievement of proposals as yet untested by experience. We cannot give such projections as much weight as the actual operating results of many years.

The Commission's final words on the subject came in a portion of the opinion dealing with the Trustees' application under § 77. Here the Commission said, 724-25:

Although earning capacity is usually looked upon as the principal criterion in valuing railroad properties, due consideration may be accorded to other relevant factors, including original cost of the properties, cost of reproduction new, cost of reproduction depreciated, actual investment therein, severance or strategic value, contributed traffic, value of particular facilities, and the effect of potential operating economies. We have given due consideration to past, present and prospective earning power as well as to other relevant factors specified in section 77(e), including reproduction and original costs, and, additionally we have considered the liquidation value of N.H. — finding that value to be $125 million, a figure comparable to the price in the purchase agreement. Moreover, the liquidation value of the N.H. properties in the context of the NH's present situation exceeds the fair upset price, and as indicated, under section 77(b)(5) all or any part of a debtor's property may be sold at not less than a fair upset price.

Those who have a valuable interest in the N.H. estate could not, under circumstances favorable to them in a liquidation, expect to effect a greater net return then would be accorded them under the plan we are approving herein for the sale of the assets to Penn-Central. All have had fair and ample opportunity to present evidence and argument on the question of valuation, and we find the record entirely adequate to make possible and to support findings as to that issue for all purposes of section 77 of the Bankruptcy Act as well as section 5(2) of the Interstate Commerce Act. Accordingly, the sale of N.H. assets to Penn-Central under the terms of our decision herein is consistent with section 77(b)(5) in that it provides an adequate means by which the trustees' plan can be executed. We also find that the consideration to be paid for the assets of N.H. constitutes just and adequate compensation. As we view the situation, the parties holding a present interest in N.H. will receive the economic equivalent of what is surrendered upon the sale of the N.H. assets.

The impression we derive from all this is that the Commission understandably avoided the hard decision whether a deficit ridden railroad could fairly be sold for less than its very substantial liquidation value by finding that under the Purchase Agreement liquidation value was received and thus the transaction was not unfair to NH. Since the validity of the latter inference would be indisputable, we shall first examine whether substantial evidence supported the Commission's finding of equivalence. If it did, we need go no further; if it did not, we must consider whether the Report can nevertheless be sustained.

II. The Liquidation Value of NH.

As indicated, the Commission made findings as to (1) the liquidation value of N.H. on December 31, 1966, (2) the value of the consideration payable by Penn Central, and (3) the cost of the acquisition to Penn Central. It found the first to be $128.9 million which it "rounded off" to around $125 million; the second to be $123.6 million which it again "rounded off" to around $125 million; and the third to around $159 million. In analyzing these findings, we shall endeavor to be ever mindful of the limited scope of our review, see Erie-Lackawanna R.R. v. United States, 279 F. Supp. at 337, and cases there cited, and of Mr. Justice Cardozo's warning, "An intelligent estimate of probable future values, * * * and even indeed of present ones, is at best an approximation. There is left in every case a reasonable margin of fluctuation and uncertainty." Dayton Power Light Co. v. Public Utility Comm'n, 292 U.S. 290, 310, 54 S.Ct. 647, 657, 78 L.Ed. 1267 (1934).

While we agree with the Government that in many instances the bondholders seek to have us substitute our judgment for the Commission's and display a disturbing tendency to add apples to cucumbers by conceiving N.H. as operating and not operating at one and the same time, we nevertheless find the Commission's figures for liquidation value significantly too low and for value of consideration significantly too high. In this section of the opinion we shall discuss the former.

(1) Failure to discount expenses of liquidation.

We begin with what appears to be an obvious error by no means unsubstantial in amount. The Commission found that liquidation of N.H. would require time, which it assumed to be six years. It projected the receipts over this period and discounted them to present value, applying a factor of 6%. Recognizing that so large a program would entail substantial expenses including continued accrual of property taxes, it estimated these for each of the six years — a total of $58.23 million. However, it did not discount the expenses to present value. If future receipts are to be discounted to present value, as they properly should, future expenses must likewise be. The obligation to pay $1,000,000 six years hence no more entails a present cost of $1,000,000 than a right to receive $1,000,000 six years hence is an asset in that amount.

The Report makes no attempt to explain its inconsistent treatment of the credit and debit sides of the ledger. The Commission's defense on brief is wholly unsatisfactory. It suggests that if a discount is to be applied to the liquidation expenses, a partially compensating downward adjustment should be made in the properties' assessed market value to reflect the fact that deferring tax payments by selling properties subject to tax liens depresses the sales price below the face value of the lien. The Trustees could, however, avoid this result simply by discharging the tax liens shortly before sale — under the contemplated liquidation schedule sufficient funds would be available for these and all other expenses. While it is difficult for us to determine the proper amount of the discount on the present record, it seems likely that, as the bondholders claim, this would be between $5 and $6 million.

(2) NH's interest in Grand Central Terminal (GCT) properties.

By far the largest item in dispute is the value of NH's interest in income derived from hotel and office buildings constructed on property acquired by the New York Central in connection with the construction of Grand Central Terminal (GCT) in New York City and the tracks leading to it. Penn Central contends that, except for repayment of construction advances on the Waldorf Astoria Hotel, $1,110,258 as of December 31, 1966, NH's interest is simply to have the net proceeds of these properties devoted to the payment of terminal expenses; hence when N.H. ceases to operate its railroad, as it would in liquidation, its interest will end. The bondholders claim N.H. is entitled to a perpetual half interest in the net proceeds of the properties undiminished by terminal expenses, appraised at $113.5 million, less some $2 million representing the amount, capitalized on an 8% basis, of $160,179.22, which on their view of the Agreements, the N.H. is obliged to pay regardless of its use of the Terminal or a net figure of $111.5 million. The Commission valued NH's settlement prospects at $13 million by a process hereafter described. Since the facts with respect to the contracts and relations concerning the GCT properties are set forth in a painstaking report rendered to the District Court for Connecticut on June 18, 1968, by Hon. John Van Voorhis, retired Judge of the New York Court of Appeals, acting as special master, we can condense the recital here. While we have read Judge Van Voorhis' report, we have not relied upon it; our conclusions are our own.

The NH's passenger tracks from the east run only as far as Woodlawn Junction in the Bronx, where they join those of the New York and Harlem Railroad (Harlem), since 1873 a leased line of the New York Central. N.H. in 1848 acquired trackage rights southward over the Harlem to the latter's terminal, then located at Fourth Avenue and 26th Street; in 1872 the Harlem granted N.H. and the Central use of the newly built Grand Central Depot. Legislation of 1903 and 1904 required that the tracks immediately north of 42nd Street be depressed below street level; this necessitated demolition of the old depot and construction of the present Terminal.

By agreement of 1907 the Central leased to N.H. "the use" of the Terminal, not to exceed 50%, in common with the Central. The Terminal was defined as including land and improvements described in annexed schedules, these encompassing certain properties on or in the vicinity of Park Avenue between 42nd and 57th Streets. The Central was bound to acquire all necessary land and to construct the Terminal at its own expense. N.H. was to pay a proportion, based on car use, of four items: (i) $485,393.69 interest on the cost of construction of the former depot; (ii) interest, at the rate of 4 1/4% on the cost of the new Terminal; (iii) annual payments to the City of New York, insurance premiums, taxes and assessments relating to the new Terminal; and (iv) the cost of maintaining and operating it. Section 14 of the lease provided that amounts owing from tenants and others should be collected by the Central, and that "all such rentals or other compensation so received, and not belonging exclusively to either Company, shall be credited to the fixed charges or to the cost of maintenance and operation of the said Railroad Terminal, as the same may be applicable."

Shortly before 1913 it became apparent that large additional revenues could be generated by constructing buildings on land that had been acquired for the Terminal. Desiring "to express more fully the intent of the parties" as to the rights of N.H. and the Central "with respect to the construction, maintenance and use of such buildings and with respect to the rentals accruing from the leasing of such buildings or portions thereof," the two companies entered into a supplemental agreement in 1913. This added a long paragraph to the section of the 1907 agreement which had originally leased the "use" of the Terminal properties only for the express purpose of accommodating the passenger service of the NH. This new paragraph made it clear that the N.H. also had the right to participate in real estate development, stating that it was

"agreed that the use of the Railroad Terminal demised by this paragraph to the New Haven Company shall include the right on the part of the New Haven Company to join with the Central Company, in accordance with agreements relating thereto, made, or from time to time to be made, by and between the Central Company and the New Haven Company, in the construction, holding, maintenance and leasing of buildings * * * upon the land included within the Railroad Terminal."

The 1913 Agreement further provided that, as to each building:

"toward the construction of which the New Haven Company shall furnish or contribute money, and until it is reimbursed therefor, the New Haven Company shall have and own an undivided interest which shall bear the same proportion to the entire building as the amount of money furnished or contributed by the New Haven Company, and for which it has not been reimbursed, bears from time to time to the entire cost for the construction of such building. At the time when the New Haven Company shall be fully reimbursed for all moneys furnished or contributed by it for or toward the construction of any such building, with interest at such rate as shall have been agreed upon, such undivided interest of the New Haven Company shall cease."

Rentals from buildings in which the New Haven participated were to be applied as follows:

"In all cases in which the New Haven Company and the Central Company shall join in the construction of any such building or buildings, the rentals to be credited to the fixed charges or to the cost of maintenance and operation of the Railroad Terminal under the provisions of Paragraph 14 of this agreement shall include all rentals accruing from the leases of the buildings or parts of buildings so constructed, except such rentals as by agreement between the New Haven Company and the Central Company shall be paid either of them as a proper return for the moneys advanced by such party to provide for the construction of any said buildings."

As to many of the buildings constructed N.H. availed itself of its right to join with the Central, either providing half of the cost or becoming jointly liable to outside institutions supplying the financing; as to three it did not. The rentals from all the buildings after deducting the cost of operation were applied against the expenses of the Terminal except that where N.H. had not participated only the ground rents were so applied; not until 1964 did these rentals produce a surplus. The N.H. trustees have sued in the New York courts to recover half these amounts but the actions have not proceeded beyond the pleading stage. Litigation between 1960 and 1962 resulted in decisions by the New York courts that, as to the Biltmore Hotel, a property toward whose construction N.H. had contributed 50%, N.H. was entitled to an equal voice with respect to the making of any lease, although the advances had all been repaid. New York Central R.R. v. New York, N.H. H.R.R., 24 Misc.2d 414, 208 N.Y.S.2d 605 (Sup.Ct.N.Y.Co. 1960), aff'd, 13 A.D.2d 309, 216 N.Y.S.2d 928 (1st Dept. 1961), aff'd, 11 N.Y.2d 1077, 230 N.Y.S.2d 226, 184 N.E.2d 194 (1962). The New York courts did not have occasion to determine whether N.H. was also entitled to an equal share in surplus rentals during its operation into GCT, much less whether any such right would cease on discontinuance. The Biltmore litigation had been conducted for N.H. by Sullivan Cromwell; in 1964 that firm rendered an opinion to the N.H. Trustees "that discontinuance of entrances into the Terminal by the New Haven would terminate the 1907 Agreement and the 1913 Agreement as well, except as to the right of the New Haven to be reimbursed for any unpaid building advances."

After reviewing this history the Commission, rejecting the polar positions of Penn Central and of the bondholders, said it would "only decide here that the claims of N.H. are not frivolous but claims of substance, and in the absence of litigating them they must be deemed to be valuable," 684. It found that, on a relative car-use basis of 40%, NH's share of surplus income would have been $1,842,401 in 1964 and $1,473,627 in 1965 — an average of $1,658,014. "Capitalized at 8 percent a year, which we believe to be a realistic rate, this annual average income would have a value of approximately $20 million." Noting Penn Central's concession that NH's claims had a "nuisance value" of $5 million, the I.C.C. apparently averaged these figures for a total of $13 million.

In our view NH's claims to share in the total net income unreduced by terminal expenses and to share in the surplus income differ toto coelo. Since the possibility of NH's ceasing operations into the Terminal was scarcely envisioned in 1913, the task of a court now required to determine NH's rights in that event is not so much to determine what the parties thought but rather to guess, with such small aid as their language gives, what they would have thought if they had.

It is scarcely conceivable that the Central would have agreed that NH, by picking up an option to share in the cost of building hotels or office buildings being constructed for the primary purpose of reducing the heavy expense of operating the Terminal, could quit at any time, take its share of the rentals and leave the Central burdened with the total expense of operating the Terminal, which would not be reduced anything like commensurately by NH's cessation of operations. Whereas only the clearest language would support such a construction, § 14 looks quite the other way. The bondholders argue against this that the Central would receive a windfall if it could apply the rentals from the properties to the expense of operating the Terminal, yet charge a successor of N.H. its pro rata share of such expense for the privilege of entrance. One answer is that there is no assurance N.H. will have a successor. Even if the State of New York should acquire NH's tracks from the Connecticut border to Woodlawn in order to maintain commuter service, which is wholly uncertain, it might well elect to operate only to a connection with the New York City subway system rather than bear the heavy expense of an elaborate terminal that was constructed to serve trunk as well as commuter traffic. At least this would be so unless N.H. were able to pass along its right to enter GCT on what is now an expense-free basis despite the agreement's anti-assignment clause. While that is a possibility, there is no proof that this privilege would so increase the profitability of NH's operations in New York State as to make them more valuable alive than dead. The windfall argument thus comes down to the point that abandonment of N.H. operations into GCT would reduce the terminal expenses in some amount, not shown of record, and thereby increase the surplus income. But here the prospect that the State of New York could acquire NH's right to have the income from the properties applied to terminal expenses, without however raising the price of NH's assets in the state above their liquidation value, works against NH. In short, while Penn Central presumably would pay something to be rid of what we deem an extravagant claim to half the total net rentals, "nuisance value" is not an inapposite expression.

We conclude by the same token that NH's claims to share in surplus income over and above all terminal expenses are exceedingly strong. Just as we cannot imagine the Central's agreeing that the complete dedication of the rentals to payment of the terminal expenses should end on NH's ceasing to operate, we likewise cannot conceive NH's agreeing — or even Central's demanding — that in what would have been deemed a wholly unlikely event the equally unlikely surplus income should fall entirely into Central's hands. Here also, only the strongest language would lead to such a construction but the only words that would support it — a provision that on reimbursement of advances "such undivided interest of the New Haven Company shall cease" — have been read by the New York courts in the Biltmore litigation as subordinate to the language of the granting clause conveying the right to use the land in common with the Central. As to this claim relating to surplus income it is the Central's position that has merely "nuisance value." Hence an allowance of only 65% of what the Commission considered the value of NH's share in the surplus income is "so unreasonable as to require judicial disapprovel." Old Colony Bondholders v. New York, N.H. H.R.R., 161 F.2d 413, 422 (2 Cir.), cert. denied, 331 U.S. 850, 67 S.Ct. 1754, 91 L.Ed. 1865 (1947).

While we are unable to say on this record what the correct figure as to NH's share in the surplus income is, we can indicate our views on certain elements. In the absence of better evidence the 1964-65 experience does not appear unreasonable as a base period unless — which the present record does not tell us although the press does — further projects yielding substantial income are fairly in view. However, Penn Central argues that the Commission erroneously failed to reduce the "surplus" income to reflect deductible expenditures in those years for hotel equipment and furnishings. It would seem that while deduction of these non-recurring items is appropriate in determining NH's share of 1964-65 income, they should not be deducted in calculating future surplus income unless similar expenses are expectable. In order to entitle N.H. to the benefits of the agreements, it would have to make the minimum annual payment of $160,000. On the other side we see no basis for determining the NH's share of the surplus income on the ground basis of car-use; when N.H. participated in building construction, it did so as an equal and should share on that basis. Finally we decline to fault the Commission for its use of an 8% capitalization rate rather than the 5% rate used by the Trustees' expert in estimating the total value of the office buildings. The surplus income is a marginal item, subject to fluctuation with economic vicissitudes and increased expenses of terminal operation, and a more conservative capitalization rate could thus be deemed in order, especially given the more volatile character of the rentals derived from the hotel properties which the same expert capitalized at 10%. The selection of a capitalization rate, within reasonable limits, is a subject for the expert knowledge of the Commission. While it is argued that the figure of the value of NH's share in surplus income thus determined is subject to discount for the expenses and uncertainties of litigation, Penn Central would likewise have expenses, and also the expense and risk of defending against NH's larger claim. In our view the various items of expense and "nuisance value" wash, leaving the estimated value of NH's claim to half the surplus income not properly subject to reduction. To this, moreover, there should be added a further sum to reflect NH's interest in surplus income accrued in 1964, 1965 and 1966. All told it would be fair to assume that these readjustments will increase the liquidation value of NH's interest in the GCT properties to at least $25 million.

(3) Harlem River and Oak Point Freight Yards.

The Commission had before it four appraisals of the liquidation value of the Harlem River and Oak Point Freight Yards, a tract of some 160 acres in the southeast Bronx not far from Randall's Island Park. One of the bondholders submitted an appraisal by Mr. Paul O'Keefe, a highly qualified real estate expert, that the best use of the property would be residential and that as such it had value of $32 million. He made another appraisal of $26 million for industrial use. The N.H. Trustees submitted two reports of Mr. Arthur McCann, valuing the property for industrial use at $22.65 and $18.1 million. Penn Central submitted an appraisal of $15.585 million, also on the basis of industrial use. The Commission took the $18.1 million figure. 668.

Although the bondholders make impressive arguments in support of Mr. O'Keefe's higher figure and we hardly would have faulted the Commission for accepting it, we cannot say there was not substantial evidence for rejecting the appraisal as "at best speculative because of the requirement of a zoning change" since in any event the change "could not be accomplished for some time to come because of outstanding long-term leases held by some of the tenants occupying the premises." Mr. O'Keefe recognized that because of the leases "no assumption * * * can be made that the property can be obtained for immediate residential development." More important, he stated repeatedly that before a zoning change could be approved, an owner must present the city with a development plan for the property. Since he also testified that such a plan could be developed for the yards only if some governmental unit agreed to subsidize housing construction, the Commission was entitled to consider the possibility too remote to constitute a sound measure of valuation.

On the other hand, we find no adequate explanation in the Report for the rejection of the higher of the two estimates submitted by Mr. McCann, for the New Haven Trustees, or, for that matter, Mr. O'Keefe's industrial use appraisal of $26 million, which in its main discussion of the yards the Commission does not mention. At another point in the opinion, 685, the agency indicated its belief that the earlier and higher McCann study was based upon the assumption that the properties were to be continued to be used as railroad yards and if this were correct, the Commission would have been justified in rejecting it as a guide to the yards' liquidation value. However, we are not at all clear that this understanding was right. Though the summary page of McCann's second and lower valuation study contains a parenthetical statement that the earlier appraisal was made on the basis that the property would be used "as a freight yard of an operating railroad," an examination of the earlier study itself indicates it was not in fact made on that basis but instead assessed the property's "Fair and Reasonable Market Value" to industrial users. The difference in the results of the two studies seems rather to have been due to the second appraisal's having reduced the properties' value by two factors — some $1.825 million in liquidation expenses, and a discount giving effect to the fact that it would take from one to three years to dispose of them. While these are proper enough, the record does not permit us to determine whether they duplicated reductions already made in the Commission's overall calculations. Our remand will enable the Commission to clarify this.

(4) Boston Providence.

A reorganization plan, whereby N.H. will become the owner of all the B P's assets, has been approved by the Commission the courts, In re Boston Providence R.R. Corp., 260 F. Supp. 415 (D.Mass. 1966), cert. denied, 389 U.S. 974, 88 S.Ct. 475, 19 L.Ed.2d 467 (1967), and the stockholders and creditors, and a hearing by the District Court for Massachusetts with regard to confirmation of the plan has recently been held. Since the appraised net liquidation value of the B P's physical assets was set by the plan at $8.1 million, the bondholders claim that NH's interest should be valued at that figure. The Commission, however, accepted the Trustees' valuation of $4.7 million, which represented the value of the stock carried on NH's books. Both the bondholders and the agency would seem mistaken. Since N.H. will be obliged under the plan to acquire 20,770 shares of B P stock now publicly held at a total cost of $2.3 million, this sum should be deducted from the $8.1 million, leaving $5.8 million — $1.1 million more than the Commission's valuation.

In an effort to justify the $4.7 million figure the Commission and Penn Central point to another feature of the reorganization plan which is claimed to reduce the value of the B P stock to NH. When the B P plan was approved by the Commission, there was thought to be substantial possibility that parts of its roadway might be sold to the Massachusetts Bay Transit Authority or be taken by that or some other governmental agency for highway purposes at prices substantially in excess of their appraised value under the reorganization plan. To take this into account, the plan provides for the issuance of Certificates of Contingent Beneficial Interest (CCBI's) as to the proceeds of any such sale, involving more than $500,000, which occurs before December 31, 1978. 48% of the CCBI's are held by the New Haven, 52% by other ex-shareholders. The Commission argues that because N.H. will be obliged to pay out 52% of the net proceeds of such sales without a deduction for the value of the properties already credited in the reorganization proceedings, the agency was justified in deducting the $1.1 million. Apart from the fact that the Report does not rely on this as the basis for the discount, it seems highly questionable, particularly in light of current market assessments of the value of the CCBI's, that any reduction is justifiable.

Before leaving liquidation value, we should discuss some of the bondholders' many criticisms we do not accept. Foremost among these is the Commission's failure to allow for "going concern" value. The bondholders submitted an estimate that a rebuilding of N.H. would require some $55 million for such intangibles as market investigations, organization costs, engineering expenses, and interest during construction. The agency said this estimate was vastly overstated, finding that, at most, 10% of the figure was justified. It refused to credit the estate even with this lesser sum, however, on the ground that "[i]t is not realistic to assume that a potential buyer would pay the liquidated value of the N.H. assets and then pay additional amounts representing elements of going concern value in the face of NH's past deficit operations and its bleak prospects for the future." 687. We agree. Broadly speaking, there is no "going concern value" in a deficit ridden railroad being liquidated as real estate or scrap; the bondholders cannot have it both ways. The argument has slightly more force — but not enough to persuade us — when presented in a much more limited form. It is possible, perhaps even likely, that rather than suffer termination of service, some of the states would condemn parts of NH's operating properties, and the bondholders rely on a recent decision of the New York Court of Appeals, Port Authority Trans-Hudson Corp. v. Hudson Rapid Transit Tubes Corp., 20 N.Y.2d 457, 472, n. 4, 285 N.Y.S.2d 24, 33 n. 4, 231 N.E.2d 734 n. 4 (1967), as requiring payment for "going concern" values regardless of anticipated losses in such event. However, that case dealt with a unique situation where the larger part of the property being condemned, tunnels under the Hudson River, had no value for any purpose save that for which they were being taken and though there are dicta supporting the bondholders' position, it is at least doubtful that the rule would be applied in other situations in which the condemnee's equities were not so compelling. Certainly federal law does not require payment of going concern value in condemnation when liquidation constitutes a property's "highest and best" use since the assets have substantial value for that purpose but no earning power. See 2 Orgel, Valuation Under Eminent Domain § 194 at 41 (1953). Kimball Laundry Co. v. United States, 338 U.S. 1, 69 S.Ct. 1434, 93 L.Ed. 1765 (1948), upon which the bondholders mistakenly rely, makes it clear that "going concern value" must be allowed only when there is reason to believe that as a result of investment in intangibles, the earning capacity of the enterprise is greater than is "commonly * * * given to such a combination of land, plant and equipment." Id. at 9, 69 S.Ct. at 1439. Consequently, the states would not be compelled to resort to condemnation since the I.C.C. could constitutionally condition any certificates of abandonment issued to N.H. on its agreeing to sell its assets at their scrap value to a purchaser intending to operate the properties as a railroad, as it has done in similar situations in the past. See, e.g., Jay Street Connecting R.R. v. United States, 174 F. Supp. 609, 617 at n. 2 (E.D.N.Y. 1959).

Somewhat similar considerations apply to the bondholders' contentions that the Commission should have included tax benefits to Penn Central in computing NH's liquidation value. These are of two different sorts. As to one, large loss carry-forwards, amounting to $82.7 million as of December 31, 1965, realization will depend on Congress' enacting legislation which the Treasury Department has opposed. The other lies in the potential for increased depreciation allowances and, on sales or abandonments of assets, either increased loss deductions or reduced capital gains arising from the excess of NH's tax basis for its properties over their cost to Penn Central. This is subject to some legal doubt, and the value of both types of benefit is also rendered uncertain by the large reservoir of tax deductions already available to Penn Central. What is decisive, however, is that benefits of this kind do not fit the notion of liquidation value. Since N.H. could not pass such benefits on to persons buying its property on a piecemeal basis, it is not entitled to have them added to liquidation value for purposes of sale to Penn Central, as it would be if the properties were being valued on an operating basis, cf. Erie-Lackawanna R.R. v. United States, supra, 279 F. Supp. at 346. The only consideration appropriate for these items under the circumstances is in somewhat reducing the cost of the acquisition to Penn Central.

For different reasons we decline to fault the Commission for rejecting the testimony of an expert witness for the bondholders that the Trustees' liquidation study did not take into account that little or no expense would be entailed in the sale of substantial sections of NH's lines since these "would be candidates for purchase by other carriers." While the point is appropriate, the witness was uncertain both as to what segments would prove to be successful "candidates" for sale "in place" and the prices at which such sales would be made. He concluded only that the Trustees' estimate was "substantially understated," without committing himself to the dollar amount of the adjustment allegedly required. The Commission was justified in not undertaking an assignment the witness had declined, especially since "the simple fact is that no other railroad has displayed an interest in purchasing any portion of NH." 679. The bondholders' criticisms of a variety of other findings — notably, the values assigned by the Commission to NH's investments in cooperative undertakings with other railroads and to passenger cars — must be rejected as seeking to have us invade the Commission's province.

IV. The Value of the Consideration to be Received.

The Commission placed the following values upon the property to be received by N.H. under the Purchase Agreement:

Millions

(1) 950,000 shares of Penn Central stock at $87.50 a share $ 83.1

(2) $23 million principal amount of Penn Central 5% bonds 20.0

(3) Assumption by Penn Central of Boston Providence mortgage 3.0

(4) Purchase Price Adjustments Under Purchase Agreement 9.5

(5) Cash 8.0 ------ 123.6

The bondholders assail items (1) and (4) as too high; they also contend the computation failed to take into account that $16.2 million of the price may have to be used to pay accrued real estate taxes in order to permit the conveyance free of liens which the Purchase Agreement requires. We reject the first criticism but sustain the other two.

The bondholders say there was no adequate basis for valuing Penn Central stock at $87.50 a share, when at the time of the I.C.C.'s decision, Pennsylvania stock — which was to be exchanged on a one for one basis for new Penn Central common — was being traded in the $60 range. In defense of its higher figure, the Commission cites the testimony of William J. Kirk, one of the N.H. Trustees, who stated that in concluding the Purchase Agreement he assumed the stock's price would be fluctuating between $75 and $100 at the time of the closing. 688-89. As his testimony does not reveal the basis for this belief, the Commission sought to establish that a proper analysis of Penn Central's prospective earning power led to a similar conclusion. It first capitalized at 8% the net income earned by the Pennsylvania and the Central separately in 1965 and thus concluded that the value of a Penn Central share, before considering merger savings, was $68. It then capitalized the $81 million in annual savings anticipated from the merger, subtracting only the $8.5 million it expected N.H. to lose annually after inclusion, thereby arriving at post-merger earnings of $8 per share, and concluded that the "inherent quality" of the stock demanded a price of $100, thus indicating that its $87.50 estimate was fair. Against this it is urged that the Commission's calculations overlook that the annual savings of $81 million were not expected to be attained until eight years after the merger was concluded, Pennsylvania R. Co. — Merger — N.Y. Central, 327 I.C.C. 475 at 501 (1966), with a great part of them dependent upon the completion of a four year modernization program. Id. at 489-90.

While this complaint might have considerable validity if the Commission had valued the Penn Central stock at $100 per share, that is not what the Commission did. If $72.5 million of annual net savings would raise the market price from $68 to $100, only some $45 million would be needed to raise this to $87.50. When the stock market's tendency to look to future earnings is borne in mind, it was not unreasonable to estimate that a price of $87.50 would be realized by the time of the closing — assuming that normal market conditions then prevailed. The increase in the price of the stock to around $85 per share at this writing at least shows that the Commission's figure is not wild speculation.

The bondholders assert also that there was no occasion to indulge in any guessing as to the value of the Penn Central stock since the I.C.C. could simply have required issuance of such number of shares, taken at average stock market prices over a period before the closing, as were needed to produce the desired total price. Due to the cyclical nature of the railroad industry this might operate quite unfairly. While if Penn Central had had a history extending over a cycle, an average during such a cycle might have been fair, no such experience was available.

On the other hand, the Commission's treatment of the consideration to be paid in satisfaction of Penn Central's obligations under the "Purchase Price Adjustment" provisions, item (4), must be regarded as erroneous. Penn Central has the right to make this payment either in cash, stock or in 5% bonds. While the agency recognized that the $23 million of Penn Central bonds that the New Haven is to receive would not be worth par and therefore valued them at only $20 million, it did not apply a similar discount to the bonds of the same type which Penn Central can use to satisfy its obligations under the "adjustment" clause. We find no force in the response that since Penn Central has the option to pay this sum in cash or stock, the I.C.C. could not assume that payment will be made in bonds. It is apparent that Penn Central will choose the cheaper way out. The Commission's appraisal is thus overstated by $1.3 million.

We likewise see no sufficient answer to the bondholders' contention as to the failure to make a deduction for the $16.2 million of accrued real estate taxes on property required to be transferred free of liens, which, they say, resulted either in an overstatement of the consideration to be received by the estate or an understatement of liquidation value. The Report does not discuss this and only Penn Contral seeks to justify it, on the basis that the reorganization court can direct a sale free of liens and it may thereafter be possible to persuade state and local taxing authorities to accept a reorganization plan providing less than full satisfaction for their claims. But even if part of the taxes were forgiven, the effect of the error would not be corrected. For the agency deducted the full $16.2 million when calculating the Estate's net liquidation value.

Before considering the third of the Commission's calculations, the cost of the acquisition to Penn Central, we should examine whether the Commission made errors in the direction of overstating NH's liquidation value that would reduce or eliminate what may be a discrepancy exceeding $45 million between that figure and the value of the consideration to be received. Only two of those suggested warrant discussion. The Commission's brief takes a deduction of $3 million for 1966 depreciation. This item is not found in the Report and its derivation is obscure. A much more significant item is raised by the claim that before the N.H. Trustees could even begin a liquidation, lengthy abandonment proceedings would be required, during which large cash losses would be incurred and the property would further depreciate. The bondholders respond that if inclusion of N.H. in Penn Central had not been in prospect, the N.H. Trustees would have been under a duty to initiate abandonment proceedings as soon as they reported, on December 31, 1963, that there was no prospect for independent operations of NH, and that December 31, 1966 is thus an entirely reasonable date for the beginning of the assumed liquidation. Indeed, they insist they are entitled to be fully protected against losses thereafter — a contention we shall discuss below. Our rejection of that contention, see section VII of this opinion, means that a considerable portion of the losses will be borne by the N.H. estate, and there would be no justification for a double deduction. Beyond that, the Commission had to pick a date for valuation and December 31, 1966 was a reasonable choice. If we were to posit a liquidation beginning only some years hence, we could not simply deduct interim cash losses and depreciation but would have to ask the Commission to revalue the assets and this might yield a higher price. 698.

V. Cost of Acquisition of N.H. to Penn Central.

The Commission found the cost to Penn Central of acquiring the assets of N.H. to be as follows:

Millions

(1) 950,000 shares of Penn Central stock at $87.50 per share $ 83.1

(2) $23 million principal amount of Penn Central 5% bonds 23.0

(3) Assumption by Penn Central of Boston Providence mortgage 3.2

(4) Purchase Price Adjustments under Purchase Agreements 9.5

(5) Cash 8.0

(6) Equipment obligations 22.6

(7) Labor protection and pension costs 9.0 ------ $158.4

We have concluded above that the Commission's determination of item (1) was within the powers accorded to it. While items (2) and (3) also find their analogues in the computation of the value of the consideration to NH, the Commission here priced them differently — taking the face value of the bonds rather than their estimated value at the time of closing, on the basis that Penn Central would ultimately have to pay this. We find this unpersuasive; no one would regard the issuance of a long term non-interest bearing note as costing the issuer the face amount, and the same principle covers a bond carrying an interest rate below that necessary to enable it to sell at par. Items (2) and (3) should thus be reduced to $20 million and $3 million respectively and, for reasons indicated above, item (4) should be reduced to $8.2 million.

The bondholders' attack on item (6) may have merit. A substantial proportion of New Haven's equipment, valued at $8.7 million, stands as collateral for $14.3 million in equipment loans. Another substantial portion, valued at $28.8 million, stands as security for only $10.1 million in loans. The bondholders argue it was unreasonable to believe that Penn Central will assume $14.3 million in equipment obligations to get the continued use of $8.7 million in equipment, when under the Purchase Agreement it can require the N.H. Trustees to disaffirm these obligations. The Commission held, however, that the Penn Central would assume all the equipment obligations since the equipment is "required to continue N.H. operations." 693. The Report does not satisfy us that the Commission fully understood the bondholders' argument; Penn Central might itself have surplus equipment or be able to get it from other sources more cheaply than by assuming obligations 165% of the equipment's worth. Unless there is some better explanation why Penn Central would follow such a seemingly ill advised course, the cost to Penn Central should be reduced by $14.3 million less the cost of additional equipment that must be secured to replace N.H. equipment that would be repossessed. On the other hand, we are not persuaded that the Commission overstated item (7).

As against the items where some reduction in the cost of acquisition to Penn Central is demanded, much larger ones were omitted. One of these is Penn Central's compulsory participation in NH's losses from the consummation of the Penn Central merger, 717-19, which, if NH's inclusion should take three years, will amount potentially and, very likely, actually to $16.5 million even on the Commission's formula. Moreover, we do not see why the burden of future N.H. losses, estimated by the Commission at $8.5 million after economies from consolidation should not likewise be considered a cost from a practical although not an accounting standpoint. Capitalized at an 8% basis this would amount to an additional $106 million. We thus find the total cost of the acquisition of N.H. to Penn Central under the arrangements approved by the Commission will exceed $250 million. This figure must, of course, be evaluated in light of the ultimate $72.5 million annual savings from the Penn Central merger after allowing for N.H. losses — a value of some $900 million, capitalized on the same basis.

VI. The Effect of our Redetermination on the Commission's Order.

In our view the Commission's finding "that the consideration is at least equivalent to the liquidation value," 697, could well be erroneous in the $45-50 million range. Hence the consideration may be only some two-thirds of liquidation value. We must now decide what the consequences of that error are.

Citing some or all the passages from the Report we have quoted at the outset of this opinion, the defendants argue that the Commission never said that N.H. must receive liquidation value, and that any such statement would be bad law. Agreeing there are passages in the Report that lend support to the former contention, we find the argument insufficient to avoid the need for a remand. This is not a case in which the agency, while understanding the basic outlines of the problem before it, explains the grounds for its decision inadequately, Phelps Dodge Corp. v. N.L.R.B., 313 U.S. 177, 196-197, 61 S.Ct. 845, 85 L.Ed. 1271 (1941), or unsoundly, SEC v. Chenery Corp., 318 U.S. 80, 63 S.Ct. 454, 87 L.Ed. 626 (1943). In finding that the consideration was the equivalent of NH's liquidation value, the agency misapprehended the facts. While remand would not be required even in such a case if the improper findings could justly be considered subsidiary unless there is "solid reason to believe that without the subsidiary finding the agency would not have arrived at the conclusion at which it did arrive," Communist Party of United States v. Subversive Activities Control Board, 367 U.S. 1, 67, 81 S.Ct. 1357, 1395, 6 L.Ed.2d 625 (1961), here the agency's mistake went to the heart of its analysis. We cannot be certain that if the Commission had known of its errors of fact, it would have approved the price provided by the Purchase Agreement, even if it had held the views of the law the defendants now attribute to it. This being so, we must send the case back. See Braniff Airways, Inc. v. CAB, 126 U.S.App.D.C. 399, 379 F.2d 453, 465-467 (1967) (Leventhal, J.); NLRB v. Reed Prince Mfg. Co., 205 F.2d 131, 139 (1st Cir.), cert. denied, 346 U.S. 887, 74 S.Ct. 139, 98 L.Ed. 391 (1953).

While we could end this section of the opinion at this point, we think it will be useful to indicate our views as to the proper legal standard. There is some initial appeal in the position that when the cost of the acquisition of N.H. to Penn Central under the Purchase Agreement, including capitalization of future losses, will exceed the NH's liquidation value by $100 million, the Commission is not required to make Penn Central pay still more; the price, it is said, must be reasonable to the buyer as well as the seller. But this ignores the constitutional right of N.H. to cease operation and liquidate. We see no reason to question the validity of Justice Holmes' decision in Brooks-Scanlon Co. v. Railroad Commission, 251 U.S. 396, 40 S.Ct. 183, 64 L.Ed. 323, forbidding the State of Louisiana to require a railroad to continue its deficit operation with no hope for profits in the foreseeable future. See also, Mora v. Mejias, 223 F.2d 814 (1 Cir. 1955); Railroad Commission v. Eastern Texas R.R., 264 U.S. 79, 85, 44 S.Ct. 247, 68 L.Ed. 569 (1924); Bullock v. State of Florida ex rel. Railroad Commission, 254 U.S. 513, 41 S.Ct. 193, 65 L.Ed. 380 (1921); cf. Pacific Telephone Teleg. Co. v. Tax Comm'n, 297 U.S. 403, 413, 56 S.Ct. 522, 80 L.Ed. 760 (1936). We cannot accept the argument that all this was changed by the addition of § 1(18), requiring the Commission's consent to abandonment, to the Interstate Commerce Act by the Transportation Act, 1920, or by the enactment of § 77 of the Bankruptcy Act in 1933. These statutes were designed to assure orderly procedures and to afford methods that might avoid the ultimate solution; they were not intended to override constitutional guarantees, cf. In re Third Avenue Transit Corp., 198 F.2d 703, 707 (2 Cir. 1952), and would not be effective if they had been. In the usual case under § 5(2)(d) the Commission doubtless is not required to fix a price as high as the road to be acquired might obtain elsewhere if its value to the acquiring road will not be that great, although we have intimated that it could lawfully do so on a showing that inclusion was required by the public interest and that this was the only way to meet the owners' "reasonable expectations, perhaps based on inclusion in another system that would be less in the public interest." Erie-Lackawanna R.R. v. United States, supra, 279 F. Supp. at 338. But when the asking price for the acquired road is what the Constitution demands the owners must receive as a condition to allowing the road to be operated in the public service, we think — subject only to what we shall say hereafter as to interim losses — that the Commission may not fix, or approve, anything less.

We recognize, however, that this legal issue is one of novel impression which only the Supreme Court can authoritatively determine. We therefore suggest in the interest of avoiding another remand that if the Commission disagrees with our view, it should feel free to formulate its own and supply the necessary subsidiary findings. The Commission is therefore directed to make detailed findings as to each element of liquidation value and of the consideration to be paid by Penn Central on a basis consistent with this opinion. If it agrees with our view that the price cannot be less than the liquidation value, it shall prescribe suitable terms. If it does not, it shall state the pricing theory it has adopted, the reasons for it, and the results, and shall make whatever detailed subsidiary findings are required to support this.

VII. The Claim of the Trustee of the Harlem River Division Mortgage for Assumption.

USTC claims it is unfair that the Harlem River bondholders should be left dependent on Penn Central's decision. It says that, alone among NH's bondholders, those under the Harlem River mortgage are fully secured, and that it would be inequitable to deprive them of their lien and leave them only with a right to assert priority over other N.H. bondholders in the proceeds of the sale. It cites, as an instance of what might lie ahead, the proposal in the N.H. Trustees' Plan of Reorganization that the Harlem River bondholders accept a package consisting of seven year 4 1/4% subordinated debentures, secured in main part by Penn Central common stock, for the principal of their claim, and even more speculative N.H. common stock for the accrued interest. USTC finds the Commission's refusal to order assumption of the Harlem River bonds by Penn Central particularly frustrating in light of the lack of objection by anyone.

While there is much force in USTC's position, we think its request deals with a matter better left to the reorganization court. There is no reason to believe that the District Court for Connecticut would approve a plan that did not provide fairly for the Harlem River bondholders. See Group of Institutional Investors v. Chicago, M., St. P. P.R.R., 318 U.S. 523, 564, 63 S.Ct. 727, 87 L.Ed. 959 (1943). While that court does not have the power that we do, in reviewing the Commission's order under § 5 of the Interstate Commerce Act, to require a change binding on Penn Central, it has ample means to see to it that fair provision is made for the Harlem River bondholders and we would not wish to impinge on its freedom of action.

VIII. Protective Conditions and Loss Sharing.

The N.H. bondholders contend the Commission erred in not requiring an immediate takeover of N.H. by Penn Central, in the form of a lease or otherwise; that the arrangements for a $25 million loan and sharing in losses which the Commission prescribed as a substitute are unfair; and that the Commission should have imposed "Appendix G" financial conditions for the benefit of N.H. similar to those it stipulated in favor of Erie-Lackawanna, D H, and B M. See 259 F. Supp. at 970-971; 279 F. Supp. at 326-329.

Our last opinion recounted the steps taken by the Commission through September 1967, to keep N.H. alive pending its inclusion in Penn Central, 279 F. Supp. at 332-335. In the Report the Commission adhered to its refusal to direct inclusion of N.H. in Penn Central without the terms having been fixed by it and passed upon by the reorganization court and found itself convinced "that a quick takeover in this case could not be accomplished by means of a lease," 705 — views it was justified in holding. It prescribed a $25 million loan with provisions for loss-sharing by Penn Central, and evidently concluded that with these arrangements in effect there was no need for Appendix G conditions. The loan can be taken down whenever the N.H. Trustees' cash balance falls below $5 million. It is to be evidenced by Trustees' certificates bearing interest at the prime rate of Morgan Guaranty Trust Co. at the time and maturing December 31, 1971; these can be retained, sold or applied toward the price of acquiring the N.H. assets, except insofar as they have been cancelled under the loss-sharing formula. This begins with the deficit in net railway operating income. That sum is to be decreased by any income from the GCT properties which Penn Central agrees to pay NH, by depreciation which Penn Central bears in any event under the Purchase Agreement, and by retirement losses. Of the deficit thus adjusted, Penn Central is to bear 100% in the first year after the Penn Central merger, 50% in the second year, and 25% in the third, but in no event more than $5.5 million in any year.

The broadest position of the N.H. bondholders is that Penn Central should have been required to shoulder 100% of the losses for whatever period is needed to effectuate inclusion of N.H. or, for that matter, until abandonment is authorized if inclusion should prove impractical. This position stems from the considerations developed and to some extent approved in Section VI of this opinion. The bondholders were entitled, they say, to have the N.H. Trustees begin abandonment proceedings in 1963 when the Trustees reported that future operations of N.H. as an independent railroad were unfeasible; that the Trustees failed to do this only because inclusion of N.H. in Penn Central appeared to offer an alternative preferable in the public although not in the bondholders' interest; that the Commission found the Penn Central merger to be in the public interest only if N.H. were included; and that it would thus be inequitable — indeed unconstitutional — if the liquidation value of N.H. as of December 31, 1966, were to be eroded by losses "which reduce the value of the assets or represent deferral of expenses that have a prior claim on the N.H. estate," 703-04, n. 33. Supporting the Commission's course Penn Central argues against this that a considerable part of the losses could be avoided if N.H. could be immediately included in Penn Central; that Penn Central cannot fairly be penalized for NH's inability to consummate the inclusion more promptly; and that to impose all the losses on it would encourage the N.H. bondholders to litigate endlessly. It adds, properly in our view, that the Commission's statement, 327 I.C.C. at 524, that the Penn Central merger would not be consistent with the public interest "without complete inclusion of NH" is by no means the same as saying it was only the inclusion of N.H. that rendered the merger in the public interest; the statement rather must be read along with the Commission's finding, that, absent inclusion, Penn Central "with new routings and improved service, could undoubtedly wean substantial traffic away from N.H. and leave that already moribund carrier in perhaps an irretrievable situation." 327 I.C.C. at 522.

While we do not find choice between these positions easy, we accept the Commission's view that it was not bound to require Penn Central to assume all losses of N.H. subsequent to the Penn Central merger. No one seriously questions that the N.H. Trustees made the right choice in pressing for inclusion in Penn Central rather than in seeking permission to abandon. To say they should have pursued the two paths simultaneously ignores the limits on human energies and on the Commission's time. If the Trustees had presented abandonment applications as an alternative to sale, the Commission would almost certainly have deferred their consideration for a reasonable period while inclusion was being explored. If inclusion had been found impracticable, the N.H. bondholders would have had to shoulder losses between the date of such a determination and the grant of authority to abandon. We cannot see why the contrary result entitles them to have the entire burden of such losses shifted to Penn Central. Although we agree that by investing in a railroad the bondholders did not surrender their constitutional right not to be required to operate the property at a prepetual loss, they did subject themselves to such interim losses as are reasonably incident to working out the solution most consistent with the public interest. See Meyers v. Jay Street Connecting R.R., 259 F.2d 532 (2 Cir. 1958), and 288 F.2d 356 (2 Cir.), cert. denied, 368 U.S. 828, 82 S.Ct. 49, 7 L.Ed.2d 31 (1961). Cf. Palmer v. Commonwealth of Massachusetts, 308 U.S. 79, 89, 60 S.Ct. 34, 84 L.Ed. 93 (1939).

If we are right in this, the Commission was justified in setting a ceiling on Penn Central's participation in losses and in providing that it bear only a fractional share. The N.H. bondholders say that even if this be true, the downward progression of Penn Central's percentage constitutes an in terrorem provision designed to thwart them from fully presenting their claims to fair treatment. We agree that desire to curtail litigation is an insufficient reason for the downward progression. The Commission should not have assumed that the bondholders' attacks would necessarily lack merit; indeed this opinion shows the contrary. Moreover Penn Central has as much opportunity as the bondholders to prolong litigation. Since the Commission expressly placed the downward slide on this basis, the formula cannot stand, see Justice Douglas' dissent in the Penn-Central Cases, 389 U.S. at 557-558, 88 S.Ct. 602, and we must remand for reconsideration. In doing this we in no way suggest that the agency is obliged to place more of the burden upon Penn Central unless, upon further reflection, it concludes that either the Constitution or equity so requires. It is not barred, for example, from following an ancient precedent and splitting the losses in half, despite the fact that its original formula required Penn Central to accept 58 1/3% of the anticipated losses if inclusion should take three years and 75% if it comes at the end of the second year, assuming that the losses do not increase. Another formula perhaps meriting consideration would be to gear the portion of the losses not assessable against Penn Central to the savings it could accomplish if inclusion had been effected. It could well be, however, that still further analysis is required. The original loss sharing formula, limiting Penn Central's burden in any year to a maximum of $5.5 million, was clearly predicated upon the agency's view that NH's fortunes had reached their nadir in 1967 and that the outlook for 1968 and beyond was somewhat less dismal 718-19. The Commission is free to consider, in the light of more current information, whether this assumption is correct. Similarly the Commission's confidence that the $25 million loan will provide for NH's cash needs pending inclusion may prove misplaced. It follows from the principle of those cases remanding administrative decisions because of staleness of the record that where, as here, remand of an issue is required on other grounds, the Commission should not be prevented from taking steps to protect the public interest by virtue of an earlier failure to predict the future accurately. Consequently, if additional funds are required to assure NH's continued operation pending inclusion, the agency should take steps to avoid the danger by appropriately recasting these provisions.

There remains the contention that the Commission was bound to prescribe Appendix G type conditions for the protection of NH. We have previously expressed our view that there was no real analogy between the three roads benefitting from such conditions, whose future relationship with Penn Central was almost certain to be highly competitive, and NH, the Pennsylvania's historic connection to New England and now to be made a component of the Penn Central system. See Oscar Gruss Son v. United States, 261 F. Supp. at 389-391. We have also said, that "The unlikelihood that Penn Central would avail itself of its opportunities for diversion from N.H. in the interval prior to inclusion would be heightened if it is now required to bear or share in operating losses." Erie-Lackawanna R.R. v. United States, 279 F. Supp. at 336. It appears also that a major rerouting of traffic between New England and points west of Albany, which is indeed in contemplation, depends on the completion of a new yard at Selkirk, N.Y. some years off. Furthermore Appendix G financial conditions might be quite unfair to Penn Central when applied in favor of a road whose freight revenues are deteriorating for reasons wholly unrelated to the Penn Central merger. Under all the circumstances we cannot say the Commission acted arbitrarily or capriciously in refusing to impose Appendix G conditions.

IX. Protective Traffic Conditions for Erie-Lackawanna at Maybrook, N.Y.

N.H. has long maintained an important interchange of New England freight traffic with two sets of carriers at Maybrook, N.Y. — the route from Maybrook crossing the Hudson River at Poughkeepsie and joining NH's main line at New Haven. One group, comprised of the Central Railroad of New Jersey (CNJ), the Reading and the Western Maryland, provided service to and from points in southeastern Pennsylvania and western Maryland and beyond, which competed primarily with the service furnished by the Pennsylvania via New York harbor. Erie-Lackawanna (EL) supplied service through Binghamton, N.Y. and beyond, competitive both with the Pennsylvania and with the New York Central's service through Albany. The EL connection has been exceedingly important, 124,435 cars having been handled with it via Maybrook in 1966 as against no more than 97,391 with CNJ (51,228 via Maybrook); and 102,980 handled by N.H. and Pennsylvania via New York harbor. N.H. has solicited preferentially for the Maybrook gateway since this generally affords its longest haul, and has worked cooperatively both with CNJ, Reading and Western Maryland and with EL to develop excellent service and connections.

At the request of CNJ, Reading and Western Maryland, the Commission imposed in their favor special traffic conditions, set forth in Appendix I to its Report, 752-53, to maintain the quality of service and equality of solicitation for the Maybrook interchange pending decision on the application for inclusion of these lines in a proposed N W-C O system. However, it denied a similar request by EL. Noting Penn Central's contention that "EL is really seeking a built-in competitive advantage for itself and the N W system," 738, the Commission stated:

By imposing conditions of the type suggested, we would not be merely sustaining a necessary operation until it could fend for itself, but we might well be creating a permanent division of markets with undesirable effects on the competition.

It should be noted that N W has not requested the imposition of any conditions in this proceeding. Evidently, through prescription of the conditions it has proposed, EL hopes to gain some protection against N W adopting a policy of employing the Mechanicsville gateway to and from New England and thereby short-hauling EL. Penn-Central should not have to shoulder the responsibility of safeguarding EL from its parent. In any event, so long as the conditions imposed to protect CNJ, Reading and WM remain in effect, they would also benefit EL.

EL petitioned for reconsideration. In its Third Supplemental Report, rendered March 1, 1968, the Commission denied this without prejudice to EL's renewing the application on the basis of actual experience of injury. The Commission amplified the reasons previously given, as follows:

(1) EL would have the benefit of the standard routing conditions set forth in Appendix I of the initial report approving the Penn Central merger, 327 I.C.C. at 564, 565.

(2) The special conditions imposed for the benefit of CNJ; Reading and Western Maryland "will redound to the benefit of EL."

(3) Penn Central's announced intention to route via Maybrook traffic emanating east of Columbus, Ohio, indicates Penn Central's self-interest in the preservation of Maybrook as a major gateway.

(4) EL has "overemphasized the effects that the N.H. inclusion will have upon rail competition in Southern New England" and preservation of rail competition in Southern New England is of secondary significance, particularly since the area's relatively low volume of traffic is now barely sufficient to support the operations of the NH.

(5) "The Appendix G conditions (traffic and indemnity provisions) provide sufficient protection of EL pending its inclusion in the N W system."

(6) "Since EL's inclusion in the N W system is now assured * * *, its ability to compete for traffic between New Haven territory and the west will be strengthened."

(7) "Affording EL the protection it now seeks might well give it and N W superior competitive advantage."

(8) The Commission's "reserved powers are adequate to provide additional protection in the future should circumstances require it."

We agree with EL that this falls short of the meaningful analysis as to anti-competitive impact on which the Supreme Court has insisted. McLean Trucking Co. v. United States, 321 U.S. 67, 87, 64 S.Ct. 370, 88 L.Ed. 544 (1944); Schaffer Transportation Co. v. United States, 355 U.S. 83, 78 S.Ct. 173, 2 L.Ed.2d 117 (1957). As to (1), while the standard routing conditions, which have been routinely imposed since Detroit, Toledo Ironton Control, 275 I.C.C. 455 (1950), require that "all routes and channels of trade via existing junctions and gateways" be kept open, with neutrality of handling traffic as among connections, they do not speak to the quality of service afforded. The Commission clearly did not consider them adequate to protect CNJ, the Reading and the Western Maryland in this case and has not so considered them in many other cases EL has cited to us. As to (2), while the special conditions imposed for the benefit of CNJ, Reading and Western Maryland may indeed assure against deterioration of the physical condition of the Poughkeepsie bridge route, they give no protection as to schedules affording effective connection with EL or many other factors important to maintaining service on a competitive basis. Similar considerations apply to (3); indeed they apply a fortiori since Penn Central is free to change its mind and would seem to have strong incentive for doing so as soon as the Selkirk facility is completed. The Commission has not shown why EL has "overemphasized" the anti-competitive effects of N.H. inclusion; anyway the point as issue is not competition "in Southern New England" but between New England and points to the west. The Appendix G conditions ceased to be effective on April 1, 1968, when EL was included in N W. While inclusion in N W will indeed strengthen EL's ability to compete by enlisting the aid of N W solicitation and affording it one system access to many new points, especially in the middle west, this will be of little value if Penn Central should cause the quality of the connecting service with EL at Maybrook to be downgraded in favor of New York harbor or, more particularly, after completion of the Selkirk facility, in favor of the Albany routing.

If the Commission's action is to be defended, it must thus be on the ground, numbered (7), that EL's inclusion in the N W system, renders the present imposition of conditions similar to Appendix I inappropriate. We think it may, at least in some degree. Penn Central and N W are now the two largest rail systems in the east, competing vigorously against each other for traffic between many hundreds of points. To require Penn Central actively to solicit traffic for N W, see, e.g., condition 8 of Appendix I, 331 I.C.C. at 753, might rationally be deemed to go beyond what the economics of the case require or what human nature could endure. On the other hand, this reasoning would not necessarily justify the denial of conditions concerning the maintenance of train schedules, standards of service and competitive rates. And if EL has made a case for this kind of protection now, we do not consider that the extracts from the Supreme Court's decision in the Penn Central Merger Cases, 389 U.S. at 497 n. 3, 514, and 519-520, 88 S.Ct. 602, which the Commission cited in its Third Supplemental Report, would justify denial on the basis of the Commission's reserved power to attempt to undo damage that has occurred.

Since the agency's opinion offers neither an analysis of the anti-competitive effect of inclusion without protective conditions or convincing evidence that pre-existing coordination with EL will be maintained, "we would shirk our duty were we summarily to approve the Commission's evaluation of the record without determining that the agency's evaluation had been made in accordance with the mandate of Congress" that the policies of the anti-trust laws be conscientiously applied. Schaffer Transportation Co. v. United States, supra, 355 U.S. at 88, 78 S.Ct. at 177. EL's request should be reconsidered by the Commission in the light of what we have said as to the inadequacy of most of the reasons advanced for the denial. Whether or not the doctrine of SEC v. Chenery Corp., 318 U.S. 80, 63 S.Ct. 454, 87 L.Ed. 626 (1943), requires us to remand for further consideration by the Commission rather than attempting to model conditions on our own part, it surely permits us to take that course.

We therefore vacate so much of the Commission's orders of November 16, 1967 and March 1, 1968, as finds the acquisition of N.H. by Penn Central on the terms provided in the Purchase Agreement and the interim loss-sharing and loan arrangements to be just and reasonable, and as denies EL's request for special protective traffic conditions at the Maybrook, N.Y. gateway, and remand the cause to the Commission for further proceedings consistent with this opinion. We urge the Commission and all parties to proceed with the utmost expedition. To the extent that new testimony is required, this should be held to a minimum; we do not mean by this opinion that parties who had a full opportunity to present evidence must be given a new turn. See Jaffe, Judicial Control of Administrative Action 715 (1965). We reiterate our suggestion that the Commission make sufficiently detailed findings to avoid need for a further remand if, after the case has returned to us, the Supreme Court on appeal should disagree with any of our views. We see no need for injunctive relief at the present time.

It is so ordered.