PATRICK E. HIGGINBOTHAM, Circuit Judge:
The Federal Energy Regulatory Commission found zone rates on a major gas pipeline to be unjust, unreasonable, and discriminatory, and, pursuant to the Natural Gas Act, the Commission ordered a change in transmission cost allocation. Pipeline customers challenge the FERC’s directive as well as the Commission’s determination of the date on which the change is to be effective. We reject the challenges and affirm.
I
Transco operates a major unidirectional natural gas pipeline that starts from its supply sources in Texas and Louisiana and ends almost 2,000 miles away near New York City. The pipeline’s service territory is divided into three zones. Zone 1 extends [1551]*1551about 522 miles and includes Mississippi, Alabama, and Georgia. Zone 2 extends 455 miles and includes the Carolinas, Virginia, and the District of Columbia. Finally, Zone 3 extends 227 miles and includes Maryland, Delaware, Pennsylvania, New Jersey, and New York.
Transco and its customers reached an agreement in 1962 that provided that Zone 2 customers would pay 2.8 cents more per thousand cubic feet of gas than would Zone 1 customers, while Zone 3 customers would pay 3.6 cents more per thousand cubic feet than Zone 2 customers. The Federal Power Commission, the FERC’s predecessor, approved the plan, and subsequent rate settlements and schedules have incorporated these zone differentials.
Transco filed an application for higher rates in 1976. Transco had not proposed any change in the 1962 zone differentials, but the parties in settlement negotiations were unable to agree on a cost allocation method. The FERC ordered a hearing on the issue, and an administrative law judge approved the 1962 differentials.1 The Commission, however, rejected the ALJ’s determination and ordered costs to be allocated by the Mcf-mile method.2 Zone rates thus would be “based upon fully allocated costs without regard to the settlement based differentials previously in effect.”3
The D.C. Circuit set aside the Commission’s cost allocation order in Public Service Commission of the State of New York v. FERC.4 The appellate court first rejected the move to Mcf because the FERC had not found that the 1962 differentials were, in the language of § 5(a) of the Natural Gas Act, “unjust, unreasonable, unduly discriminatory, or preferential.”5 The Commission bore the burden of making such proof by substantial evidence.6 “It is not enough that the petitioners failed to prove that the zone differentials are just and reasonable. It is the Commission which must adduce substantial evidence tending to show that the existing zone rates are unjust and unreasonable.”7 The D.C. Circuit also found that the FERC had not supported the allocation change with substantial evidence.8
Transco filed an application for a general rate increase on March 31, 1982. The parties reached a settlement, which reserved the issue of the zone differentials.9 In reviewing the settlement, the AU found the current differentials to be unjust and discriminatory and, after finding distance to be the primary cost factor on the transmission system, ordered partial implementation of an Mcf-mile cost allocation method.10 Transmission costs incurred in transporting gas from off-shore gathering points to compressor Station 65 were excluded from Mcf allocation.11 The Commission affirmed in part and modified in part the AU’s decision in Opinion No. 260.12 The Commission denied in part and granted [1552]*1552in part rehearing in Opinion No. 260-A.13
The pipeline customers raise multiple challenges on appeal. The Zone 3 customers contend that the AU unlawfully placed the burden of proof on them. Zone 3 also argues that neither the unlawfulness of the 1962 differentials nor the appropriateness of Mcf-mile allocation were supported by substantial evidence. The Zone 2 customers attack the exclusion of upstream production costs from Mcf-mile allocation. Both zones, as well as the North Carolina Utilities Commission, contest the date the Commission set for the new differentials to take effect.
II
A
Zone 3 first contends that the AU unlawfully shifted the burden of proof to it, despite PSCNY’s statement that the party advocating the change bears the burden to show the existing rates are unjust and discriminatory. Zone 3 of course recognizes that the AU appreciated the importance of PSCNY:
Since Public Service has not been explicitly overruled, that decision must be regarded as precedent and its dictates complied with in the present proceeding. Advocates of a change in the present zone differentials must show by substantial evidence both that the existing rate differentials are unlawful and that the proposed zone differentials based on the Mcf-mile method are lawful and in compliance with the strictures of the Act.14
However, after then looking at Commission precedent, which determined that distance was the primary factor in selecting a proper cost allocation method, the AU stated, “As framed by these Commission decisions, the question to be answered is whether any of the factors identified by the Zone 3 Group present such 'extreme variations’ as to outweigh distance as the primary cost factor on Transco’s transmission system.” 15 It is this framing of the question that Zone 3 contends erroneously shifted the burden of proof.
We disagree. First, it is clear that both the AU and the Commission were well aware of the proper allocation of the burden of proof.16 Second, the precedent cited by the AU recognizes a general economic argument that Zone 3 does not dispute. “It is a simple economic fact that the delivery cost of natural gas increases in close proportion to the length of the transmission line of any given size.”17 Third, the AU clearly believed this “simple economic fact” was true on this particular pipeline, and his reference to the testimony of Mr. Clay, one of Zone 2’s witnesses,18 demonstrates his awareness that substantial evidence existed to support the finding that distance was the primary cost factor on Transco. We thus reject Zone 3’s contention that the AU improperly relied on “general [FERC] principles” to shift the burden of proof.
B
After finding distance to be the primary cost factor on the Transco system, the AU directly addressed the legality of the 1962 differentials:
Assuming arguendo that the historic differentials perfectly allocated transmission costs at the time at which they were agreed upon, the contention that they still do so today, some twenty years later is contrary to the record as well as to common sense. A pipeline’s transmission costs are composed of a number of components, including compression costs, and costs related to the pipeline [1553]*1553facilities themselves such as upkeep and taxes. As recounted above, according to the testimony of Mr. Clay, if the zone differentials were adjusted to take into account the increases in Transco’s transmission costs the differentials between zone 1 and 2 would have to be more than doubled and that between zone 2 and 3 would have to be more than tripled. To argue as do the Zone 3 Group, that all of these costs have remained the same or that increases in some costs have been completely offset by decreases in others is contrary to substantial record evidence.19
Zone 3, however, believes the ALJ’s finding that the historic zone differentials were unlawful was not supported by substantial evidence. First, Zone 3 argues that the FERC should have considered the historic rationale for the differentials. Second, Zone 3 offers a variety of factors that it believes outweigh distance as the primary cost determinant on the Transco system.20
Zone 3 first stresses the historic underpinnings of the differentials, arguing that the pipeline never would have been built or expanded but for the existence of the Zone 3 market. In emphasizing this historic basis, Zone 3 relies on the Supreme Court’s discussion in Colorado Interstate Gas Co. v. FPC: 21
The Commission found that but for the direct industrial market at Pueblo, Colorado and the wholesale market at Denver, the pipeline would not have been constructed. It is therefore obviously fair to determine transmission costs for the pipeline as a whole and not to compute them on a mileage demand basis. In that way the beneficiaries of the entire project share equitably in the cost. To allow the costs to accumulate the closer the gas gets to Denver would be to assume that the extension to Denver was a separate project on which the earlier customers were in no way dependent. These circumstances illustrate that considerations of fairness, not mere mathematics, govern the allocation of costs.22
The Commission, however, correctly points out that Colorado Interstate states that on those facts it was fair not to compute costs on a mileage demand. basis.23 Colorado Interstate does not foreclose the ALJ finding on the facts before him that the 1962 rates were unlawful and that Mcf-mile cost allocation was appropriate.24 That Zone 3 may have been the catalyst for the system being built does not negate the fact that the other zones also are important and viable markets; nor do the historic reasons justify an inequitable distribution of costs.
Zone 3 next contends that non-mileage factors predominate on the Transco system, focusing first on terminal zone storage. According to Zone 3, storage “permits more efficient pipeline operations” and eliminates the need to construct additional facilities to serve the peak demands of the other two zones.25 We, however, [1554]*1554agree with the AU that the storage issue does not bear as direct a relationship to costs on the Transco system as does distance.26
Zone 3 next looks at load factor and load concentration. As the FERC explains with regard to load factor, the more fully a system is used, the lower the per unit fixed costs. Thus, the argument goes, high load factor customers should share a lesser portion of the fixed costs associated with transmission capacity than low load customers. High load concentration permits the use of economies of scale. According to Zone 3, its load factor is 78.32% and its load concentration is 75%, thus justifying a departure from Mcf-mile cost allocation.27 These figures, however, are far below the extreme variations the Commission has found to outweigh distance in other cases.28 More importantly, as the AU stated, “[Tjhese calculations shed no light on ... whether the cost savings associated with zone 3’s load factor and load concentration exceed the costs associated with the extra distance involved in serving zone 3 customers.” 29
Zone 3 also urges that distance should not be relied on when there are supply sources near the terminal zone. However, present Canadian supplies to Transco are insignificant, accounting for less than 2% of Transco’s total sales; these sales are certified only for a limited time and future sales are speculative.30
Finally, Zone 3 tells us that “fully 12 major pipelines charge system-wide rates that do not, through rate differentials, reflect the varying miles of haul from gas supply sources to the markets served. In addition, there are four pipelines that, like Transco, utilize fixed zone rate differentials that are not adjusted to reflect mileage costs of transmission.” The Commission responds that other major long line systems do use Mcf-mile cost allocation. Regardless of which party’s analysis of other pipelines is accurate, the proper question is whether substantial evidence supports the use of Mcf-mile cost allocation on the Transco system. We are persuaded that it does.
C
Zone 3 separately attacks the AU’s adoption of the mileage cost approach on the basis that the AU should have considered Zone 3’s load factor, load concentration, and storage facilities. We previously have discussed those factors and see no need to do so again. Substantial evidence shows that distance is the primary cost-determinant on the system and that the Mcf-mile approach measures distance; substantial evidence thus supports the AU’s determination of the legality of Mcf cost allocation. While this explanation seems simplistic, it is a product of our standard of review and not an ignoring of true complexities.
[1555]*1555III
The ALJ, however, did not adopt an Mcf mileage allocation in its entirety. Instead, the AU excluded from Mcf-mile calculations those costs incurred in transporting gas from central gathering points in the Gulf of Mexico to compressor station 65, located at the Louisiana-Mississippi state border; the parties refer to these as upstream costs.31 This approach to the upstream costs is known as the zone gate method, under which costs are calculated and allocated separately for each zone.
Zone 2 contends the FERC erred in excluding upstream costs from the Mcf-mile method. Zone 2 first argues that use of the Mcf-mile method on the entire system would have resulted in a proper allocation of costs.
We, of course, look to whether substantial evidence supports the AU’s determination, and conclude that it does. As the AU found, upstream costs are higher on average than non-production area transmission costs. If the Mcf-mile method was used for upstream costs, Zone 3 would pay a greater proportion of these costs. However, the AU found that these costs did not bear any relationship to distance.32 As the FERC tells us, and we agree, these costs are incurred to serve everyone downstream and thus it is fair to allocate these costs volumetrically.
Zone 2 also argues that the “gas supply area investment factor” does not justify exclusion. This argument evidently was proffered by one of Zone 3’s witnesses, and Zone 2 assumes that the AU and the Commission adopted it. However, we see no mention of this factor in the decisions. In any event, this argument does not refute the substantial evidence supporting the AU.
Zone 2 next argues that the Commission’s approach is contrary to Commission precedent, citing Southern Natural Gas Co.33 The FERC there stated:
Even if we accept the proposition that a significant portion of Southern’s transmission investment relates to transmission supply lines whose costs do not vary with the distance of haul, it does not follow that the system-wide cost allocation should be retained. At best, this argument would justify the reclassification of that portion of the transmission costs related to supply to the commodity component of Southern’s rates. This argument has no bearing on the allocation of the cost of the transmission facilities not so reclassified.34
Zone 2 believes that Southern Natural stands for the proposition that all transmission costs are to be included under the Mcf-mile method unless there is evidence that those costs should be reclassified as a commodity component of rates. Southern Natural, however, established no such per se rule. The statement is a generality and does not address the issue directly before us.
Finally, Zone 2 argues that even if the upstream costs are not transmission costs but rather gas supply costs, then they should have been reclassified as gas supply costs and allocated volumetrically. As Zone 2 puts it, upstream costs should not be “functionalized to transmission [and] be excluded from mileaging as ‘gas supply costs,’ while on the other hand continue to be treated as transmission costs for purposes of cost classification and rate design.” Instead, the costs should have been reclassified as gas supply costs and allocated volumetrically in commodity rates.
The logic of this argument is not immediately apparent, and thus some explanation is necessary. Pipeline costs are “function-alized” into production, gathering, storage, and transmission; gas supply costs are en[1556]*1556compassed in the production and gathering functions. Costs are either fixed or variable and apportioned between demand and commodity rate components. Demand charges are assessed on the basis of each customer’s peak and contractual entitlements, while commodity charges are assessed on the basis of each customer’s actual purchases. Under the Commission’s approach, production area costs are allocated to the commodity component, while transmission costs are allocated to both demand and commodity. The hybrid approach allocates upstream costs volumet-rically, like gas supply costs, but apportions them to the demand and commodity components, unlike gas supply costs. Reclassification of upstream costs as gas supply would transfer the costs to the commodity component, benefiting Zone 2 because it makes small actual purchases.
Zone 2, however, has not presented us with any reason for reclassifying these costs. That the upstream costs are not allocated on a mileage basis does not make them gas supply costs. If Zone 2 wants these costs reclassified, it should present evidence supporting that reclassification. We agree with the Commission that
North Carolina has confounded two distinct issues. The first is functionalization of costs as transmission or gas costs. The second is classification of the functionalized costs into the demand and commodity components. As North Carolina observed, “no party has challenged the propriety of Transco’s functionalization of those costs to transmission.” Once functionalized to transmission, the costs should be classified under the MFY method to the demand and commodity components.35
IV
We finally turn to the effective date of the new cost allocation, which all parties contest. Article X of a 1982 Settlement Agreement between the parties reads in pertinent part:
In this proceeding, numerous issues have been raised with respect to the design of the aforesaid rates as well as other rate design, cost classification and cost allocation issues. As a consequence, the parties agree that such issues are hereby reserved for hearing and decision, with any changes from the methods and principles utilized in this Settlement Agreement to be effective prospectively only from the first day of the month following 30 days after the effective date of the Commission’s order requiring any such change whether or not the rates herein have been superseded by a new general rate increase filing.
Opinion No. 260-A set the effective date of the order as April 1, 1987, explaining that April 1 was the effective date for Transco rate increases in Docket No. RP87-7.36 However, the Commission in mid-October changed the effective date to October 1, 1987, relying on the settlement.37 The Commission affirmed that date in January 1988.38
We first address Zone 3’s arguments. Zone 3 first argues that the Commission has misconstrued the settlement, which calls for the rates to be effective “prospectively only.” Zone 3 urges that the settlement must be interpreted in light of section 5(a) of the NGA, which directs the Commission to “determine the just and reasonable rate ... [and] fix the same by order.”39 Thus, according to Zone 3, “the effective date [of the rates] cannot be any earlier than the date on which the Commission accepts or adopts, and thereby ‘fixes’ the rates after receipt of the compliance filing.” For this proposition Zone 3 cites Electrical District No. 1 v. FERC,40 in [1557]*1557which the D.C. Circuit stated that to fix a rate under a parallel provision in the Federal Power Act, the Commission must specify the rate. “[I]t is [not] enough to prescribe the legal and accounting principles which, properly applied, will yield one particular rate.”41 Thus, the October 1, 1987, date violates the settlement and the NGA because the Commission must wait until it accepts Transco’s compliance filing before it can set an effective date for new rates.
This situation, however, is controlled by the Settlement Agreement, to which Electrical District does not purport to apply. The agreement refers specifically to the FERC’s order requiring changes in methods and principles. Unlike § 5(a), it does not refer to the date the Commission fixes the rates. When the parties signed the settlement, they agreed to the date specified in it, regardless of the dictates of § 5.
Zone 3 also argues that Commission policy expressly precludes retroactive changes in rate design. Regardless of previous Commission policy, however, it is the settlement that controls.
Zone 2 and the North Carolina Utilities Commission also attack the effective date of the new rates. The FERC interpreted the settlement language “ ‘to be made effective prospectively only from the first day of the month following 30 days after the effective date of the Commission order’ to mean ‘from the date the rehearing order is issued.’ ”42 In interpreting the language as such, the FERC relied on Panhandle Eastern Pipe Line Co.,43 in which the Commission interpreted settlement language that read “prospectively from and after a Commission order disposing of [the] issues” as referring to a final disposition on rehearing.44 Thus, the FERC calculated the time from Opinion No. 260-A, the rehearing order, rather than Opinion No. 260.
In protesting that time should be calculated from Opinion No. 260, Zone 2 and the NCUC believe the FERC misreads the settlement, which calls for calculation from the Commission order “requiring such change.” Zone 2 also points to a footnote in the October 16, 1987, order denying in part and granting in part rehearing; the footnote reads in part, “It is true that Opinion No. 260 required the change and rehearing was denied in Opinion No. 260-A.”45 Finally, Zone 2 disputes reliance on Panhandle Eastern, although it is unclear why Zone 2 believes the reliance to be misplaced.
We agree with the FERC that Opinion No. 260-A is the order to which the settlement refers. As the FERC stated in the portion of the footnote Zone 2 deigned not quote, “[T]he issue was not finally resolved until issuance of Opinion No. 260-A.”46 The FERC did not consider the issue final until the rehearing order. Without clear settlement language to the contrary and because the issues in the case were not settled until Opinion No. 260-A, it was not unreasonable for the Commission to conclude that the parties meant a final disposition of the issue.47
Zone 2 next argues that the October date results in undue discrimination. Zone 2 notes that an extremely long time passed between the AU’s initial decision finding the rates to be discriminatory and the effective date of the new rates; during this time Zone 2 continued to pay unreasonable rates. This is coupled with the fact that April 1, 1987, was the ordered effective date for the switch to a modified fixed-variable rate design (MFV) provided for in Opinion No. 260; according to Zone 2, this change shifted more fixed costs related to the unjust rates to Zone 2. The parties, [1558]*1558however, provided for the rate differentials' effective date in the settlement, and we see nothing unduly discriminatory in holding the parties to their bargain.48
The NCUC argues that the Commission fails to give proper consideration to the Natural Gas Act’s purpose of protecting consumers from unjust rates. However, as we already have found the Commission’s interpretation of the settlement to be reasonable, we reject this contention.
The NCUC also argues that the settlement no longer controls. In October 1986, in Docket No. RP87-7, Transco proposed a rate design change to MFV. This was prior to issuance of Opinion Nos. 260 and 260-A, and the NCUC argued that Article X of the Settlement Agreement precluded the October 1986 filing.
The FERC, however, decided that Article X had been affected by a settlement in another docket in 1984, in which Transco had agreed not to propose any general rate increase or change in the basic design of its rate structure until 1985. This was Article XIV, which read:
As part of this Agreement, Transco agrees that it shall not make effective a general rate increase (as defined in Section 154.63(a)(2) of the Commission’s Regulation) in the basic design of its sales rate schedules to be effective prior to April 1, 1985.49
Because the Commission interpreted Article XIV to mean that Transco could make such a filing after that date, it permitted the filing in Docket No. RP87-7, subject to the outcome in Docket No. RP82-55, which eventually became Nos. 260 and 260-A. The RP87-7 rates became effective subject to refund on April 1, 1987. The FERC specifically noted that the 1984 settlement “did not provide for the continued effectiveness of Article X of the settlement in Docket No. RP 82-55-000” but instead only barred changes in rate design prior to April 1, 1985.50
Although the NCUC argues that this language renders Article X ineffective, we disagree. Instead, we adopt the Commission’s discussion of the issue:
The Commission believes that its statement that Article X of the 1982 settlement did not continue in effect should be limited to the context of the issue before it in the suspension order and should not be read as a determination that Article X of the 1982 settlement was thereby terminated. This narrow interpretation of the November 5, 1986 suspension order is supported by two different reasons. First, the settlement language in Docket No. RP83-137-000 arose in response to a controversy over whether the settlement in the instant docket prohibited Transco from filing rate design changes. Article XIV of the settlement in Docket No. RP83-137-00 resolved that issue. The Commission believes that was all it did. It did not terminate Article X of the 1982 settlement. The second related reason is that an interpretation that Article X was terminated would lead to absurd results. Article X not only addressed the effective date issue but reserved numerous issues for hearing. If Article X were terminated, the Commission could not have resolved those issues in Opinion Nos. 260 and 260-A. The Commission concludes that Article X continued in effect except with respect to any prohibition on the filing of rate design changes by Transco.51
Finally, the NCUC argues that the 1982 settlement does not bar a challenge to rates in subsequent filings. As the NCUC correctly notes in its reply brief, however, this “matter is not one directly before the [cjourt in this appeal.”
AFFIRMED.