R. LANIER ANDERSON, III, Circuit Judge:

These two consolidated petitions for review of an order by the Interstate Commerce Commission ("ICC" or "Commission") involve a single set of facts and require that this circuit for the first time address a provision added to the Interstate Commerce Act by the Railroad Revitalization and Regulatory Reform Act of 1976 ("Reform Act"), and regulations promulgated by the Commission interpreting a second provision added by the Reform Act. The new statutory provision requiring interpretation, Section 206 of the Reform Act, now revised and codified at 49 U.S.C.A. § 10729, provides for expedited review by the ICC of capital incentive rates filed by rail carriers subject to the ICC's jurisdiction. This section further provides that if a capital incentive rate becomes effective either through Commission inaction or approval, "the Commission may not, for 5 years, suspend or set it aside as violating" specified provisions of the Interstate Commerce Act. § 10729(b). The regulations requiring interpretation are found at 49 C.F.R. 1109.1 and establish rebuttable presumptions of "market dominance" as defined by § 10709(a). A prerequisite to the Commission's having jurisdiction to approve or disapprove a capital incentive rate is that the carrier have market dominance over the transportation to which the rate applies. § 10709(c).

The petitioners challenge the ICC's decision dated January 15, 1980, in which the Commission refused to review a capital incentive rate proposed by the railroad intervenors for the transportation of coal because the Commission found no market dominance over that transportation. They claim that the ICC erred in (1) finding the railroads' proposed rate qualified for consideration as a capital incentive rate under § 10729, (2) finding that the railroads did not possess market dominance with respect to the particular transportation of coal, (3) finding that even if the presumptions of market dominance had been triggered, such presumptions were rebutted by the existence of geographic competition, (4) failing to adequately state the rationale of its decision, (5) failing to make any findings with respect to discriminatory pricing, and (6) failing to require the railroads to produce evidence within their possession.

These petitions, through no fault of petitioners, have come before this court in a most unusual and unsatisfactory posture. The Commission has declined to file a brief in support of its decision. Instead, after deciding on April 30, 1980, that this case merited further thought, the Commission moved this court to decline to address the merits and to remand these petitions so that it might institute a reconsideration on the record as it now exists. (Hereinafter, the phrase "voluntary remand" shall refer to this ICC — requested remand without consideration of the merits, while the phrase "court-generated remand" shall refer to a remand after consideration of the merits.) The petitioners and the railroads join forces against their former mediator in opposing this motion for a voluntary remand, but for different reasons. The railroads argue that a voluntary remand is neither necessary nor permitted, and moreover, that any court-generated remand is limited in scope as to what the Commission may do to correct any legal errors. The petitioners argue that a voluntary remand is not permitted under § 10729, but that a court-generated remand is necessary and appropriate in this case.

We conclude that we cannot grant the Commission's motion for a voluntary remand, nut must address the merits of these petitions. We find that the railroads' proposed rate does qualify for consideration as a capital incentive rate. However, we vacate the Commission's findings on market dominance and geographic competition and remand.

Our discussion of the issues will proceed according to the following outline:

I. Facts

II. Statutory Framework

III. Standard of Review

IV. Qualification for Treatment Under § 10729

V. Voluntary Remand

VI. Scope of Commission's Authority Upon Remand

VII. Presumptions

A. Rate Bureau Presumption

B. Market Share Presumption

C. Revenue/Cost Presumption

(i) Use of Additives

(ii) Rate of Return in Incremental Fixed Plant Investment Additive

(iii) Double Count

D. Substantial Investment Presumption

(i) Significance of One Terminating Carrier

(ii) Finding of Competing Carriers for the Axial to Coleto Creek Movement

(iii) Finding of Alternative Domestic Sources

(iv) Finding of Alternative Modes

VIII. Geographic Competition

A. Statutory Restriction on Consideration of Geographic Competition

B. Restriction on Consideration of Geographic Competition Within Regulations

C. Commission's Finding of Geographic Competition After the Colowyo Contract

D. Availability of Geographic Competition Before the Colowyo Contract

IX. Section 10741 Attack on Rate as Discriminatory; and Railroads' Duty to Produce Evidence

I. FACTS

Central Power Light Company ("CPL") is an electric utility engaging in the generation, transmission, and sale of electric power to 200 communities in south Texas including the City of Corpus Christi. Since its inception in 1916, CPL has relied upon oil and gas to fuel its boilers. Following the OPEC oil embargo in 1973 and the passage of legislation regulating the use of oil and gas as a boiler fuel, CPL decided to build a coal — fired generating plant at a site near Coleto, Texas. This plant will eventually consist of two generating units. Each unit has a generating capacity of 550,000 kilowatts and will require 30,000 tons of coal per week, or approximately 1,500,000 tons annually. The first unit has been completed and has undergone preliminary start-up operations. It was scheduled to go on line the first quarter of 1980. The second unit is scheduled for completion in 1988.

In 1974, CPL began investigating possible sources of coal to supply the first unit at Coleto Creek. The railroads have placed in the record certain evidence concerning alleged contacts CPL had with various domestic mines as it began looking for a source of coal. Among the sources allegedly contacted were mines in Wyoming, Colorado, Texas and Kentucky as well as in foreign locales. Although CPL was concerned at what it considered unreasonably high prices quoted by carriers and despite not having reached agreement with any carrier on a price for transporting coal, CPL entered in December, 1976, a 25-year contract with Colowyo Coal Company ("Colowyo") for a total of 30 million tons of low-sulfur coal mined in Axial, Colorado.

The transportation of the coal from Axial necessitates a rail movement in unit coal trains beginning with the Denver Rio Grande Western Railroad Company ("DRGW"), the only rail carrier serving the Axial mine. DRGW carries the coal 375 miles, where the Atchison, Topeka and Santa Fe Railway Company ("Santa Fe") continues the movement for the next 865 miles. The Southern Pacific Transportation Company ("SP") is the only carrier serving the plant at Coleto Creek and completes the final 156 miles of the movement. (DRGW, Santa Fe and SP are hereinafter sometimes collectively referred to as "railroads" or "carriers").

After entering its contract with Colowyo, CPL was still unable to negotiate a mutually acceptable price with the railroads. In 1979, CPL began operating the first unit at Coleto Creek on a preliminary start-up basis, using South African coal shipped to the port at Corpus Christi, Texas, and then trucked from the port to the Coleto Creek plant, a distance of 88 miles. CPL used this South African coal for its start-up operations because its delivered price is substantially less than the delivered price of the Colowyo coal at the rates requested by the railroads.

Because the movement from Axial would begin with the start-up of full scale operation in 1980, the railroads broke the impasse with CPL by filing on July 19, 1979, their Notice of Intention to file a proposed capital incentive rate under § 10729 for the movement at $20.85 per ton in shipper-supplied cars. On August 9, 1979, CPL filed a protest, contending that the proposed rate did not qualify for capital incentive rate treatment, that the rate was unlawfully high under § 10709, and that the rate was unlawfully discriminatory under § 10741. The State of Texas was granted leave to intervene in the proceedings before the Commission in support of CPL. After an investigation of the proposed rate, the Commission concluded that it had no jurisdiction over the question of the rate reasonableness because CPL had failed in its burden of proving that the carriers have market dominance. Both Texas and CPL have petitioned this court for review of this decision, which petitions have been consolidated and expedited. The Department of Justice ("Justice") represents the government in this petition as statutory respondent.

II. STATUTORY FRAMEWORK

The 94th Congress in 1976 passed the Reform Act in order to restore the financial stability of the railway system, to enable railroads to rehabilitate and maintain their physical facilities, to improve their operation and structure, and to promote their revitalization. The Reform Act contains two major reforms in the regulatory process to implement this goal of revitalization. First, it mandates the deregulation of rates which were not the product of market dominance. Second, it imposes time limits on the Commission in considering rail rate proposals.

The railroads in this case chose to proceed under the capital incentive rate provision added to the Interstate Commerce Act by the Reform Act and now codified at § 10729. This provision is an important part of Congress' attempt to expedite rail rates proposed to the Commission. Section 10729 provides that a rail carrier may file a notice of intention to file a schedule establishing a new rate "requiring a total capital investment of at least $1,000,000 to implement." The procedural advantage to carriers in filing under this section is that strictly compressed time limits are imposed upon the parties for the presentation of their opposition to the proposed rate. Under § 10729, unless the Commission determines within 180 days after filing the notice that the proposed rate would be unlawful for any reason under the Interstate Commerce Act, the rail carrier may publish the rate, which may not then be suspended or set aside as unlawful for a period of five years from its effective date (except for situations not pertinent here). Section 10729 gives an additional advantage to the carriers in that in any hearing pertaining to a capital incentive rate the burden of proof is upon the protestant. Ex parte No. 327, Rate Incentive for Investment Capital, 353 I.C.C. 754, 767 (1977).

CPL and Texas chose to attack the proposed rate under two sections. First, they attacked the rate under § 10709(c) as exceeding "a reasonable maximum." As part of the attempt to deregulate rail rates, Congress established as a prerequisite for a finding that a rate exceeds a reasonable maximum a requirement that the protestant must establish the carrier has "market dominance over the transportation to which the rate applies." § 10709(c). "Market dominance" is defined as "an absence of effective competition from other carriers or modes of transportation for the transportation to which a rate applies." § 10709(a). Section 202(b) of the Reform Act requires the Commission to establish rules, standards and procedures to determine when a carrier possesses market dominance over a service rendered. This section requires the Commission to design such rules to "provide for a practical determination without administrative delay." In response to this requirement, the Commission in a continuing proceeding, entitled Ex parte No. 320, Special Procedures for Findings of Market Dominance, promulgated the regulations now found at 49 C.F.R. § 1109.1 (1979). These regulations establish four rebuttable presumptions to aid the Commission in ascertaining whether market dominance exists. 49 C.F.R. § 1109.1(f) and (g). The rebuttable presumptions are (1) that "the rate in issue has been discussed, considered or approved upon a rate bureau agreement," (2) that "the proponent carrier has handled 70 percent or more of the involved traffic or movement during the preceding year," (3) that "the rate in issue exceeds the variable cost of providing the service by 60 percent or more," and (4) that the "affected shippers or consignees have made a substantial investment in rail-related equipment or facilities which prevents or make impractical the use of another carrier or mode." The Commission in the case at hand found none of these rebuttable presumptions to have been met, and concluded that no market dominance existed. Accordingly, it found that it had no jurisdiction to determine whether the rate was unreasonably high. CPL attacks this finding, contending that each of the presumptions applies.

CPL made a second attack on the proposed rate under § 10741, alleging that the rate was discriminatory. Section 10741(a) states that a carrier "may not charge or receive from a person a different compensation (by using a special rate, rebate, drawback or other means) for a service rendered, or to be rendered, in transportation the carrier may perform under this subtitle than it charges or receives from another person for performing a like and contemporaneous service in the transportation of a like kind of traffic under substantially similar circumstances." Although there is no prerequisite that a protestant show market dominance before a rate may be found discriminatory under this section, the Commission failed to address CPL's allegations of discriminatory pricing.

III. STANDARD OF REVIEW

In reviewing the Commission's decisions, we are guided by familiar standards. Basic is the rule that the reviewing court must consider whether the decision was based on relevant factors and whether there has been clear error of judgment. Citizens to Preserve Overton Park v. Volpe, 401 U.S. 402, 416, 91 S.Ct. 814, 823, 28 L.Ed.2d 136 (1971). While our "inquiry into the facts is to be searching and careful, the ultimate standard of review is a narrow one. The court is not empowered to substitute its judgment for that of the agency." Ibid; see Bowman Transportation v. Arkansas-Best Freight System, 419 U.S. 281, 285, 95 S.Ct. 438, 441, 42 L.Ed.2d 447 (1974). In the context of rate making, particularly appropriate to one aspect of this opinion, the Supreme Court has noted:

Such decisions `are not to be disturbed by the courts except upon a showing that they are unsupported by evidence, were made without a hearing, exceed constitutional limits, or for some other reason amount to an abuse of power.' Manufacturers R. Co. v. United States, 246 U.S. 457, 481 [ 38 S.Ct. 383, 389, 62 L.Ed. 831] (1918). As this Court has observed, `The process of rate making is essentially empiric. The stuff of the process is fluid and changing — the resultant of factors that must be valued as well as weighed. Congress has therefore delegated the enforcement of transportation policy to a permanent expert body and has charged it with the duty of being responsive to the dynamic character of transportation problems.' Board of Trade of Kansas City v. United States, 314 U.S. 534, 546 [ 62 S.Ct. 366, 372, 86 L.Ed. 432] (1942).

Atchison, Topeka Santa Fe Ry. Co. v. Wichita Board of Trade, 412 U.S. 800, 806, 93 S.Ct. 2367, 2374, 37 L.Ed.2d 350 (1973). (footnote omitted).

Tempering this deference afforded agency action is the requirement that a reviewing court set aside agency action found to be arbitrary, capricious, an abuse of discretion, or otherwise not in accordance with law. 5 U.S.C.A. § 706(2)(A) (West 1977). Variations on this theme inform us that we must be able to discern from a decision the reasons for the Commission's conclusions and the policies it is pursuing. Potomac Electric Power Co. v. United States, 584 F.2d 1058 (D.C. Cir. 1978). "We must know what a decision means before the duty becomes ours to say whether it is right or wrong." United States v. Chicago, Milwaukee, St. Paul Pacific Railroad, 294 U.S. 499, 511, 55 S.Ct. 462, 467, 79 L.Ed. 1023 (1935). Also, an agency must either conform itself to its prior norms and decisions or explain the reason for its departure. Secretary of Agriculture v. United States, 347 U.S. 645, 74 S.Ct. 826, 98 L.Ed. 1015 (1954); Mitchell Energy Corp. v. Federal Energy Regulatory Commission, 580 F.2d 763 (5th Cir. 1978); NLRB v. Sunnyland Packing Co., 557 F.2d 1157, 1160 (5th Cir. 1977); see Frozen Food Express, Inc. v. United States, 535 F.2d 877 (5th Cir. 1976). With these standards in mind, we turn to the issues.

IV. QUALIFICATION FOR TREATMENT UNDER § 10729

Only CPL argues on appeal that the Commission erred in determining that the rate is a capital incentive rate under § 10729 with its five year protection against attack. We treat this question first not only because of its procedural priority but also because a determination of this question bears upon whether a voluntary remand is proper.

The pertinent part of § 10729(a) establishing the criteria to be met before the special rate incentive procedure applies, reads:

A proposed rate, classification, rule, or practice for transportation by a rail carrier . . . requiring a total capital investment of at least $1,000,000 to implement shall be established and become effective under this section. This section applies whether the investment is made individually or collectively by the carrier or by a shipper, receiver, or agent for any of them, or by a third party.

In previous western coal unit train cases initiated under § 10729, railroads have typically relied upon investments in additional locomotives and in construction and improvements on track and roadbed required by a movement to establish the $1,000,000 investment threshold. Incentive Rate on Coal-Cordero, Wyoming to Smithers Lake, Texas, 358 I.C.C. 537 (1977) ( "Smithers Lake"); Incentive Rate on Coal-Gallup, New Mexico to Cochise, Arizona, 357 I.C.C. 683 (1977) ( "Cochise"); Incentive Rate on Coal-Hayden, Colorado to Kings Mill, Texas, 359 I.C.C. 749 (1979) ( "Kings Mill"), petition for review pending; cf. Incentive Rate on Coal-Belle Ayr, Wyoming to Council Bluffs, Iowa, 359 I.C.C. 201, 207 (1978) ( "Council Bluffs"). The railroads here followed form, alleging that the Coleto Creek movement would require the purchase of 37 locomotives, the construction of a new connection between the Santa Fe and SP lines at Caldwell, Texas, extension of sidings and passing tracks along the Santa Fe line, and substantial upgrading of track of both the Santa Fe and SP. The required upgrading alleged by SP was on 92 miles of branch line, on much of which rail would have to be replaced, as well as ties under the rails and ballast under the ties, with stabilization of the subgrade under the ballast, in order safely to accommodate the heavy coal trains.

CPL argues that these investments do not qualify the railroads' proposed schedule for capital incentive treatment because (1) the service for which the schedule is proposed is neither innovative nor of a new type, (2) the proposed expenditures are not required for the CPL traffic, but at most amount to conveniences, (3) the expenditures claimed by SP for upgrading its track are actually operating expenditures rather than capital investments, and (4) the railroads refused to produce evidence requested by CPL in order to accurately determine the amount of claimed investments which could be properly attributed to CPL's traffic as opposed to the railroads' normal traffic.

The Commission did not, as it did in prior coal cases, include the alleged purchases of locomotives in its computation of required capital investments (Joint Appendix ("J.A.") III, p. 1939), but nevertheless found the $1,000,000 threshold requirement to have been exceeded by a "substantial margin." (J.A. III, p. 1940). The Commission specifically found that the combined investment by the Santa Fe and SP in constructing the interchange at Caldwell, Texas, and the expenditures by SP to upgrade its branch line track would each exceed $1,000,000. Ibid.

With respect to CPL's argument that the capital incentive rate provision applies only to innovative types of service, we note that there is no such requirement in the language of § 10729. CPL, though, perceives in the legislative history of § 10729 such a requirement. Although the District of Columbia Circuit in Houston Lighting Power Co. v. United States, 606 F.2d 1131 (D.C. Cir. 1979), cert. denied, 444 U.S. 1073, 100 S.Ct. 1019, 62 L.Ed.2d 755, extensively reviewed the legislative history of § 10729 and found no requirement that the service be innovative, CPL offers no reason to us why the District of Columbia Circuit was mistaken in its conclusion. After reading the legislative history, we agree with the District of Columbia Circuit that the service need not be innovative or of a new type for the reasons they enunciated. 606 F.2d at 1137-1139. In so holding, we follow our recent decision in Celanese Chemical Co., Inc. v. United States, 632 F.2d 568 (5th Cir. 1980).

The remainder of CPL's objections that the railroads failed to qualify for treatment under § 10729 are disposed of by focusing upon the interchange to be constructed between the Santa Fe and SP lines at Caldwell, Texas. CPL attacked this proposed interchange as unnecessary, constituting a convenience only in the movement of coal trains from the Santa Fe line to SP's line. Presently, the Santa Fe line passes over SP's line on an overpass at Caldwell, Texas. (J.A. II, p. 1098). There exists an interchange at this junction, but it is a little used connection neither constructed, nor maintained, for the engineering standards necessitated by coal train operation. ( (Ibid). Moreover, it is placed in such a way that the unit coal train proceeding off the Santa Fe track to the SP track would be headed in the wrong direction. To put the train in the right direction would necessitate blocking traffic on SP's intercontinental main track for two hours or more while the locations of lead locomotives, helper locomotives (located after the first 2/3 of the train), and the caboose are switched. A witness for CPL stated that a new connection at Caldwell, while a convenience, was not necessary. An SP engineer countered with testimony that it would be hazardous, impractical and inefficient to operate over the existing connection. (J.A. III, p. 1112). In the light of such testimony, the Commission's conclusion that the Caldwell interchange is required by CPL's coal traffic is reasonable. Moreover, there is clearly sufficient evidence in the record that the total cost of this interchange would exceed $1,000,000. (J.A. II, pp. 1080 and 1114). Since the Caldwell interchange alone suffices to qualify the proposed rate for capital incentive treatment, we do not address CPL's remaining arguments.

V. VOLUNTARY REMAND

The Commission, as noted above, has not filed a brief on the merits. On April 30, 1980, it voted to reopen this proceeding subject to approval by this court. On May 6, 1980, before the conclusion of the briefing schedule and before oral argument, the Commission moved this court to remand without consideration of the merits to enable the Commission to reconsider the present record. The Commission wishes to reconsider primarily the issue of market dominance, and especially its findings on each of the three presumptions set out in 49 C.F.R. 1109.1(g). It also wishes to reconsider whether the railroads submitted sufficient evidence of competitive alternatives to rebut any presumption of market dominance. If the Commission on reconsideration finds market dominance, it will need then to examine the reasonableness of the rate. The Commission also wishes to reconsider CPL's allegation of rate discrimination. Finally, the Commission requests that the remand not be limited to reconsideration of the above issues, but that we give the Commission authority to consider related issues which it may deem necessary to resolve in order to render an adequate decision in this proceeding.

As we read the Commission's motion, it approaches, for all practical concerns, a request for remand of this entire proceeding for reconsideration. CPL, Justice, Texas and the railroads all join in opposing the Commission's motion. They maintain that a voluntary remand is impractical now when the decision is ripe for appellate review. The railroads further argue that § 10729 of the Reform Act precludes such a voluntary remand. After a review of the Reform Act, we agree that the modifications added by that legislation to Commission procedure in rail cases preclude a voluntary remand in this case.

The Commission argues that it has authority under 49 U.S.C.A. § 10323(a) and under American Farm Lines v. Black Ball Freight Service, 397 U.S. 532, 90 S.Ct. 1288, 25 L.Ed.2d 547 (1970) and United States v. Benmar Transport Leasing Corp., 444 U.S. 4, 100 S.Ct. 16, 62 L.Ed.2d 5 (1979), to reconsider a decision even after the petitions for judicial review have been filed. We do not consider American Farm Lines or Benmar Transport authoritative in this situation since neither case deals with a rail carrier, and the Reform Act has established special provisions governing Commission procedure in rail cases.

We further conclude that the Commission is mistaken that § 10323(a) is the appropriate section governing reconsideration in rail cases. Instead, the appropriate section governing procedure in rail cases is § 10327, which was added to the Interstate Commerce Act by the Reform Act. Section 10327(a), clearly establishes § 10327's procedure in rail cases, reading in pertinent part:

(a) Notwithstanding sections 10322, 10323, and 10324(c) of this title, this section applies to a matter before the Interstate Commerce Commission involving a rail carrier providing transportation subject to the jurisdiction of the Commission under subchapter I of chapter 105 of this title.

Section 10327(g)(1), which deals specifically with reconsideration, states:

(g)(1) The Commission may, at any time on its own initiative because of material error, new evidence, or substantially changed circumstances —

(A) reopen a proceeding;

(B) grant rehearing, reargument, or reconsideration of an action of the Commission; and

(C) change an action of the Commission.

An interested party may petition to reopen and reconsider an action of the Commission under this paragraph under regulations of the Commission.

We note that this section limits Commission reconsideration on at its own initiative to those cases where there is "material error, new evidence, or substantially changed circumstances," a requirement lacking in § 10323(a) governing reconsideration of other cases. The Commission in requesting a remand for reconsideration has cited neither material error, new evidence, nor substantially changed circumstances in this case. Because it relies upon none of these circumstances, the Commission has no authority in this case to reconsider its decision on its own motion. Moreover, this limitation on Commission instigated reconsideration in rail cases is consistent with Congress' concern that regulatory delays were a major cause of our nation's railroads' difficulties. To allow the Commission to reconsider without citing any of these specified reasons would delay the determination of whether a rate may be published.

Our second reason for refusing a voluntary remand arises out of the fact that this is a capital incentive rate case brought under § 10729. We believe that the language and policy of § 10729, as well as the structure of the Reform Act, precludes the Commission from reopening a case on its own initiative after the 180-day time limit to rule on a capital incentive rate. The part of § 10729 pertinent to the problem of voluntary remand reads:

Once a rate, classification, rule, or practice becomes effective under this section, the Commission may not, for 5 years, suspend or set it aside as violating section 10701, 10726, 10741-10744, or 11103 of this title. However, the Commission may order the rate, classification, rule or practice to be revised to a level equal to the variable costs of providing the transportation when the Commission finds the level then in effect reduces the going concern value of the carrier.

(emphasis added). In this case, the Commission wishes to reconsider, among other things, whether the rate for CPL's traffic violates § 10741, prohibiting discriminatory rates; and if the Commission finds market dominance to exist on voluntary remand, it would have to determine whether the rate is reasonable under § 10701. Both provisions are included in § 10729's five year prohibition against Commission action.

The language of § 10729 sweeps broadly, placing no limitations on the preclusion imposed upon the Commission not to suspend an incentive rate for five years. It might be argued that § 10729 prohibits the Commission from reconsidering a rate only at the instigation of a shipper under § 10707, but does not preclude the Commission itself from reconsidering a rate. We do not perceive such a qualification in the language of § 10729. If anything, the sentence in § 10729 giving the Commission authority to reconsider a rate which reduces the going concern value of the carrier implies that in other situations the Commission lacks such authority to set a rate aside. We agree with the District of Columbia Circuit, when in a slightly different context, it stated:

In our view, what the statute means is that the Commission has no authority to provide agency-generated reconsideration of its approval of a rate — whether the approval is expressed positively, in an order, or passively, by failure to intercede.

Houston Lighting Power Co. v. United States, 606 F.2d at 1144. The Commission attempts to distinguish this language by arguing that a voluntary remand still requires an order by this court, and thus is not "agency-generated." We think this distinction is without merit and ignores the reality of the Commission's motion. A court would have no reason to remand without consideration of the merits absent a request by the Commission. In effect, the Commission, in making its motion, is asking that the court authorize it to do what it otherwise could not do.

There is nothing in the legislative history addressed to the particular question of whether the Commission could, on its own initiative, seek a voluntary remand in a rate incentive case. However, we do perceive in the legislative history a policy which would be violated by a voluntary remand. The Reform Act traces its roots back to the 93rd Congress. The House Committee on Interstate and Foreign Commerce in that Congress, after considering a bill sponsored by the Department of Transportation, reported to the House a bill, H.R. 5386, with a rate incentive provision essentially the same as that found in § 10729, the primary difference being that the House bill offered 3 years, instead of 5 years, protection to an incentive rate. In the debate on the floor of the House, Rep. Adams explained that the rate incentive provision was known as the "Big John" provision because it was aimed at ameliorating problems such as that encountered by the Southern Railway in its efforts to publish new rates for shipment of grain in innovative "Big John" cars it had developed. Rep. Adams stated this provision would:

give some assurance to a carrier or a shipper that if he invests a million dollars in a new service, he will receive a prompt decision one way or another on that rate from the ICC and that if a rate is permitted it will remain in effect for 3 years.

93rd Cong., 2d Sess., 120 Cong.Rec. 38,736 (1974). This bill passed only the House and was never enacted. In the 94th Congress, bills were again introduced in both the House and Senate to revitalize the nation's railroads. These bills, which were eventually combined and enacted as the Reform Act, again contained essentially the same rate incentive provision. The committee reports to both bills emphasized that they were meant to remedy the serious regulatory lag hampering rail rate cases. In pursuit of this goal, special incentive rate cases brought under § 10729, as well as regular rail rate cases brought under § 10707, are subject to time limits for consideration, both types of cases requiring prompt Commission action. Were we to permit this requested voluntary remand, the goal of prompt Commission resolution in rail cases would be frustrated. In effect, we would be giving the Commission additional time beyond the 180-day requirement of § 10729 in which to determine a capital incentive rate. Although we do not believe that the Commission in this case is using a voluntary remand as a ploy to circumvent § 10729's time limitations, it would be an obvious threat to the procedural integrity of § 10729 and the concerns of Congress if voluntary remands were allowed. Accordingly, we believe they are barred by § 10729.

VI. SCOPE OF COMMISSION'S AUTHORITY UPON REMAND

The railroads do not contest the jurisdiction of this court to review the Commission's decision in this case. However, the railroads argue that, after a court-generated remand, the Commission has no authority, because of the five year prohibition of § 10729, to reconsider, or revise its findings. The railroads admit that this limitation on the Commission's authority would as a practical matter make judicial review largely a pointless exercise. But they steadfastly argue that the Commission's power, even on court-generated remand, is limited under the language of § 10729 to changing the rate only after 5 years have passed.

The same argument was made to the Houston Lighting Power court:

In intervenors' [railroads'] view, this language [in § 10729 limiting Commission's authority to suspend a rate] deprives the Commission of authority to find a capital incentive rate unlawful and to set it aside even when a reviewing court has remanded the Commission's initial decision approving the rate because of defects under the Administrative Procedure Act.

606 F.2d at 1143. The court there rejected the argument and we agree. To have judicial review of an agency action with no authority in that agency to correct errors is a futile exercise. We do not perceive in the legislative history of § 10729, the "clear and convincing evidence" necessary to conclude that Congress intended to restrict normal judicial review. Dunlop v. Bachowski, 421 U.S. 560, 567, 95 S.Ct. 1851, 1857, 44 L.Ed.2d 377 (1975), Abbott Laboratories, Inc. v. Gardner, 387 U.S. 136, 141, 87 S.Ct. 1507, 1511, 18 L.Ed.2d 681 (1967). While Congress wished to give railroads assurances that if they made substantial investments, rates associated with these investments would remain undisturbed for 5 years, there is no hint it wished to protect erroneous Commission action. On remand, accordingly, the Commission shall have authority to take whatever action our opinion requires.

VII. PRESUMPTIONS

As noted above, the Commission has promulgated four rebuttable presumptions found at 49 C.F.R. 1101.1(f) and (g) to aid it in speedily determining whether market dominance exists. The Commission found that none of these presumptions applied to CPL's movement; we must review the Commission's action with respect to each.

A. Rate Bureau Presumption

CPL claimed before the Commission that one factor indicating market dominance was that the carriers considered the schedule within their rate bureau, the Southwestern Freight Bureau. Establishing that a rate has been discussed, considered, or approved upon a rate bureau agreement does not establish a rebuttable presumption of market dominance per se. Instead, it establishes a rebuttable presumption that a carrier participating in the rate or in such discussion or consideration does not provide effective competition to the proponent rail carrier for the involved traffic. 49 C.F.R. 1109.1(f). The evidence CPL placed in the record on this point is admittedly sparse, consisting only of a notice published on July 12, 1977, that the proposed rate had been docketed, inviting interested parties to comment and to request a public hearing, and a subsequent withdrawal of the rate from the public docket on September 26, 1979. (J.A. III, pp. 1622-3).

The Commission did not address CPL's allegation of rate bureau activity. The railroads defend the Commission's failure to address this issue by pointing to the language of the rate bureau presumption which provides that the presumption arises when the "evidence addressed establishes that the rate in issue has been discussed, considered, or approved upon a rate bureau agreement." 49 C.F.R. 1109.1(f). The railroads argue that docketing the proposed rate for independent action is not the same as discussion, consideration or approval.

We are not convinced that the rate bureau presumption should be read as narrowly as the railroads request. In Ex parte No. 320, Interim Report, the Commission stated:

Ordinarily, no presumption will attach when a proposal is docketed for independent action pursuant to the rate bureau procedures. We do note, however, that the presumption will apply when two or more carriers which have previously discussed or considered a conference rate later publish the same or similar rates at or about the same time, pursuant to their right of independent action. The circumstances underlying independent action are too varied to permit any uniform inference, and the circumstances surrounding such publication will be considered on a case-by-case basis.

353 I.C.C. at 891. The use of the adverb "ordinarily" implies that in certain circumstances docketing a proposal for independent action pursuant to rate bureau procedures will establish the rate bureau presumption. Also the Commission makes clear that the circumstances underlying independent action within a rate bureau are too varied to permit uniform inference. Significantly, in Ex parte No. 320, Final Report, in which the Commission promulgated the market dominance presumptions, the Commission added to § 1109.1(a) a requirement that a carrier furnish information concerning whether a rate has been docketed, as well as whether it has been discussed, considered or approved. Admittedly, in explaining the reason for this addition, the Commission stated that evidence of docketing procedure would enable it to better determine the existence of tacit collusion where two or more rates are published at or about the same time. Ex parte No. 320, Final Report, 355 I.C.C. at 14. But the Commission did not limit the relevance of a docketed rate to this situation.

We read the Commission's regulation together with its comments on the regulation as evidence of its concern that docketing a rate may have some relevance in determining whether the rate bureau presumption has been satisfied. We cannot ascertain the possible implications of docketing a rate with the particular rate bureau involved in this case as there is nothing in the record as to the procedure followed by the Southwestern Freight Rate Bureau. Although § 10706(a)(3)(A) provides that there be a final disposition of a rate docketed with a rail rate bureau by the 120th day after it is docketed, there is no indication as to whether such approval was obtained in this case or whether such approval is automatic in the rate bureau involved. While we defer to the expertise of the Commission to make the initial determination of what relevance docketing a rate may have in this case on market dominance, we conclude that in light of the Commission's own requirement for information concerning docketing of rates, and its statement that circumstances underlying independent action are too varied to permit uniform inference, the Commission should have addressed CPL's allegation. Cf. Pitre Brothers Transfer, Inc. v. United States, 580 F.2d 140 (5th Cir. 1978). On remand, the Commission shall do so.

B. Market Share Presumption

Under the Commission's market share presumption of market dominance, carriers are presumed to enjoy market dominance where they have handled at least 70% of the involved traffic during the preceding year. 49 C.F.R. § 1109.1(g)(1). Here no traffic has moved in the past from Axial to Coleto Creek, but when it commences, the railroads will have 100% of the market. The Commission found that the market share presumption was irrelevant in this case, noting:

In the movement under consideration here, no traffic has existed in the past. Although the utility company has used coal from sources in South Africa, it has not done this on a regular basis. Accordingly, the first presumptive test is not relevant to the situation presented here.

(J.A. III, p. 1942).

While the Commission is correct in literally reading the market share presumption to be inapplicable to a new movement such as the movement to Coleto Creek, the Commission has in the past weighed the application of the market share presumption in new movements of coal and has found that the "market share allegation does merit serious consideration." Kings Mill, 359 I.C.C. at 761. Kings Mill is very similar to this case, involving unit-train movements of coal in shipper-supplied equipment from a mine with which the shipper had contracted for large quantities over a long period of time. It is especially significant for the case at hand because the only stated reason the Commission gave for finding market dominance in Kings Mill was that the shipper's market share allegation had been demonstrated in light of the fact that a long-term commitment to a particular source had been made. Ibid.

This same reasoning was used in an even earlier case, Smithers Lake, again involving unit-train shipments of coal in shipper-supplied cars from a contract source. In Smithers Lake, the railroads expressly argued that the market share presumption could not apply since there had been no prior traffic between the origin and destination. 358 I.C.C. at 542. The Commission rejected this reasoning, stating:

[A]lthough the market share presumption is literally inapplicable to new movements, the fact that all of the subject traffic will be handled by the respondents is clearly an important factor. No evidence has been presented which would indicate that respondents will control less than 70 percent of the market from Cordero [the mine] to Smithers Lake [the generating plant] once actual movements commence.

358 I.C.C. at 555. This reasoning was a significant factor in Smithers Lake as to why the Commission found market dominance. Accordingly, on remand, the Commission shall either adhere to its prior decisions such as Kings Mill and Smithers Lake or explain its deviation.

C. Revenue/Cost Presumption

Under the revenue/cost presumption, market dominance is presumed to exist where the rate in issue exceeds the variable cost of providing the service by 60 percent or more. 49 C.F.R. § 1109.1(g)(3). The crucial concept is that of variable costs. All the petitioners allege one or more mistakes by the Commission in calculating the railroads' variable costs for CPL's traffic. Because this is such a technical subject, a review of the Commission's rules and practices for determining costs will be beneficial before discussion of the substance of the petitioners' objections.

In proposing the revenue/cost presumption in Ex parte No. 320, Interim Report, the Commission stated that variable costs are expenses which over a long-term period, fluctuate with the volume of traffic handled. 353 I.C.C. at 911. The Commission noted that such costs include operating expenses, rents, taxes, and an allowance for the cost (apparently to be calculated at the imputed interest level) of equity capital invested in transportation property, plus interest on borrowed capital invested in such property. 353 I.C.C. at 911, n. 44. The Commission has indicated that in establishing variable costs for the revenue/cost presumption, shippers with the responsibility of providing evidence of costs may rely upon a formula known as Rail Form A. Ex parte No. 320, Interim Report, 353 I.C.C. at 913. In justifying the 60% figure used in the presumption, the Commission itself relied upon Rail Form A variable costs. Ex parte No. 320, Clarification, 359 I.C.C. at 738.

Rail Form A is a formula by which various levels of carrier costs for a movement may be determined. It can be utilized both to determine the variable cost associated with a particular movement and to allocate a portion of constant costs to a movement. The variable cost incurred by a movement is calculated by Rail Form A on the basis of historical systemwide averages and the volume of the movement in question. It projects an increase in operating expense as well as an increase in fixed plant investment and equipment investment incurred by a movement. Fifty percent of this incremental fixed plant investment and one hundred percent of this incremental equipment investment is deemed variable by Rail Form A. Accordingly, the return on the capital (at the imputed interest level) justified by fifty percent of the incremental fixed plant investment and one hundred percent of the incremental equipment investment is deemed a variable cost and is included in the total variable cost derived by Rail Form A.

With respect to unit coal train cases, the Commission has consistently modified Rail Form A in calculating costs for the purposes of rate-making. I S, No. 9199, Unit Train Rates on Coal-Burlington Northern, Inc., Decided July 13, 1979 (unprinted) ( "BN-Iowa"); Arkansas Power Light Co. v. Burlington Northern, Inc., 361 I.C.C. 504 (1979) ( "Arkansas Power"), petition for review pending; San Antonio, Texas v. Burlington Northern, Inc., 361 I.C.C. 482 (1979) ( "San Antonio I"), vacated and remanded sub nom. San Antonio v. United States, 631 F.2d 831 (D.C. Cir. 1980) ( "San Antonio II"); Annual Volume Rates on Coal-Wyoming to Flint Creek, Arkansas, Docket No. 36970 and Southwestern Electric Power Company v. Burlington Northern, Inc., Docket No. 36980, combined, Decision Served May 25, 1979 (unprinted) ( "Flint Creek"); Kings Mill, supra; Smithers Lake, supra. This modification has taken the form of the use of additives, one for additional equipment and the other for incremental fixed plant investment necessitated by unit coal train movements. These additives are added to the figure derived by Rail Form A to give what the Commission considers to be a more accurate projection of costs. These additives are not based upon systemwide historical averages, but are based upon actual incremental investments necessitated by a movement. These additives calculate both an operating expense portion associated with a movement and an allowance for the cost of the capital investment necessitated by a movement. These costs are directly allocated to a given movement by means of the additive. (J.A. III, pp. 2050-1). Thus, for example, instead of relying on Rail Form A's averages, the additive takes the actual capital investment in fixed plant necessitated by a movement, calculates the cost of that capital investment and directly allocates this cost to the movement. (J.A. III, pp. 2050-1). In order to eliminate a double count on the elements included within the additives, the Commission purportedly removes from the Rail Form A calculation the incremental fixed plant and equipment investments associated with the movement under consideration.

The petitioners make several objections to the Commission's modification to Rail Form A, but primarily complain that the Commission's method of removing incremental fixed plant investment from Rail Form A is defective and that a double count of this element remains. Although CPL objected to the use of the additives, it calculated a cost figure adding the additives to its Rail Form A calculation to arrive at a variable cost figure of $12.12 per ton. (J.A. III, p. 1724). This gives a revenue/cost ratio of 172%, well above the threshold level of the presumption. The Commission, though, rejected CPL's estimates of some of the inputs into the Rail Form A formula and also rejected CPL's calculation of the incremental fixed plant investment additive to arrive at an estimate of variable costs of $13.90 per ton. This amount yields a revenue/cost ratio of 150%, a ratio insufficient to establish the presumption.

(i) Use of Additives

CPL and Justice argue that any use of additives violates the Commission's rules for determining variable costs in applying the revenue/cost presumption. They contend that both Rules to Govern Assembling Presenting Cost Evidence, 337 I.C.C. 298 (1970) ( "Rules for Costs") and Ex parte No. 320, Interim Report, mandate strict adherence to Rail Form A methodology without modification. We believe neither of these proceedings requires such rigid methodology in deriving costs.

Rules for Costs was a proceeding in which the Commission discussed methods for determining variable costs in various transportation modes. It explained and noted that parties could rely on the use of Rail Form A to determine variable costs for rail carriers. 337 I.C.C. at 324. The Commission in Rules for Costs, however, was careful to preserve flexibility in the formula to derive variable costs, indicating that circumstances can exist in which the usual methodology for determining costs might be inadequate. The Commission left open the possibility of adapting the usual procedure to a special situation. The Commission emphasized:

[T]his proceeding [ Rules for Costs] would not determine whether any particular costs are more valid in certain cases than others, and that the parties would still be free to employ other methods of estimating costs.

337 I.C.C. at 300.

In Ex parte No. 320, Interim Report, the Commission noted that while shippers could rely on Rail Form A, this formula "could be adjusted to fit the traffic or movement to which the rate in issue applies." 353 I.C.C. at 913. Thus, Ex parte No. 320, Interim Report, as does Rules for Costs, indicates that flexibility is the watchword in determining the costs of a movement.

We conclude that in Rules for Costs and Ex parte No. 320, Interim Report the Commission clearly indicated that Rail Form A could be modified when necessary to determine variable costs. Accordingly, CPL's and Justice's argument must fail. However, we do not endorse the use of the particular modifications used by the Commission in this case for the reasons stated below.

(ii) Rate of Return on Incremental Fixed Plant Investment Additive

The incremental fixed plant investment additive is calculated as an annuity spread over a series of equal future payments during the service life of the new investment. (J.A. III, p. 2090). The Commission here, as in prior unit coal train cases, calculated the incremental fixed plant investment additive at what is called the "revenue need level," utilizing a weighted average of the cost of equity capital as well as debt capital to determine the cost of capital associated with this additive. (J.A. III, pp. 1968, n. 8, and 2090-99). Thus the Commission figured the cost of equity capital at a level somewhere between the traditional return on equity, which includes a profit or incentive for risk of ownership element, and a level we have referred to as the imputed interest level, i. e., the level similar to the cost of debt capital. CPL and Texas object to the Commission's treatment of the return-on-equity factor in calculating this additive, but for different reasons. We think both objections are valid; we conclude that the Commission erred to the extent it used a return-on-equity factor in excess of the cost of debt capital.

CPL objects that the Commission's calculation of the return-on-equity factor in the additive at the "revenue need level" violates the Rules for Costs. While we have noted above that the Rules for Costs provides for flexibility in cost methodology, it clearly states that, in determining variable costs, the risk of ownership element in the return on equity is not appropriate, and that a return on equity is an appropriate element of variable costs only to the extent of imputed interest similar to that charged for borrowed capital:

Allowances for return on investment . . . to the extent that they are admittedly intended to serve as an encouragement and incentive for continued or risk-bearing ownership, are in the nature of so-called pure economic profits and should not be taken into account as an element falling within the restrictive construction given here to costs. This is not to say, however, that such an allowance for profit is not to be considered as a factor in ratemaking — it should if transportation is to remain under private ownership and control . . . However, this so-called pure profit should be clearly distinguished from opportunity costs. The latter does include a return on investment, simple, in the sense of, and as the equivalent to, cost of capital, but in no way connected with the risk or incentive of ownership. It is with respect to this limited effect to be given to return on investment as an element of cost that the examiner generally agrees . . . as to why equity capital invested in carrier facilities should be treated in the same manner as similar debt capital. Such equity capital is then entitled to receive imputed interest, similar to that charged for borrowed money . . . .

337 I.C.C. at 393. The Commission has repeatedly stated that an allowance for the traditional return on equity investment is not appropriate in determining the variable costs of a service. Flint Creek at pp. 20-21; BN-Iowa at p. 46. When it has utilized a traditional return on equity, it has been careful to emphasize that such an inclusion has been to calculate revenue need costs in the context of rate-making, and not in determination of variable costs. Flint Creek at pp. 20-21; BN-Iowa at p. 46. Because we believe this principle to be well-founded in Commission cases, we conclude that the Commission has acted unreasonably in deviating from this principle of determining variable cost, at least in the absence of full explanation of the reasons for deviating. In light of the Commission's prior norms, we hold that the Commission erred in calculating the cost of equity capital at the "revenue need level," rather than at the proper cost of debt capital level.

Texas complains that inclusion of a return-on-equity factor was improper in light of the reasoning the Commission used in establishing the 160% ratio in the revenue/cost presumption. Texas rightly notes that the traditional notion of return on equity capital constitutes profit. It notes that in justifying the 160% ratio, the Commission included a factor for profit in estimating the level that would trigger the revenue/cost presumption.

We assumed that any percentage test would have to account for both a railroad's constant costs and its variable costs. We first determined from the best available evidence that nationwide railroad fully allocated costs approximated 129 percent of variable costs. Subsequent publications show this as 127 percent at variable costs.

We then increased this figure to take into account all railroad expenses and a reasonable profit, as well as to provide a wide margin for error.

359 I.C.C. at 737. (footnote omitted). Since a profit factor was incorporated in determining the maximum appropriate spread between variable costs and revenue, we believe it imperative that the Commission adhere to the methodology presupposed by the revenue/cost presumption. That methodology requires the Commission to determine variable costs without incorporating a traditional return on equity in applying the revenue/cost presumption. We do not imply by this holding, however, that the Commission may not impute a return on equity capital similar to the return on debt capital in calculating variable costs for the presumption. We have noted above that Rules for Costs indicates that equity capital is entitled to receive imputed interest similar to that charged for borrowed money. 337 I.C.C. at 393. See also § 202, Staggers Rail Act of 1980.

We conclude, both for the reason asserted by CPL and the reason asserted by Texas, that the Commission, in calculating variable costs for purposes of the revenue/cost presumption, has acted unreasonably to the extent that it utilized a return-on-equity factor in excess of the cost of debt capital. Should the Commission on remand wish to continue utilizing a return-on-equity factor in excess of the cost of debt capital, it must explain why it is deviating from its well established norm announced in Rules for Costs. Accordingly, we vacate and remand.

(iii) Double Count

Because with any new movement Rail Form A normally projects an increment in fixed plant investment based on system averages, the inclusion of the incremental fixed plant investment additive would result in a double count unless appropriate adjustments are made to Rail Form A. A recurring argument before the Commission has been the proper methodology for eliminating this double count. See BN-Iowa, Arkansas Power, San Antonio I, Flint Creek, Kings Mill, Smithers Lake. Shippers have consistently maintained unsuccessfully before the Commission that its modifications to Rail Form A calculations are not adequate to alleviate this problem.

The Commission in this case admitted that the use of the additives resulted in a slight overstatement of costs to an unknown degree, but added that it believed the overstatement not to be significant because, adopting the methodology of the railroads, it had excluded the incremental fixed plant investment related exclusively to coal traffic from its Rail Form A calculations. It noted that while some "remaining system investment" is left in Rail Form A, such investment allocated to CPL is small and CPL would, in any event, utilize a portion of existing fixed plant. (J.A. III, pp. 1942 and 2050-51). This is essentially the same justification the Commission has given to its methodology in the rate-making context. BN-Iowa, Arkansas Power, San Antonio I, Flint Creek, Kings Mill, Smithers Lake.

We are unable to ascertain from the Commission's opinion whether the overstatement it admits results from a double count of incremental fixed plant investment or whether it results for some other reason. Whatever the case, we vacate and remand because we are unable to discern the reasoning of the Commission.

A reason sufficient unto itself for our decision is that the Commission admits that its method results in a slight overstatement of variable costs, meaning that the actual revenue/cost ratio is slightly higher than the 150% the Commission calculated. Although we appreciate the fact that it may be difficult or even impossible to precisely calculate the overstatement, we believe that it is possible to provide a better explanation than that now before us. We also hope that the balance of our discussion in this paragraph (iii) will provide guidance as to some of the specifics that can be clarified. Accordingly, the Commission on remand should explain the duplication (including a discussion of the items or categories that are duplicated, and a discussion of how appropriate costs are eliminated from Rail Form A to eliminate the duplication), why it believes the duplication to be insignificant, and why it believes the 160% ratio not to have been exceeded.

Our second reason for vacating and remanding on this point is that the Commission has failed to adequately address the serious arguments raised by CPL that the Commission's methodology results in a double count. To better understand these objections, we need to explain more fully how Rail Form A predicts incremental fixed plant investment, and how the Commission modified Rail Form A in this case.

We understand Rail Form A to predict incremental fixed plant investment in essentially the following manner. An average of incremental fixed plant investment per increase in volume of traffic is first established on the basis of historical systemwide data. Then, for a movement in question, a prediction of incremental fixed plant investment is derived by multiplying the projected increase in volume for that movement by the historical average incremental fixed plant investment per volume. Fifty percent of the total predicted incremental fixed plant investment is then deemed to be a variable cost.

Because of the Commission's brevity in its explanation of why it rejects CPL's double count objection, we are not certain that the following accurately describes the modifications to Rail Form A which the Commission adopted. As we understand it, the Commission made a distinction between (i) the incremental fixed plant investment that is occasioned solely by unit coal trains, such as heavier than normal rail, or the addition of ties and ballast necessitated to upgrade a line solely because of CPL's movement, and (ii) the incremental fixed plant investment necessitated by system traffic, including unit coal trains. (J.A. II, p. 1380). In other words, an attempt is made to segregate incremental fixed plant into that which would not occur but for the special characteristics of coal trains and that which would occur merely because of the increase in volume resulting from the coal train traffic considered as normal traffic. The latter incremental fixed plant investment is that which can be considered to be appropriate for normal traffic. This distinction is thought necessary by the railroads because unit coal trains require capital outlays heavier than the average capital outlay normally resulting from volume increase. By segregating the incremental fixed plant investment necessitated solely by CPL's movement, the railroads attempt to segregate this heavier than average increment and to attribute it directly to CPL's movement. Within the additive is included the incremental fixed plant investment occasioned solely by the unit coal train movement, directly calculated without use of Rail Form A, while the Rail Form A calculation would include that incremental investment attributable to the volume increase occasioned by normal traffic, including CPL's traffic. Because the additive includes incremental fixed plant investment which is greater than average, an adjustment is made to the factor normally used in Rail Form A determining the portion of incremental fixed plant investment which is deemed variable. In this case, the railroads reduced the percentage from 50% to 47.4%. (J.A. II, p. 1382). This reduction in the percentage of incremental fixed plant investment deemed variable by Rail Form A is the only modification to the Rail Form A formula specified in the record. It is presumably this modification the Commission relies upon when it states that the railroads removed the incremental fixed plant investment related exclusively to coal traffic from the Rail Form A calculation.

Having explained the adjustment to Rail Form A, we can more clearly state CPL's objection to the Commission's methodology. First, CPL objected below that in all the prior cases in which the Commission has modified Rail Form A methodology, it has failed to come to grips with the fact that Rail Form A predicts an increment in fixed plant investment for a specific movement on the basis of volume output. The significance of this is that the Rail Form A measurement of increment in fixed plant is proportional to the volume increase of a movement. CPL maintains that because Rail Form A averages include historical cases of heavier than average outlays, the Rail Form A calculation, by itself and without any additive, is sufficient to predict accurately the increment incurred, even in cases of greater than average investment, as in the instant coal traffic. The thrust of this objection is that because historical averages include cases of greater than average incremental fixed plant investment, Rail Form A's calculation without modification is the most meaningful and representative determination of true incremental fixed plant costs. While the Commission may have a ready answer to this argument, we believe it merits consideration and answering.

A second objection by CPL, and one more directly to the point of a double count, is that the distinction between incremental fixed plant occasioned solely by CPL's movement and incremental fixed plant occasioned by traffic increase, including CPL's traffic, is not a workable distinction to eliminate the double count and is a distinction which has not been made. It argues for example that there are no workable criteria to separate ties and ballast utilized on a line which are occasioned solely by a unit coal train and that which is occasioned by volume increase of normal traffic including CPL's lines. We are not as certain as CPL that this distinction is meaningless or can never be made to work, but we perceive in the Commission's opinion no reasons as to why it believes the distinction utilized by the railroads, and which it adopted without comment, is appropriate. In adopting the railroads' methodology without addressing this concern of CPL, the Commission has made it impossible for us to adequately review its findings.

We are also concerned that the distinction, if meaningful, was not applied in this case to eliminate a double count. This can be seen, for example, in the handling of the increment along a portion of SP's line from Flatonia to Coleto Creek. Within the additive was included all the ties and ballast necessary to upgrade this line because this line was apparently scheduled for abandonment before CPL's movement. (J.A. II, pp. 1357 and 1415 and J.A. III, pp. 1717 and 1973). Since all of the cost of the ties and ballast was included in the additive, it would seem that a proper modification of Rail Form A would exclude all the incremental fixed plant investment relating to ties and ballast attributable to the volume increase along SP's line from Flatonia to Coleto Creek. We are unable to determine whether the modification of Rail Form A approved by the Commission accomplished this. On remand, the Commission shall address this point, as well as a similar discussion of other costs included in the additive and a demonstration that Rail Form A has been modified to eliminate any double count with respect to such costs.

In Celanese Chemical Co., Inc. v. United States, supra, we recently quoted from San Antonio II and held that the Commission's procedure — adding a fixed plant investment additive to the Rail Form A costs — constituted a double count and, therefore, was erroneous. We follow Celanese and hold that the Commission on remand shall be required to eliminate any double count. We do not require at this time that the Commission abandon its use of additives. It very well may be that additives are necessary to calculate accurate estimates of variable costs with respect to unit coal trains. Instead, on remand, the Commission should explain more fully its methodology, address CPL's objections so that we may rationally assess the propriety of its method, eliminate the double count, and demonstrate how the elimination has been accomplished.

D. Substantial Investment Presumption

Under the substantial investment test, presumption of market dominance arises when a shipper makes a substantial investment in rail-related equipment or facilities which prevents or makes impractical the use of another carrier or mode. 49 C.F.R. § 1109.1(g)(3). The Commission here found that CPL has made an investment of $15,500,000 in rail-related equipment. The Commission nevertheless found this presumption inapplicable. The Commission found that: (1) the investment did not preclude, economically or otherwise, the use of competing rail carriers, (2) CPL was not precluded from using its investment to transport coal from other domestic sources, and (3) CPL was not precluded from using rail links to Texas seaports, where CPL was obtaining coal from South Africa for its shakedown operations. The Commission supported these findings with only the general remark that the record demonstrates the existence of considerable competition to supply both the coal and the transportation of coal without specifying where in the record such support occurs. Because we fail to find support in the record for several of the Commission's findings, we vacate these portions of its decision. With respect to other findings, we vacate and remand in order that the Commission may more adequately explain its findings. In paragraph (i) below, we discuss the significance of the fact that Coleto Creek is served by only one terminating rail carrier, SP. Then, in paragraph (ii), we discuss the finding of competing rail lines for the Axial-Coleto Creek movement. In paragraph (iii), the Commission's finding of alternative domestic sources is considered. Finally, in paragraph (iv), we look at the finding of the existence of alternative transportation modes in the form of a sea-rail link to the Coleto Creek facility.

(i) Significance of one terminating carrier.

Our first reason for vacating these findings of the Commission is a general one relating to the interpretation of the presumption and applies across the board to all three findings of the Commission. It is conceded that SP is the only rail line which can serve the Coleto Creek facility for any movement of coal by rail, and that CPL would have to deal with SP if it is to use the equipment in which it has invested. The fact that SP has a monopoly position in the termination of any coal movement raises a question in our mind as to whether CPL can practically look to other carriers in the initial transportation of coal. The language of the presumption states that market dominance is presumed if the investment makes impractical the use of "another carrier," using the singular form. We do not perceive in the Commission's analysis an articulated reason as to why the language of the presumption permits a finding that "another carrier" may be used when the shipper is locked into a terminating carrier. If SP uses its monopoly position to charge monopoly prices for its portion of any movement, then it very well may be that any other movement using rail would be economically unfeasible. This, of course, may depend partly upon the length of the movement by SP which would be necessitated in using other originating carriers and the bargaining power of SP in setting the rate for the movement. On remand the Commission should explain why the monopoly position of SP does not establish this presumption, given CPL's investments.

(ii) Finding of competing carriers for the Axial to Coleto Creek movement.

In finding that CPL was not precluded by its investment from using competing rail carriers, the Commission made a finding not supported by the evidence. As noted above, SP is the only rail line capable of serving the Coleto Creek facility. Similarly, DRGW is the only line capable of serving the mine at Axial. Accordingly, any competition of competing rail carriers would be found in the segment served by Santa Fe, the bridge portion. We are not sure there is record evidence to support a conclusion that there existed competing rail lines for this segment. Most of the evidence concerning competition consists of cursory descriptions of alleged requests for quotation of rail rates and some quotations, all occurring before CPL entered its Colowyo contract. (J.A. II, pp. 798, 805, 823 5, 840-1, 857-862, 882-885). This evidence relates to transportation by the railroads here involved, and therefore cannot support a finding that other rail carriers were available competition for the bridge portion of this movement. However, because there is some suggestion of the availability of competition for the bridge portion, we decline at this time to hold flatly that the railroads have failed to carry their burden of showing competing rail service. On remand, the Commission, if it again finds that competing rail carriers are available with respect to the bridge portion of the movement, shall make specific findings which will enable this court to review the sufficiency thereof. Finally, if the Commission on remand does find competition with respect to the bridge portion, it shall demonstrate why such competition is sufficient to rebut the presumption in light of the undisputed fact that there is competition neither on the originating leg, which is served only by DRGW, nor on the terminating leg, which is served only by SP.

(iii) Finding of alternative domestic sources.

The Commission also found that the investment by CPL did not bar it from turning to alternative domestic sources. This is an allusion to the availability of geographic competition, an issue we discuss more fully below. The Commission's thinking in looking to alternative domestic sources is that if such sources are available, then a shipper may utilize his rail related equipment in purchasing from these alternative sources, and thus not be locked into one carrier. We vacate this portion of the Commission's decision because, as with its finding of competing rail carriers, its finding is unsupported by evidence in the record. The Commission failed to specify in its decision what alternative sources are available. The only evidence in the record of alternative domestic sources is a brief survey of requests for quotation of rail rates, and some actual quotes, from places other than Axial, all before CPL entered the Colowyo contract. (J.A. II, pp. 798, 823-5, 840-1, 857-62, 882-5). Such evidence may provide some inference, albeit weak, that CPL was considering coal from such other places. This inference is bolstered somewhat by other evidence, primarily railroad witnesses' assertions that CPL acknowledged contacting other mines, suggesting that CPL was considering other sources of coal. (J.A. II, pp. 798, 859-60, 882-5). However, there is no indication of the extent of negotiation for other coal. There is no indication that these sources offered coal suitable to CPL's needs, in the quantities required, or otherwise competitive with the Colowyo contract. Nor is there an indication that these alternative sources remained available after CPL had made its investment in rail-related equipment. There was also an assertion by a railroad witness that CPL had a 30% interest in a lignite field and that the Coleto Creek unit was capable of utilizing lignite. (J.A. II, pp. 882-5). There is nothing in the record, though, to indicate that this lignite field had reserves, committed or uncommitted, sufficient to serve CPL's needs or any other information concerning the availability or suitability of the lignite. Given the large amounts of coal required and the unique specifications required, we do not believe it is rational for the Commission to assume on the basis of such thin evidence that alternative sources are available. We conclude that the evidence is insufficient to support a finding of the availability of alternative domestic sources, and therefore the railroads have failed to carry their burden with respect to this particular attempted rebuttal of the presumption.

(iv) Finding of alternative modes.

Finally, the Commission found that CPL's investment did not preclude it from utilizing a rail link to Texas seaports, thus allowing alternative modes of transportation. The Commission did not specify what in the record supported this finding although it did note that in its shakedown operation, CPL was using South African coal being imported through the Corpus Christi port. Insofar as the Commission grounded its finding of the availability of a water mode alternative on CPL's shakedown operation, this was not an adequate basis on which to find the sea-rail mode feasible. The record clearly indicates that CPL was using trucks, and not rail, to transport the coal from the harbor to its plant. Use of a truck-sea mode would require abandonment of CPL's large investment in rail cars. Thus the shakedown operation, i. e., the use of a water-truck mode, does not provide an alternative use for CPL's rail car investment, and therefore does not rebut the substantial investment presumption.

The railroads argue that CPL has the burden of proving that a shift to a different mode of transportation would result in a substantial economic loss. They argue that CPL has the burden of proving that it could not lease or resell its rail cars to avoid economic loss. We reject the railroads' argument. Ex parte No. 320, Interim Report indicates that proof that a substantial investment will not result in economic loss is part of the railroads' burden of rebutting the presumption. In the rule making proceedings before the Commission, the railroads argued that a substantial investment presumption was not appropriate because specialized rail equipment was readily marketable. The Commission rejected that argument, and adopted the substantial investment presumption. In Ex parte No. 320, Interim Report, the Commission stated:

Shippers must be able to make the choice to use an alternative service without absorbing substantial economic loss. The greater the cost of making a shift in carriers, the greater the chance that the carrier will be in a position to extract substantial premiums without fear of diverting traffic to other carriers.

353 I.C.C. at 916. The Commission further stated:

It is true that at some point the shipper may attempt to minimize his losses by selling his equipment, but such a change in operations usually cannot be accomplished without a substantial loss to the shipper.

Ibid. Finally, the Commission, in noting what rebuttal evidence a carrier may produce, remarked:

The carrier could also show that the shipper has not made such a substantial investment that it is effectively dependent on rail transport.

353 I.C.C. at 917.

There is other evidence in the record, however, upon which the Commission may have been relying in finding the water-mode alternative to exist. A witness for SP indicated that it had quoted a price to CPL on moving coal from the Corpus Christi port to the generating plant, and was prepared to provide a rail-link if a suitable price could be negotiated. (J.A. II, pp. 865-866) Insofar as the Commission was relying upon this evidence, we must vacate and remand because the Commission failed to address an obvious problem with using a rail link with Texas ports in light of CPL's purchase of 347 cars. This purchase was for a movement from Axial, Colorado, covering an approximately 2,600 — mile round trip, while the round trip distance to and from the Corpus Christi port is approximately 176 miles. Clearly, a rail connection with the Corpus Christi port would require far fewer cars. Yet, the Commission did not determine how many cars would be needed with the rail-seaport mode or how the use of such a mode could avoid substantial economic loss resulting from a surplus of unused rail cars. On remand, the Commission shall explain why it believes reliance by CPL on a sea-rail mode would not require the mothballing or resale of substantial numbers of the cars purchased by CPL, and whether a sufficient number of rail cars would have to be abandoned to constitute the abandonment of a substantial investment, sufficient to trigger the presumption.

VIII. GEOGRAPHIC COMPETITION

The Commission found that even if CPL and Texas had established the revenue/cost presumption of market dominance, the railroads had rebutted the presumption by specific evidence of competitive alternatives. It concluded that both before and after CPL had entered its contract with Colowyo, geographic competitive alternatives existed. The Commission found that with respect to the period before the Colowyo contract, CPL had its choice of numerous sources of coal, carriers and varying modes of carriers, without specifying any of these alternative sources or carriers. In an apparent attempt to blunt the effect of CPL's long-term contract with Colowyo, the Commission reasoned that CPL had retained considerable flexibility under the terms of the contract, especially flexibility resulting from the assignment clause of the contract. The Commission did not specify what flexibility, other than possible assignment, there was in the contract. It made no findings as to the practicality or likelihood of CPL finding an assignee, nor did it discuss the significance of the provision within the assignment clause that CPL would remain liable after any assignment. Nowhere in the opinion did the Commission indicate what alternative sources of coal would now be available should CPL assign its contract.

Three objections are raised to the Commission's consideration of geographic competition to rebut any showing of market dominance through the presumptions. First, CPL and Texas argue that the language of § 10709(a) precludes the Commission from considering geographic competition and limits it to consideration of competition in the form of other rail carriers from Axial, Colorado, to Coleto Creek, Texas, or other modes of transportation between those two points. We discuss this argument in Subpart A below. Second, CPL, but not Texas, maintains that even if the Commission could, under the language of § 10709 consider geographic competition, it is precluded from doing so under its decision in Ex parte No. 320, Interim Report and other prior cases, defining the criteria for establishing market dominance. We discuss this argument in Subpart B. Finally, CPL, Texas and Justice all contend that even if the Commission could consider geographic competition under § 10709 and Ex parte No. 320, Interim Report, its finding in this case that geographic competition exists was not based on a reasoned analysis of all the relevant factors and was contrary to its prior approach in factually similar cases. We discuss this argument in Subparts C and D.

A. Statutory Restriction on Consideration of Geographic Competition.

We decline at this time to address the question of whether a reasonable interpretation of "market dominance" as used in § 10709 permits the consideration of geographic competition. As we note immediately below, the Commission at one time clearly construed the statute as barring consideration of geographic competition, but has attempted to abandon this interpretation. Our reasons for making no decision at this time are set out more fully in the immediately following subpart, but essentially involve the fact that we do not have the benefit of Commission thinking as to why the language and history of § 10709 do not preclude consideration of competition other than that between the point of origin and destination.

B. Restriction on Consideration of Geographic Competition Within Regulations.

CPL argues that even if the language of § 10709 permits the Commission to consider geographic competition, the Commission's holding in Ex parte No. 320, Interim Report precludes such evidence.

In proposing the market share presumption, the Commission engaged in a lengthy discussion of the relevant market. The Justice Department, the Department of Transportation, and the railroads proposed a broad market definition, including competition of substitute products and geographic competition. Ex parte No. 320, Interim Report, 353 I.C.C. at 900. The Department of Transportation in particular urged that geographic competition be used to rebut any presumption arising under the market share test. 353 I.C.C. at 901.

Although its discussion occurred in its consideration of the market share presumption, the Commission clearly rejected the proposals of the Justice Department, the Department of Transportation, and the railroads on the broad basis of the language of § 202(b) of the Reform Act, as well as the policy behind the statute. The Commission held that § 202(b) of the Reform Act precluded the Commission from looking to geographic competition. The Commission was clearly interpreting the broad, statutory concept of "market dominance" as defined in § 202(b), and not just the definition of "market" as used in the market share presumption. In issuing Ex parte No. 320, Final Report, adopting the present regulations, the Commission in no way modified its interpretation of "market dominance" as used in the statute. This is the statutory interpretation which CPL urges us to adopt as the only reasonable interpretation of § 202(b); this is the interpretation which the District of Columbia Circuit upheld in Atchison, Topeka Santa Fe Railway Co. v. ICC, 580 F.2d at 634.

Since the issuance of Ex parte No. 320, Final Report, the Commission has apparently reversed itself, indicating in very brief language that evidence of geographic competition is relevant in determining market dominance. In its clarification of the revenue/cost presumption required by the remand in Atchison, Topeka Santa Fe Railway Co. v. ICC, supra, the Commission in a footnote stated that there had been a misunderstanding of the Commission's willingness to consider all evidence relevant in determining market dominance, noting:

While certain evidence is not germane to computing market share, proponents of a rate may introduce evidence of potential competition, competition from private carriage, alternative product competition or geographic competition to show that effective competition exists.

359 I.C.C. at 736, n. 7. In a case decided at approximately the time this case was filed, the Commission again noted without explanation that geographic competition is relevant in rebutting any presumption. BN-Iowa at 28. In a case decided after the record was closed in this case, the Commission noted that it had requested the parties to supply information concerning substitute coal origins. Increased Rates on Coal, Midwest Railroads August, 1979, Docket No. 37246, Served Nov. 15, 1979 (unprinted). Finally, in this case the Commission stated that its prior remarks prohibiting the consideration of geographic competition were limited to the definition of "market" as used in the market share presumption.

These statements by the Commission indicate its desire to alter its prior statutory interpretation of "market dominance" found in Ex parte No. 320, Interim Report. However, the Commission has not given reasons why it believes its initial interpretation of § 202(b) of the Reform Act is incorrect. When an agency desires to depart from its prior norms or from a prior statutory interpretation, it must clearly set forth its reasons. Atchison, Topeka Santa Fe Railway Co. v. Wichita Board of Trade, 412 U.S. 800, 808, 93 S.Ct. 2367, 2375, 37 L.Ed.2d 350 (1978); NLRB v. Sunnyland Packing Co., 557 F.2d 1157 (5th Cir. 1977). The Commission at one time set forth extensive reasons for its interpretation that § 202 of the Reform Act precluded consideration of geographic competition, and cannot now reject that well-developed reasoning without explanation.

Two additional problems immediately come to mind as a result of this unexplained change. First, were we to uphold the Commission's consideration of geographic competition in this case, we would in effect be approving a new interpretation of "market dominance" as having a reasonable basis in law without the benefit of the Commission's thoughts on this matter. In the normal course of events, an administrative agency pronounces its understanding of a statute, which in turn is afforded deference by the reviewing court. On this appeal we decline to consider whether the language or history of § 10709 permits consideration of geographic competition. We prefer to have the benefit of the Commission's considered analysis of its prior reasoning in Ex parte No. 320, Interim Report and the reasons, if any, for revising that interpretation. Since we are remanding this case for reconsideration of other issues, obtaining the Commission's reasoning on this matter will not delay this proceeding.

Second, this court is ill-equipped to address in the first instance one aspect involved in the determination of whether § 202(b) of the Reform Act permits consideration of geographic competition. One of the requirements imposed by § 202 of the Reform Act was that the Commission establish rules designed to provide for "a practical determination [of market dominance] without administrative delay." Thus, in determining the scope of "market dominance," as used in the statute, a relevant factor is the rapidity with which the Commission can determine the existence of market dominance. The Commission in Ex parte No. 320, Interim Report noted that consideration of geographic competition and other forms of competition would require lengthy antitrust-type litigation and would perhaps frustrate its ability to make market dominance determinations promptly. 353 I.C.C. at 905. We perceive nothing in the Commission's later holdings — that geographic competition is relevant in determining market dominance — explaining why it now believes such evidence would not frustrate the statutory mandate for quick determination. Not being involved in the agency stages of rate litigation, this court is an inappropriate forum to determine in the first instance and without the benefit of the Commission's thinking, whether the quick determination requirement can be met if geographic competition is considered.

C. Commission's Finding of Geographic Competition After the Colowyo Contract.

Because we have not precluded the Commission on remand from adopting, subject of course to judicial review on subsequent appeal, an interpretation of market dominance including considerations of geographic competition, we must address the Commission's specific finding that the railroads had produced evidence of geographic competition sufficient to rebut the presumption. A threshold deficiency in the Commission's approach is its failure to cite or distinguish its past cases involving attempts to utilize geographic competition when a shipper had entered into a long-term contract. In every case where parties have asked the Commission to find geographic competition because of alternative sources available to a shipper before entering a contract, the Commission has ruled that such alternative sources are irrelevant once a shipper is committed to one source. Kings Mill, 359 I.C.C. at 761; Smithers Lake, 358 I.C.C. at 543 and 555; cf. Flint Creek, at p. 9 ("Before the shipper has signed a contract with a mine, it may still have bargaining power."); cf. Council Bluffs, 359 I.C.C. at 206 ("Once all these commitments were made and all effective competition was eliminated . . . ."). Smithers Lake makes clear that alternatives available before entering a contract are not relevant in determining effective competition after a contract has been entered when it states:

We reject respondents' contention that the past competitive circumstances surrounding other possible sources of coal should be considered. While the decision to ship from the Jacob's Ranch Mine may well have been arrived at in a competitive atmosphere, once protestant made that commitment and agreed to a supply contract it could not change origins. This being the case, `market competition' provides little protection to the shipper if the railroads were to attempt to exact an unreasonably high rate.

358 I.C.C. at 555. These cases establish as a general principle the Commission's recognition that a long-term contract can effectively lock in a shipper and render competitive alternatives ineffective. On remand, the Commission shall consider the foregoing cases, either follow them, demonstrate why they are not apposite, or explain why it is changing the rule announced therein.

It is possible that the Commission was attempting to distinguish sub silentio the above line of cases by finding that the railroads had made and supported several allegations of "flexibility" in the Colowyo contract. Because the Commission did not indicate which one or more of the railroads' several allegations of "flexibility" were found by the Commission to have support in the record, we discuss the several allegations seriatim.

First, the railroads asserted below, and the Commission apparently found, that geographic competition was available because the Colowyo contract is assignable. (J.A. II, pp. 896-897). We find no evidence in the record to support a finding that CPL could readily find an assignee for the contract or that assignment is a realistic possibility. The Commission apparently looked only to the fact that the contract had an assignment clause. We believe that it is arbitrary to find geographic competition solely on the basis of the assignability of the contract, without any evidence that assignment is a realistic alternative. Moreover, to do so flies in the face of the Commission's previous decision in BN-Iowa. There, the shipper, a public utility which was adding a fourth generating unit, entered into a long-term contract for coal. Although the contract had several option clauses and termination clauses affecting its length, the Commission assumed at the time of its decision that the projected length of the contract was 5 years. The assignment clause was much the same as in the contract before us. The Commission concluded in BN-Iowa that geographic competition did not exist despite the assignment clause. Accordingly, we vacate this aspect of the Commission's decision.

Second, the railroads below asserted that the contract was "flexible" because only 77% of CPL's needs for coal, representing a delivery of approximately 1.35 million tons per year, would be supplied pursuant to the Colowyo contract, and because of a clause in the contract allowing a variation of delivery of 10% each year. The railroad witness also noted that in 1990-1994, delivery would drop to 900,000 tons per year. (J.A. II, pp. 892-894). We do not believe these facts provide any evidence of "flexibility." CPL is obligated under the contract to take delivery of substantial quantities of coal each year. That coal must be moved from Axial, Colorado, to Coleto Creek, Texas. Accordingly, the Commission need not consider on remand the issue of "flexibility" of the contract because only 77% of CPL's needs are to be supplied.

Third, the railroads below asserted that the Colowyo contract can be terminated if the coal fails to meet certain specifications. (J.A. II, p. 898). There is no evidence in the record to suggest the likelihood that the coal will not in fact meet the contract specifications. Accordingly, we vacate the Commission's finding in this regard.

Fourth, the railroads below asserted that the contract was "flexible" because CPL could resell the coal. The argument is that, if CPL could easily resell the coal, it would then be free to contract for coal from other sources, thus introducing geographic competition. The evidence does include the testimony of a railroad witness to the effect that the Colowyo coal is "desirable coal" and by western coal standards is of "high quality." (J.A. II, pp. 897-8). The witness testified that the high quality of this coal, combined with CPL's access to South African coal, made resale a reasonable possibility. Ibid. We have considerable doubt that the witness' testimony alone could support a finding of flexibility; but we do not have to determine this question at this time. As noted, CPL did present a witness who rebutted the railroads' testimony. This witness testified as to the implausibility of resale, noting that the only resale which would do CPL any good would be of all the coal, a possibility the witness stated was unlikely. (J.A. II, pp. 1925-1926). On remand, if the Commission reaches the rebuttal stage of consideration of geographic competition, it should explain why the evidence concerning resale supports a finding of flexibility, and should do so in light of the evidence presented by CPL.

D. Availability of Geographic Competition Before the Colowyo Contract.

The Commission also found that competitive alternatives were available to CPL before it entered into the Colowyo contract. The only evidence in the record concerning the existence of alternative sources before the Colowyo contract is the brief summaries by railroad witnesses implying possible contacts CPL may have had with other domestic coal mines as it began its search for a reliable and suitable source, as well as railroad witnesses' assertions that CPL acknowledged contacts with other mines. (J.A. II, pp. 798 and 859-60). As we noted above in our discussion of the substantial investment presumption, there is no indication that these other mines had coal in the quantities or quality necessary for CPL's Coleto Creek plant. There is no indication as to how serious negotiations were with these other mines or how far they progressed. As such, we believe this evidence is insufficient to support the Commission's findings of the existence of geographic competition before the Colowyo contract, and those findings are accordingly vacated.

IX. SECTION 10741 ATTACK ON RATE AS DISCRIMINATORY; AND RAILROADS' DUTY TO PRODUCE EVIDENCE

CPL alleged that the railroads' proposed tariff would result in undue and unlawful discrimination in violation of § 10741. In support of its allegation, it presented substantial evidence. (J.A. I, p. 392). Although a finding of market dominance is not a prerequisite to Commission authority to determine whether a rate is discriminatory, the Commission failed to address CPL's contention even in a perfunctory manner. We believe the Commission owes a duty to CPL and Texas to address this argument and should do so on remand. Pitre Brothers Transfer, Inc. v. United States, 580 F.2d 140 (5th Cir. 1978).

CPL also complained strongly to the Commission that despite a request for evidence at the initiation of this proceeding, the railroads had failed to supply CPL sufficient evidence concerning the costs of its movement. (J.A. III, pp. 1501-1505). Such evidence would be relevant in determining the applicability of the revenue/cost presumption. The Commission has held that although shippers have the burden of proof in capital incentive rate cases, the burden of producing evidence remains with the party best able to present particular evidence. Ex parte No. 327, Rate Incentives for Capital Investment, 353 I.C.C. 754, 766 (1977); Ex parte No. 320, Final Report, 355 I.C.C. at 15-16. Because cost data would normally be in the hands of the railroads, they normally would have the duty to produce such information. This duty assumes great importance given the time constraints of a capital incentive case. Although CPL made a clear objection with respect to this matter, the Commission did not deal with this complaint. It shall do so on remand. Pitre Brothers Transfer, Inc. v. United States, supra.

CONCLUSION

A brief summary of our holdings will make clearer the task remaining on remand. The Commission's finding that the railroads have made a capital investment of over $1,000,000 necessitated by CPL's movement is amply supported by the record. Accordingly, we hold in Part IV that the rate established by the railroads or set by the Commission is entitled to the capital rate incentive treatment under § 10729. In Part V we reject the voluntary remand requested by the Commission, and accordingly proceed to address the merits of this appeal. In Part VI we reject the railroads' contention that § 10729 restricts the Commission's authority on remand to review and revise its opinion in accordance with our remand instructions. We hold that the Commission must readdress the four presumptions for market dominance in light of our opinion. The rate bureau presumption shall be reconsidered pursuant to Part VII.A. of this opinion. The market share presumption shall be reconsidered pursuant to Part VII.B. of this opinion. With respect to the revenue/cost presumption discussed in Part VII.C., we hold in paragraph (i) that the Commission did not err in using additives to Rail Form A; in paragraph (ii) we hold that the Commission, in light of its prior norms, erred in incorporating into its incremental fixed plant investment additive a return-on-equity factor which was not limited to the cost of debt capital; and in paragraph (iii) we remand for reconsideration of the double count issues. With respect to the substantial investment presumption discussed in Part VII.D., in paragraph (i) we remand for reconsideration by the Commission of the significance of SP's monopoly position with respect to rail service to the CPL plant; in paragraph (ii) we remand for reconsideration by the Commission of its apparent finding that there were competing rail carriers for the Axial to Coleto Creek movement; in paragraph (iii) we hold that there is not sufficient evidence in the record to support the Commission's finding that alternative domestic sources of coal were available to CPL; and in paragraph (iv) we remand for reconsideration by the Commission of its finding relating to the seaport-rail mode as an alternative mode of transportation rebutting the substantial investment presumption. With respect to the Commission's finding that any presumption had been rebutted by the railroads' evidence of geographic competition, discussed in Part VIII of our opinion, in Subpart A, we remand for the Commission to address in the first instance CPL's argument that the statutory definition of market dominance precludes consideration of geographic competition; in Subpart B we remand for the Commission to reconsider its conclusions concerning geographic competition in light of the Commission's prior decisions; in Subpart C we remand for the Commission to reconsider the possible lock-in effect of the Colowyo contract, again in light of prior Commission decisions, and we also reject three of the four possible grounds for the Commission's finding that the Colowyo contract is "flexible" and remand for reconsideration of the fourth ground; in Subpart D, we hold that there is not sufficient evidence in the record to support the Commission's finding that alternative competitive sources of coal were available to CPL before it entered into the Colowyo contract. In Part IX we remand for the Commission to address CPL's assertion that the rate is discriminatory, and to address CPL's request that the railroads provide certain cost information.

In no instance have we either directed that the Commission take additional evidence or refrain from doing so. We leave those matters to its discretion. F.C.C. v. Pottsville Broadcasting Co., 309 U.S. 134, 60 S.Ct. 437, 84 L.Ed. 656 (1940). Of course, the Commission may take into account the delay required for taking such evidence, the need for expedited decision, the adequacy of the opportunities previously afforded a party who wishes to adduce additional evidence and any other factors it may consider appropriate in determining, on an issue-by-issue basis, whether to rely upon the present record or to require or permit additional evidence.

Recognizing that this is a capital incentive rate case, the Commission suggested a 90-day time limitation in its request for voluntary remand. Because of the complexity of this case, however, we decline to impose a strict time limitation, but rather we urge the Commission to afford this case the highest possible priority to the end that it might be disposed of expeditiously as intended by Congress.

AFFIRMED IN PART, REVERSED IN PART, VACATED IN PART, AND REMANDED.

. . . . .

(4) Whether and to what extent the rate in issue has been docketed, discussed, considered, or approved before a rate bureau acting under an agreement filed with and approved by the Commission pursuant to section 5a or 5b of the Interstate Commerce Act: application number(s), date proposal docketed with rate bureau, and final disposition of proposal and date thereof; and the share of the market, or an estimate thereof, presently held by such participating carriers.