Cellular telephone service continues its slow march from the regulatory arena to the marketplace. On March 6, 1984, thirty-five years after the Federal Communications Commission first allocated a small band of the radio spectrum for the development of common carrier mobile radio communications and over one year after the grant of licenses to wireline carriers, the first nonwireline cellular telephone comparative proceedings were decided by the Commission. In this case (Pittsburgh) and its companion case, Cellular Mobile Systems of Illinois, Inc. v. FCC (Chicago),1 unsuccessful competitors seek review of the first two of a series of FCC awards of nonwireline cellular licenses in the top thirty metropolitan markets.

The principal question presented in this appeal by Cellular Mobile Services of Pennsylvania, Inc. (CMS), one of two unsuccessful competitors for the Pittsburgh nonwireline license, is whether the FCC's award of comparative preferences and its ultimate grant of the Pittsburgh license to MCI Cellular Telephone Company were arbitrary and capricious, an abuse of discretion, or unsupported by substantial evidence. Also before us is whether the exercise of discretion by the administrative law judge, under special procedures crafted by the FCC for cellular proceedings, violated federal statutory or due process requirements. Finally, we have for determination the question whether the comparative standards announced in the FCC's cellular rulemaking, Cellular Communications Systems, 86 F.C.C.2d 469 (1981) (Report and Order), modified, 89 F.C.C.2d 58, further modified, 90 F.C.C.2d 571 (1982), petition for review dismissed sub nom. United States v. FCC, No. 82-1526 (D.C.Cir. Mar. 3, 1983) (hereinafter Cellular Rulemaking), failed to provide adequate notice of the actual comparative criteria employed by the FCC.

We affirm. As to CMS's procedural attack, we conclude that the expedited cellular procedures as applied in the Pittsburgh proceeding were lawful.2 We further hold that the rulemaking Report and Order provided ample notice of the bases for agency comparison of contending applicants. Finally, we are persuaded that the agency's ultimate decision, which the FCC acknowledged to be "close," embodied reasoned decisionmaking and is supported by substantial evidence.

* The technological benefits and regulatory history of cellular telephone service have already been amply rehearsed in prior opinions of this court, MCI Cellular Telephone Co. v. FCC, 738 F.2d 1322 (D.C.Cir.1984); National Association of Regulatory Utility Commissioners v. FCC, 525 F.2d 630 (D.C.Cir.1976), and need not be repeated here.

In its 1980 cellular rulemaking, the FCC expressed concern that the traditional comparative process might cause significant delay in implementing cellular service; accordingly, the Commission invited comments on alternative procedures for streamlining the comparative process. The FCC initially proposed a "hybrid" two-stage plan, under which applicants who met strict qualification requirements would then be designated for a comparative hearing. If the applications failed to justify a comparative preference, then a tie-breaking lottery or drawing would be held. Cellular Communications Systems, 78 F.C.C.2d 984, 1000-01 (1980) (Notice of Inquiry and Notice of Proposed Rulemaking).3

Comments were also solicited on the comparative criteria to be employed. Observing that the traditional comparative criteria "focused on the nature and extent of the proposed service ..., the areas and populations to be served, and the need for the proposed service in these areas," id. at 1000, the Commission sought suggestions for other possible criteria, such as plans for phased growth and marketing plans and capabilities. Id.

* In 1981, the FCC issued its Report and Order, announcing that "[t]he pressing need for improved service compels us to seek a solution that will bring relief to the greatest number of people in the shortest possible time." Cellular Rulemaking, 86 F.C.C.2d at 490. Rejecting the hybrid procedures proposed in the NOPR, as well as a lottery approach, the FCC embraced the structure of comparative proceedings for evaluating cellular applicants, concluding that "there may be significant differences among competing applications." Id. at 499.4 To reduce the delays attendant upon traditional hearings, "paper" procedures were prescribed; the FCC reasoned that such streamlining would serve the public interest in expeditious disposition of these matters while safeguarding applicants' procedural interests. Id.

Under the procedures as adopted, applications would first be measured against the technical and financial basic qualifications requirements. Parties satisfying those requirements would then be designated for comparative consideration. Briefs and evidence seeking to demonstrate superiority would then be evaluated by an administrative law judge.

In response to comments decrying the practice of later applicants engaging in "one-upmanship" over earlier applicants, the Commission, on reconsideration, emphasized that all applicants were to submit their direct cases along with their applications, including all evidence upon which the applicant intended to rely in a comparative hearing. Cellular Communications Systems, 89 F . C.C.2d 58, 89 n. 50, 91 (1981) (Order on Reconsideration). Moreover, the applications for all thirty markets were given a due date of June 7, 1982 in order to avoid the perceived evil of one-upmanship:

We want all participants to file applications which represent their best view of a service plan for the named SMSA. To do so, we do not find it necessary for participants to consult the plans of their potential competitors. Setting up a plan which would allow applicants to revise their filings after viewing the applications of others would encourage applicants to engage in "one-upmanship," which has harmful consequences. This would undermine our ability to compare proposals with some measure of confidence that the applicant had participated in its development. Plans based on another proposal would no longer represent the applicants' best idea of how to serve a given area but would, instead, represent applicants' use of the administrative process to obtain an advantage over competitors. Furthermore, allowing opportunity for one-upmanship would needlessly encumber an administrative process which we must streamline to its essentials if the American public is to receive cellular service without unnecessary delay.

Id. at 89 (footnote omitted).

Further pleadings, the Commission decreed, would be limited to Petitions to Deny and Replies. In a Designation Order, the FCC would slate mutually exclusive applications for comparative hearing; additionally, any "unresolved, substantial and material questions of fact" could be designated for hearing as well. Id. at 92. After the Designation Order, a second stage of evidentiary submissions--written rebuttal and rebuttal evidence--would be permitted. Whereas the Report and Order provided for discovery, the FCC on reconsideration significantly curtailed its availability; discovery was to be permitted only "if the designation order points out an area where additional underlying data is needed." Id.; 47 C.F.R. Sec. 22.916(b)(6) (1984).

Opportunities for cross-examination and for objections to both direct and rebuttal evidence would be provided, within the discretion of the ALJ, in a hearing format. Under the enumerated procedures, the ALJ was vested with discretion to permit oral testimony in the "unusual situations" requiring it,5 and to permit cross-examination or other procedures "upon a substantial showing that a party will be prejudiced by the submission of all the evidence in written form." Cellular Rulemaking, 86 F.C.C.2d at 499 (citing 5 U.S.C. Sec. 556(d)); see also 47 C.F.R. Sec. 22.916(b)(6) (1984). The governing standard required a showing of specific reasons for the necessity of an oral hearing or cross-examination.6 Additional testimony might be permitted when "needed to provide a complete record for the Commission," but the Commission warned that such further sessions were to be expeditious and efficient. Order on Reconsideration, 89 F.C.C.2d at 92.7

On further reconsideration, the Commission acknowledged the arguments advanced for adding a third evidentiary stage (to respond to disputes created by the rebuttal filings), but declined to extend the pleading cycle. The FCC observed that "if ... there is a need to obtain additional information from any of the parties, the Administrative Law Judge already has the discretion to require it, either in written or oral form." Cellular Communications Systems, 90 F.C.C.2d 571, 576 (1982) (Order on Further Reconsideration).

The comparative criteria articulated by the Commission expanded upon that used in the traditional Domestic Public Land Mobile Radio Service, see Notice of Proposed Rulemaking, 78 F.C.C.2d at 1000, reflecting the goals emphasized for cellular implementation. Cellular Rulemaking, 86 F.C.C.2d at 502.8 Three rather general criteria were outlined. The first, and primary, criterion was a comparison of the respective applicants' initial systems' geographical coverage:

Because nationwide availability of service is a primary goal, a major basis of comparison will be the geographic area that an applicant proposes to serve. In comparing proposed service areas, other significant factors to be considered will be the presence of densely populated regions, highways, and areas likely to have high mobile usage characteristics, as well as indications of a substantial public need for the services proposed, including the results of public need surveys.

Cellular Rulemaking, 86 F.C.C.2d at 502 (emphasis added).

The geographic area to be served, or "cellular geographic service area" (CGSA), was to be defined by the applicant. Id. at 509. While the size of the CGSA was not limited initially, an applicant was required to propose reliable service to 75 percent of the CGSA, that is, the land area within its aggregate 39 dBu reliable service area contours. Id.9 The 75 percent coverage was required to be maintained as the system evolved. Id. In keeping with this regulatory emphasis upon geographical coverage, the only system changes deemed to constitute "major applications," which trigger new cellular licensing proceedings, were those that would alter the CGSA boundaries. Id. This procedure reflected the FCC's policy determination that "the key to our regulatory structure is the geographic service area of a cellular system." Id. (emphasis added).

On reconsideration, the Commission clarified the 75 percent requirement, indicating that it must be met within three years. 89 F.C.C.2d at 86 n. 44. Expressing reluctance "to interfere in any way with an applicant's discretion to frame its proposal in response to its perception of the market," id. at 86, the Commission imposed an outer limit on applicants' CGSA's to be measured by the boundary of the Standard Metropolitan Statistical Area (SMSA). More precisely, while a CGSA could exceed the SMSA, the extensions could be no more than de minimis. Id. at 87. The purpose of this limitation was to prevent large comparative hearings pitting expansive yet divergent CGSA's against each other simply because they overlapped in a common SMSA.

The second major comparative factor enumerated by the FCC was

the applicant's ability to expand its system capacity in a coordinated manner within the proposed service area in order to serve an increasing number of local subscribers and roamers as demand warrants.

Cellular Rulemaking, 86 F.C.C.2d at 502-03. (emphasis added).10 The Commission added that expansion might take place through the addition of transmitters or cell locations and the use of smaller cells. A desirable characteristic of coordinated expansion would be "efficient frequency use," id. at 503, as illustrated by an applicant's plans to split cells and add channels; the ability to coordinate channel use with other cellular systems; and the system's degree of frequency use. Id. This criterion was intended to reflect the FCC's emphasis on the factor of flexibility which is "inherent in the cellular concept."11

The third comparative factor was actually a combination of operational aspects, such as facilities, maintenance, personnel, rates, charges, and regulations. Id. However, these areas of comparison were considered insignificant in most cases. As a result, the first and second comparative factors far outweighed the third; standing above the other two, however, was the first factor, and especially the specific matter of geographic coverage.

On June 7, 1982, applications and direct case exhibits were filed for the 30 largest markets, including Pittsburgh. Petitions to Deny and Replies were filed, asserting various deficiencies in the competing proposals. The Commission's Designation Order followed on December 6, 1982. Having satisfied the basic technical and financial requirements,12 all three nonwireline applicants for the Pittsburgh market, CMS, MCI Cellular Telephone Co., and Celcom Communications Corporation of Pittsburgh,13 were declared "qualified" to construct and operate the cellular systems proposed. Advanced Mobile Phone Service, Inc., 52 Rad.Reg.2d (P & F) 1104 (Common Carrier Bureau 1982) (Pittsburgh Designation Order). The three applications were designated for comparative hearing on the following issues:

(a) to determine on a comparative basis the geographic area and population that each applicant proposes to serve,14 to determine and compare the relative demand for the services proposed in said area; and to determine and compare the ability of each applicant's cellular system to accommodate the anticipated demand for both local and roamer service;

(b) to determine on a comparative basis each applicant's proposal for expanding its system capacity in a coordinated manner within its proposed CGSA in order to meet anticipated increasing demand for local and roamer service;15

(c) to determine on a comparative basis the nature and extent of the service proposed by each applicant, including each applicant's proposed rates, charges, maintenance, personnel, practices, classifications, regulations and facilities (including switching capabilities); and

(d) to determine, in light of the evidence adduced under the foregoing issues, what disposition of the referenced applications would best serve the public interest, convenience and necessity.

Id. at 1110 (footnotes omitted). The Order directed the applicants to file their respective rebuttal cases within 45 days. Id. at 1111. Subsequently, the Commission's then-Chief Administrative Law Judge, Lenore G. Ehrig, was designated to hear the case.

CMS, joined by MCI, promptly requested a prehearing conference to alter the prescribed schedule.16 CMS urged the ALJ to permit discovery in order fully to develop the evidentiary record, while at the same time avoiding extensive use of cross-examination at the hearing itself. CMS also requested rulings on the admissibility of direct case exhibits prior to the filing of rebuttal cases; notwithstanding the FCC's rejection of this procedure in the main, CMS additionally requested that a third stage (surrebuttal) be added to the pleading cycle, on the asserted ground that parties should enjoy an opportunity to respond to new matters presented in rebuttal exhibits.

Five days later, Chief Judge Ehrig issued an order rejecting the request for discovery, on the ground that the Designation Order had not pointed to any area where additional underlying data were needed, citing the FCC's Order on Reconsideration, 89 F.C.C.2d at 92, and the governing regulation, 47 C.F.R. Sec. 22.916(b)(6). MCI Cellular Telephone Co., FCC 82M-4023, at 4 (Dec. 17, 1982) (Order Prior to First Hearing Session). The ALJ directed that memoranda of objections and requests for cross-examination be filed within 25 days. The objections were to be set forth "with particularity." In requesting cross-examination, the parties were to (1) demonstrate that written evidence was ineffectual to develop proof; (2) specify their objectives on cross-examination; and (3) identify the evidence which cross-examination was expected to adduce and the testimony which was expected to be discredited. Similar memoranda were to follow the rebuttal case filings. No responsive pleadings were to be permitted. Id. at 2.

On January 27, 1983, the applicants filed their respective rebuttal cases. Because the ALJ was "unable to clearly ascertain each applicant's proposals for either the first year of operation or plans for expansion thereafter," the parties were ordered to supplement their cases with memoranda which explained these plans with greater "particularity," yet to rely only on evidence already proffered. MCI Cellular Telephone Co., FCC 83M-953, at 3 (Mar. 22, 1983) (Order).

The sole hearing before the ALJ in the course of these proceedings occurred on April 9, 1983.17 At the hearing, rulings on evidentiary objections were announced; in addition, all requests for cross-examination were denied inasmuch as the parties' written requests, according to the ALJ, had "failed to persuasively demonstrate that the written evidence is ineffectual to develop proof." Transcript at 67. CMS again requested surrebuttal, for the stated purpose of responding to "matters [which] have been raised for the first time in rebuttal." Id. at 68. The ALJ denied the request, stating that the record was sufficiently complete to permit a "considered judgment." Id. at 72; see 47 C.F.R. Sec. 22.916(b)(6) (1984). The record was, accordingly, closed on the day of hearing. The parties filed Proposed Findings of Fact and Conclusions of Law on May 25, 1983; the presiding judge thereafter permitted Reply Findings and Conclusions. See 47 C.F.R. Sec. 22.916(b)(7) (1984).

The ALJ's Initial Decision was rendered in July 1983. MCI Cellular Telephone Co., 96 F.C.C.2d 1040 (1983) (Pittsburgh Initial Decision). In a lengthy opinion, the Presiding Judge concluded that MCI's proposal was comparatively superior and would best serve the public interest, convenience and necessity. Id. at 1068.1

Under the first--and pivotal--element of Issue (a), geographic and population coverage, the ALJ compared the areas within the parties' respective 39 dBu contours. MCI's proposal covered 89.6 percent (2,610 sq. mi.) of the Pittsburgh SMSA, Celcom's proposal covered 70.4 percent (2,197 sq. mi.), and CMS's proposal covered 79.3 percent (2,324 sq. mi.). Their respective population coverage was 96.96 percent (2,195,146 people) for MCI, 94.4 percent (2,136,847 people) for CMS, and 92.8 percent (2,101,315 people) for Celcom. Id. at 1043, 1060.18 None of the proposals covered the entire SMSA. MCI thus had the edge over its competitors in both geographic and population coverage.

In comparing the presence of densely populated regions, highways, and areas likely to have high mobile usage characteristics, the judge found that the non-common areas or "pockets" served by MCI were superior by virtue of the fact that CMS and Celcom had excluded more significant pockets than MCI. Id. at 1060. Overall, MCI was awarded a slight preference on this sub-issue.19

MCI was also awarded a preference under the "relative demand for service" component of Issue (a),20 a point, as we shall presently see, which is vehemently contested by CMS. The MCI demand forecast was based upon three studies--a public need study, a selected industry survey, and "focus groups"--as well as a demographic analysis. The public need survey consisted of a telephone survey of 796 randomly selected Pittsburgh households with incomes of at least $20,000. Data on likelihood of subscription at three different cost levels were collected. MCI applied an "Intent Translation Scale" to derive realized demand from gross potential demand and then reduced the demand values by a variable which reflected the common-sense assumption that survey respondents tended to express a greater willingness to subscribe than would occur in practice. MCI also considered, among other factors, the impact of price changes over time, the introduction of new service features, and population and industrial development. Id. at 1044. This survey alone was the source of MCI's realized demand forecast.

MCI also conducted a selected industry survey--a telephone survey of 100 Pittsburgh businesses chosen from five industry sectors expected to have a need for cellular service. This survey sought information such as potential subscriber attitudes, likely usage patterns, price sensitivity, and likely business user characteristics. Id. MCI's third study consisted of "focus group" sessions, conducted outside the Pittsburgh SMSA, which were open-ended group interviews with selected potential subscribers. These were designed to gather only information to identify consumer preferences, and had no impact on the demand or subscriber estimates. Id. at 1045.

In contrast to MCI, CMS commissioned six nationwide market research studies. All six were premised on the assumption, criticized by the ALJ, that businesses would be the primary users of cellular services. Pivotally for what ultimately transpired, three of the studies relied on the assumption that radio pager service constituted a precursor of cellular radio, an assumption sharply faulted by the ALJ. Id. at 1061.

The primary survey employed by CMS, the nationwide Claritas survey, was designed to estimate demand based on pager penetration in the various markets throughout the country in which CMS's parent, Graphic Scanning Corporation, offered paging services. Two mail surveys were conducted in Pittsburgh, as well as other potential markets, one of selected businesses and the other of Graphic Scanning's paging customers. These surveys collected information concerning monthly charges and anticipated cellular usage and habits. Rounding out this bevy of CMS studies, three telephone surveys based on the same questionnaire were conducted to verify the demand figures. One survey was of the business population, another was of paging customers, and the third was a survey of businesses by Westat Corporation. The Westat study was conducted in the top 30 cellular markets, including Pittsburgh. In that study, the calculation of Pittsburgh's demand was not based solely on Pittsburgh's responses; the national market data was factored in as well. Id. at 1046. The Westat survey estimated a fifth-year need of 42,300 cellular units, which CMS interpreted as reflective of interest, not of actual use. Id. at 1047.

CMS did not, however, utilize any of its various survey demand figures as its final demand calculation; instead, CMS's estimated demand reflected the results of a model employed in all thirty markets. That is, the Claritas study of paging demand yielded paging penetration data for the top thirty SMSA's. CMS Direct Exh. II, at 12. The top thirty markets were then grouped into three categories according to the level of paging demand indicated for each city. CMS Direct Exh. II, at 14 & Figure II-7. Then, each of the three categories was assigned a value, apparently reflecting the median paging demand for all cities in that category. Id. This value, called the "cellular penetration factor," Pittsburgh, 96 F.C.C.2d at 1023, or "need factor," was expressed as the projected need for cellular units per 1,000 population. CMS Direct Exh. II, at 14. Thus, Pittsburgh was assigned a need factor of 0.006, or 6 cellular units per 1,000 population, as were thirteen other SMSA's. 96 F.C.C.2d at 1047; CMS Direct Exh. II, at Figure II-7. As the last step, the final demand estimate was achieved by multiplying this figure by the SMSA census population. 96 F.C.C.2d at 1047; CMS Direct Exh. II, at 14, 16.

Based on this welter of conflicting studies, the ALJ concluded that CMS had failed to document the nexus between the six surveys and its final demand projections. 96 F.C.C.2d at 1061.21 An additional flaw discerned by the ALJ was CMS's apparent failure to consider the presence of a wireline carrier in Pittsburgh. Id.

MCI was also awarded a preference under the third aspect of Issue (a), namely the ability of the applicant initially to accommodate demand. The ALJ found that all three applicants had designed CGSA's with cells covering the obvious areas where need could be anticipated, such as downtown Pittsburgh, the airport, and major highways. Id. CMS's proposal was criticized by the ALJ, however, for excessive cell overlap in Pittsburgh--specifically, CMS allocated eight cells to the city, compared to MCI's four cells and Celcom's two cells. Stating that "fundamental principles of cellular design indicate that the spacing between cells should follow as uniform a pattern as possible to achieve optimal coverage with minimal intra-system co-channel interference," Judge Ehrig found that the record reflected "serious potential co-channel interference problems" in CMS's plan. Id. at 1054-55 & n. 13, 1062. The ALJ added that the lack of uniformity evident in CMS's proposal would result in the preclusion of rapid and inexpensive channel reallocation under its cell sectorization scheme. Id. The preference in this respect was thus awarded to MCI because its proposal "far better reflects an understanding of demand for service in the Pittsburgh SMSA," while CMS had failed to show "any nexus whatsoever between its proposed service area and its determination of need." Id. at 1063.

Contrary to CMS's assertion before us, the judge found that "depth of coverage"22 had not been established as a ground for preference by the Common Carrier Bureau, and that MCI had also accommodated the higher level of demand in downtown Pittsburgh for which CMS had sought a "depth of coverage" preference. CMS, the ALJ further concluded, had not demonstrated that it covered any important areas not covered by the other two applicants. CMS was further faulted for its unique initial system design, a plan deemed "not ... entirely consistent with the Commission's mandate that applicants design their systems to guarantee rapid implementation and expansion of cellular service. Rather than being a system which evolves from the initial stages to a more mature system, CMS has designed one likely to require a substantial amount of retuning and reassessment as demand increases." Id.23

In sum, Judge Ehrig concluded in this regard that MCI had proposed a geographic area "which most closely identifies, resembles, and meets the needs of the Pittsburgh [SMSA]," deserving a substantial comparative preference. Id. at 1063-64.

MCI was also awarded a preference under Issue (b) for its proposal for expansion of system capacity in a coordinated manner to meet the anticipated increase in demand. The ALJ concluded that MCI's plan reflected the "best understanding of cellular design principles and principles of expansion." Id. at 1065.

In contrast, CMS's proposal was extensively criticized by Judge Ehrig on the ground that it "would require significant fine-tuning before implementation of its expansion plans would be possible." Id. at 1064. The judge found that, while CMS planned an "essentially omni-directional" system, the frequency plan was designed for a highly sectorized system. Id. Unlike CMS, the ALJ concluded, MCI and Celcom demonstrated "an understanding of cellular design principles and reflect[ed] bona fide cellular systems and expansion plans." Id. Neither violated channel re-use criteria and both assigned frequencies according to a standard frequency reuse pattern. Id. Because those two applicants did not propose to build all their cells at the outset, the expansion of their respective systems could be achieved with a minimum of "costly engineering modifications" such as cell-splitting and sectorization. Such a plan was deemed to allow greater flexibility to expand and alter the system as further information about demand in the SMSA was acquired. Id. at 1064-65.

The ALJ declined, however, to award either MCI or Celcom a preference solely on the basis of greater system flexibility, inasmuch as neither proposal was set forth with "sufficient particularity." Id. at 1064. Instead, the factor enumerated as the basis for MCI's Issue (b) preference was its superior proposal for monitoring and anticipating congestion. CMS and Celcom, on the other hand, were faulted for proposing to allow the grade of service to deteriorate to an unacceptable level before remedying the problem. Id. at 1065.

Issue (c) provided CMS with its solitary preference, by virtue of its proposed rates and charges.24 Id. The CMS preference was based not on its actual charges but on its rate structure, which, the ALJ concluded, would encourage efficient cellular use. Responding to criticisms that CMS's rates were unrealistically low, the judge refused to impose on CMS the comparative demerit urged by its opponents. None of the other areas of consideration under Issue (c) were found to warrant preferences.

Based upon this analysis, the ALJ ultimately concluded that MCI's substantial preference under Issue (a) outweighed CMS's Issue (c) preference for rates and charges. Id. at 1068. CMS timely sought Commission review of the Initial Decision, Exceptions of CMS (Sept. 2, 1983), and also petitioned the Commission to reopen the record.25

In its Final Decision,26 released March 6, 1984, the Commission affirmed Judge Ehrig's Initial Decision, but with substantial modifications. MCI Cellular Telephone Co., 96 F.C.C.2d 1015 (Pittsburgh ), reconsideration denied, 56 Rad.Reg.2d (P & F) 936 (1984). Several of the preferences awarded to MCI were reduced in weight; two were eliminated entirely. Moreover, significant adverse conclusions as to the adequacy of CMS's design and the likelihood of intra-system interference were reversed.27 Although the margin of difference between the candidates was thus narrowed considerably, the FCC nonetheless discerned material remaining differences "sufficient to make a choice under comparative criteria previously established." Accordingly, the Commission affirmed the grant to MCI. Id. at 1018, 1035.

* First, and critical to its ultimate determination, the FCC affirmed the slight preference awarded to MCI for geographic and population coverage. Because nationwide service is a primary goal of cellular service licensing, geographic coverage was deemed to be a major basis of comparison. Id. at 1019; Cellular Rulemaking, 86 F.C.C.2d at 502. The FCC found that MCI proposed to serve over 200 square miles, or 13.6% more than CMS, and would include 33,713 more people.28 Pittsburgh, 96 F.C.C.2d at 1019. In assessing CMS's challenge that these differences are inconsequential, the FCC acknowledged that proper analysis required examination of more than raw numbers. In examining other factors declared relevant in Cellular Rulemaking, namely, presence of densely populated areas, highways, and areas likely to have high mobile usage characteristics, the FCC found further support for MCI's preference in its more comprehensive highway coverage and in its "pockets" or non-common areas. Id. at 1019-20.29 MCI's pockets also reflected a higher level of demand. Id.30

The FCC likewise affirmed MCI's preference for its determination and analysis of demand. As the ALJ had concluded, the Commission deemed certain of CMS's assumptions in its demand studies to be materially flawed. For example, CMS's assumption that paging is a product precursor of cellular radio significantly limited its sampling; while reasons for the decision to measure paging demand were indeed advanced, CMS had failed, the FCC determined, to establish a "direct link" between paging demand and cellular demand. Id. at 1023. The FCC affirmed the ALJ's finding that the record did not demonstrate that CMS employed its survey results to develop demand projections and distribution; the Commission stated that "even if we were not to find flaws in CMS's studies, we still would not credit it with reasonable market projections." Id.31

In contrast, MCI's assumptions and methodology were found to be clearly stated and free of serious error. In response to CMS's attack on MCI's survey assumptions, the Commission observed generally that "any market demand study must contain certain assumptions which are not susceptible to proof in order to account for all relevant factors." Id. at 1023-24. Nonetheless, MCI's forecasts were deemed more reliable because the scope of its CGSA was determined on the basis of its survey data, as were its frequency and channelization plans.32 MCI's demand projections also considered costs, population and industrial development, expected demographic changes, and likely usage patterns.

On the other hand, CMS's challenge to the ALJ's findings with respect to its ability to accommodate demand was, to a great extent, successful. Importantly, the FCC reversed the ALJ's conclusion regarding CMS's design and rejected the finding of potential co-channel interference. Citing the judge's findings that CMS would be able to accommodate first-year demand, the Commission concluded that the record failed to show CMS's inability to accommodate demand. The Commission did not, however, reverse MCI's preference on this issue; instead, the FCC downgraded it to a slight preference, finding that only MCI has "designed its system, distributed demand throughout the MSA, and determined its channel allocation plans based on its demand distribution forecasts." Id. at 1026.33

Overall, the FCC reduced MCI's award under Issue (a) from a substantial to a slight preference. This result was based primarily on the Commission's finding that CMS and Celcom had been "unduly penalized" by the ALJ. Id. at 1027, 1036.

So too, CMS successfully persuaded the Commission to vacate the preference awarded MCI under Issue (b). Specifically, the FCC emphatically rejected the ALJ's implication that CMS had not proposed a bona fide cellular design, emphasizing:

[W]e do not favor one system design over another.... In adopting its cellular rules and policies, the Commission sought to avoid imposing any rigid system design concepts, in order to permit applicants the flexibility (within certain parameters) to design their systems in different ways.... We will not penalize CMS merely because it chooses not to follow the conventional wisdom. Nor will CMS's system be disfavored because it may require more fine-tuning than other systems.

Id. at 1029 (emphasis added). The Commission observed that any cellular system was expected to require some adjustment in response to usage patterns or availability of choice transmitter sites. Id. at 1029 & n. 21. The FCC discerned no evidence supporting the finding that intrasystem interference would plague CMS's initial system, adding that later intra-system interference did not warrant a demerit. Id. at 1030. Moreover, the Commission vacated the finding that CMS, in accomplishing frequency reuse, might impermissibly have to reduce the coverage of its CGSA. Additionally, the Commission rejected the significant distinctions drawn by the ALJ as to the various proposals to monitor system traffic congestion, stating that MCI's proposal under Issue (b) was, in the end, neither better nor worse.

At the same time, the FCC rejected CMS's boosterism-style claims that its system design was, in fact, superior to those of its two competitors. In the FCC's view, CMS failed to demonstrate that service to portable telephones and in-building service would be inadequate without CMS's proposed "depth of coverage" system. Id. at 1030. While the MCI and Celcom proposals appeared to be more flexible, the CMS proposal apparently offset this slight advantage by being "slightly more detailed." Id. at 1031. Thus, all of the proposals were deemed adequate, but none contained sufficient detail to warrant a preference. Ultimately, no preference was awarded under Issue (b), because "the record [did] not establish that any single applicant's expansion plan will be able to accommodate the expected growth in demand for service while the others will not." Id.

CMS was also denied the additional preferences which it sought under Issue (c).34 Although CMS retained a preference for its rate structure (because, as the ALJ concluded, it promoted usage-sensitive, cost-based pricing), the preference was downgraded to slight. Id. at 1033.353

Turning to the various procedural issues raised by the two unsuccessful contenders, the FCC concluded that no demonstration had been made that the issues on which cross-examination was denied were "decisionally significant." Id. at 1035. Moreover, the issues on which CMS was seeking surrebuttal, by way of reopening the record, were reversed by the Commission. In consequence, the FCC concluded that any prejudice suffered by CMS was thereby eliminated. Id.

Ultimately, the FCC's modifications greatly reduced the disparity discerned by the ALJ between MCI and CMS. As we have seen, MCI's preference under Issue (a), the major issue, was diminished from substantial to slight; under Issue (b), MCI lost its preference altogether; and CMS's preference under Issue (c), explicitly made the weakest of the three, was downgraded from a preference to a slight preference. Nonetheless, the comparative difference between MCI's slight preference under Issue (a) and CMS's slight preference under Issue (c) was deemed sufficient to warrant a reasoned choice.36 The award to MCI was therefore affirmed as modified. Id. at 1036.

Celcom promptly filed a Petition for Reconsideration. CMS did not seek reconsideration, although it filed a reply to Celcom's petition. In August 1984, the FCC released a Memorandum Opinion and Order acknowledging this to be a "close case" but affirming its Final Decision on all points. MCI Cellular Telephone Co., 56 Rad.Reg.2d (P & F) 936 (1984) (Pittsburgh Reconsideration Order).

The Commission's proceedings and disposition have been described at some length because we are obliged carefully to insure that the comparative process employed in this "lead proceeding" complied with applicable standards of law. Our review in this respect begins with the threshold requirement of reasonable and fair administrative procedures, that is, fair notice and adequate opportunity for the parties to present their respective cases. CMS strenuously protests that the ALJ's denial of cross-examination and surrebuttal constituted an abuse of discretion.37 Additionally, CMS claims that it utterly lacked notice as to the Commission's intention to consider as a point of comparison the reliability of cellular applicants' demand surveys (under Issue (a)).

As we have previously observed, streamlined procedures were adopted by the Commission for the explicit purpose of expediting cellular proceedings. Cellular Rulemaking, 86 F.C.C.2d at 501.38 Recognizing this, CMS contends that the specific scheduling pressures imposed by the FCC had a destructive effect on the proceeding's fairness and rationality; the rush to judgment, CMS believes, vitiated in this proceeding fundamental precepts of procedural fairness and fatally infected the purported reasonableness of the Commission's award to MCI.

Upon analysis, however, we cannot fault the Commission for its entirely proper concern with hastening long-delayed service to the public. The intended rapid deployment of cellular radio communications has long been plagued by a glacial regulatory pace. It was thus entirely reasonable for the FCC to seek to avoid the specter of protracted administrative litigation which threatened to stall yet further the seemingly endless regulatory process.

It is important to recall in this respect that Congress has expressly empowered the Commission to "conduct its proceedings in such manner as will best conduce to the proper dispatch of business and to the ends of justice." 47 U.S.C. Sec. 154(j) (1982). The Supreme Court has previously observed that the Communications Act is a "supple instrument for the exercise of discretion by the expert body which Congress has charged to carry out its legislative policy." FCC v. Pottsville Broadcasting Co., 309 U.S. 134, 138, 60 S. Ct. 437, 439, 84 L. Ed. 656 (1940). True to the teaching of Pottsville Broadcasting, courts have since recognized that "[a]s technology develops and the field of communications changes, procedural, as well as substantive, policy must be flexible." Bell Telephone Co. of Pennsylvania v. FCC, 503 F.2d 1250, 1265 (3d Cir.1974), cert. denied, 422 U.S. 1026, 95 S. Ct. 2620, 45 L. Ed. 2d 684 (1974); see RCA Global Communications, Inc. v. FCC, 559 F.2d 881, 885-86 (2d Cir.), on reh'g, 563 F.2d 1 (2nd Cir.1977), subsequent appeal, 574 F.2d 727 (2d Cir.1978).

* The procedural standards followed by the Commission in devising its cellular rules are spelled out both statutorily (in the Communications Act and the APA) and in the FCC's Report and Order, Cellular Rulemaking. The Communications Act requires that mutually exclusive radio license applications receive a "full hearing." 47 U.S.C. Sec. 309(e) (1982). That statute elaborates on the "full hearing" requirement by mandating that all interested parties be permitted to participate and by imposing on the applicant the burdens of proof and going forward with the evidence. "Full hearing," however, does not vouchsafe an inalienable right to cross-examination or surrebuttal. The Supreme Court has construed section 309 to mean that every party

shall have the right to present his case or defense by oral or documentary evidence, to submit rebuttal evidence, and to conduct such cross-examination as may be required for a full and true disclosure of facts.

U.S. v. Storer Broadcasting Co., 351 U.S. 192, 202, 76 S. Ct. 763, 770, 100 L. Ed. 1081 (1956) (emphasis added). The Act does not, by its terms, require a traditional, trial-type evidentiary hearing at all;39 quite to the contrary, the Commission could, under settled law, determine that the "demands of the public interest were so urgent as to preclude the delay which would be occasioned by a hearing." Ashbacker Radio Corp. v. FCC, 326 U.S. 327, 333, 66 S. Ct. 148, 151, 90 L. Ed. 108 (1945). See supra p. 5. The watershed decision in Ashbacker requires a meaningful comparative hearing and a reasoned determination of the public interest; it does not inexorably require a full-blown, trial-type proceeding. Thus, in United States v. FCC, 652 F.2d 72, 88-96 (D.C.Cir.1980) (en banc ), an analogous case, this court upheld FCC satellite licensing without formal-type evidentiary hearings because of the public interest in rapid institution of service.

* In addition to the Communications Act, the APA expressly authorizes "paper hearings" in licensing cases when a party will not be prejudiced by that procedure, 5 U.S.C. Sec. 556(d) (1982). Cross-examination is therefore not an automatic right conferred by the APA; instead, its necessity must be established under specific circumstances by the party seeking it. The APA, echoed by Storer Broadcasting, provides only for "such cross-examination as may be required for a full and true disclosure of the facts." Id. In this case-by-case analysis, cross-examination is appropriately denied if the party fails to "point to any specific weakness in the proof which might have been explored or developed more fully by that technique than by the procedures adopted by the Commission," or fails specifically to "suggest what questions were necessary" to explore the general issues to be examined, or fails to explain why written submissions, including rebuttal material, were ineffectual. Absent such a showing, no prejudice has been established. American Public Gas Association v. Federal Power Commission, 498 F.2d 718, 723 (D.C.Cir.1974).

In exercising the broad discretion conferred by Congress, the FCC has repeatedly held that cross-examination is not essential in all cases; instead, this traditional device of truth-finding is deemed to lie within the discretion of the ALJ, whose decision will not be reversed by the Commission if it reflects sound discretion.40 Even if a party successfully argues under this standard that cross-examination should have been permitted, the second inquiry is whether the matters sought to be cross-examined bore significantly upon the Commission's factfinding and drawing of conclusions. If not, then the party has not been prejudiced. Deep South Broadcasting Co. v. FCC, 347 F.2d 459, 463 (D.C.Cir.1965).

As we have seen, the standards established by the FCC in Cellular Rulemaking left the availability vel non of cross-examination to the "sound judicial discretion of the judge, the prevailing standard being whether the person requesting cross-examination has persuasively demonstrated that written evidence is ineffectual to develop proof." Order on Reconsideration, 89 F.C.C.2d at 92; 47 C.F.R. Sec. 22.916 (1984) (citations omitted). What is more, both the Designation Order and the Pre-hearing Order expressly called the parties' attention to this requirement. MCI Cellular Telephone Co., FCC 82M-4023, at 2 (Dec. 17, 1982) (Order Prior to First Hearing Session); Pittsburgh Designation Order, 52 Rad.Reg. at 1111; see RCA Global Communications, 559 F.2d at 886; American Public Gas Association, 498 F.2d at 723; Order on Reconsideration, 89 F.C.C.2d at 92 n. 59. Under the applicable orders, moreover, surrebuttal might be permitted if the ALJ finds it is "needed to provide a complete record." Order on Reconsideration, 89 F.C.C.2d at 92; MCI Cellular Telephone Co., FCC 82M-4120, at 2 (Dec. 28, 1982) (Order).

In the face of clearly articulated standards governing when cross-examination may be available, CMS failed, in our view, to satisfy these prerequisites. Indeed, CMS itself admitted that cross-examination would be an ineffective method of developing the record. CMS previously argued to the ALJ that the use of cross-examination "to elicit the facts necessary to complete the record ... is an inefficient means of highlighting contested areas; it is not conducive to the development of an evidentiary record adequate to support a meaningful analysis of the comparative merits of competing proposals." CMS Request for Prehearing Conference, at 3 (Dec. 10, 1982).

Thus, it cannot seriously be argued that the ALJ improperly closed the door on these procedures in advance of the hearing. Rather, the ALJ dutifully followed the Commission's clear pronouncements in this respect.

CMS nonetheless challenges the denial of cross-examination as to two issues: (1) the reliability of its market demand analysis, CMS Brief at 30-32; and (2) the flexibility of its system design and intra-system interference. Id. at 63-64.

The first issue is indeed significant because the award of two preferences to MCI under Issue (a), namely, the determination of demand and the ability to accommodate demand, hinged on the reliability of the parties' market-demand analysis. Aside from its clear-cut greater geographic and population coverage, it will be recalled that MCI was preferred under Issue (a) for the stated reasons that (1) CMS had failed adequately to explain its market-research methodology, Pittsburgh, 96 F.C.C.2d at 1022-23; (2) MCI had clearly established a nexus between its demand forecasts and system design, id. at 1026; and, conversely, (3) CMS had failed to demonstrate such a nexus. Pittsburgh Reconsideration Order, 56 Rad.Reg.2d at 940.

We disagree that cross-examination would likely have altered the decisionmaker's views with respect to the ALJ's first and third reasons. In the first instance, cross-examination is not the appropriate office for articulating one's own methodology in comparative cellular proceedings, nor is it an appropriate means of explaining how one's design is based on market research. These matters are, instead, properly the subject of an applicant's direct case and, if inadequately fleshed out from a comparative perspective, of rebuttal. In our view, CMS failed adequately to employ the mechanisms provided under the FCC's procedures. Moreover, cross-examination might, at best, alter the probative force of MCI's asserted nexus, as demonstrated in its direct case, between market research and system design-planning. Yet this sort of attack, especially from a comparative focus, was precisely the purpose of the rebuttal-case procedures established by the Commission as routine in comparative proceedings. Besides these broader difficulties with its position, CMS in the heat of comparative battle did not even seek to cross-examine the MCI expert who was responsible for the latter's market research exhibits. CMS Objections to Direct Exhibits and Requests for Cross-Examination, at 18-19 (Jan. 11, 1983).41

CMS also challenges the denial of cross-examination of Mr. Whitty (MCI), and Messrs. Larson and Rose (Celcom) regarding their testimony on intra-system interference and flexibility under Issue (b). CMS Brief at 63-64. The requests, however, lacked the requisite specificity and made no realistic attempt to demonstrate that written submittals were ineffectual.42 In any event, the issue of intra-system interference was pretermitted by the Commission's reversal of the ALJ on the merits of that question; CMS thus cannot be heard to claim that denial of cross-examination on that particular issue was still prejudicial.

CMS next asserts another procedural infirmity--that the ALJ's rulings denying surrebuttal precluded all opportunity to respond to the other applicants' rebuttal criticisms of CMS's proposal. CMS contends that surrebuttal on the issues of market research, under Issue (a), and flexibility, under Issue (b), was necessary to overcome the denial of cross-examination and the alleged lack of notice to applicants of the decisional relevance of such studies.

As with cross-examination, no automatic entitlement to surrebuttal lies in comparative proceedings. Rather, the cellular rules vest discretion in the ALJ to permit or deny such additional submissions. The prescribed commonsensical standard authorizes surrebuttal when the ALJ finds it is needed to complete the record. Order on Reconsideration, 89 F.C.C.2d at 92; 47 C.F.R. Sec. 22.916(b)(6) (1984). In this instance, CMS's initial request for surrebuttal was tendered prior to the filing of rebuttal submissions, CMS Request for Prehearing Conference, at 3-4 (Dec. 10, 1982); indeed, the request was made prior to the parties' objections and request for cross-examination. Given this early-bird timing, CMS's initial request was necessarily a mere generalized request for a third round of pleadings, rather than a specifically crafted request addressed to the rebuttal submissions. As a result, the ALJ could not realistically have peered into the future and determined at such an early juncture that additional testimony would in fact be needed. Moreover, as with cross-examination, Judge Ehrig did not unconditionally deny all surrebuttal at this stage, see Order (Dec. 17, 1983), but instead left the question to future resolution.43

CMS's first specific request for surrebuttal occurred at the hearing itself. Counsel for CMS admitted that the record was adequate to permit argument and briefing on whatever points CMS needed to make, Transcript at 68, with the exception of only one unanswered allegation, that of intra-system interference. Id. at 69-70. The two other applicants, however, expressed reluctance to craft and be besieged with yet another round of evidentiary submissions. Id. at 70-71.

Given CMS's inadequate justifications, Judge Ehrig expressly found that the record was sufficient to permit a considered judgment without a third round. We cannot say, on this record, that this determination constituted an abuse of discretion. But at all events, we would once again be unable to find that CMS was prejudiced, inasmuch as the Commission subsequently rejected both the ALJ's findings as to the purported intra-system interference and since the subject itself was one of negligible import to the evaluation under Issue (b).

CMS's next attempts to obtain surrebuttal were set forth in two motions to Reopen the Record, one filed over four months after the record closed and after the ALJ's Initial Decision. Both motions sought to address only the issue of intra-system interference, which, as we have seen, was later washed out by the Commission. CMS never requested surrebuttal at any time--either before or after the Initial Decision--on the issues of market demand studies or flexibility, the gist of its complaint here.

Moreover, CMS's proffer in this respect hardly presented the sort of newly-discovered evidence or compelling circumstances warranting re-opening a record after oral argument and initial decision.44 The Commission has long held that, absent these unusual circumstances, a record should not be reopened "merely to permit an applicant an additional opportunity to present evidence which has been continuously available in order to make a better case for a grant of its application." Heart of Georgia Broadcasting Co., 19 F.C.C.2d 20, 30-31 (Rev.Bd.1969).

CMS now claims that, had surrebuttal been permitted, it would have shown that its system was no less flexible than the others; that it had incorporated several features designed to enhance its flexibility to accommodate future demand; and that, if anything, its system was actually more flexible than the others. CMS Brief at 64-65. We are persuaded, however, that these are precisely the showings that properly belonged in CMS's direct case, in claiming a preference under Issue (b), or in its rebuttal case, when it comparatively critiqued the two competing systems and sought to demonstrate its own superiority.

Given the myriad of opportunities the parties had to present evidence, in direct and rebuttal cases and in argumentation before the ALJ (in Petitions to Deny the other applications, supplemental memoranda,45 Proposed Findings and Conclusions, Reply Findings and Conclusions, Exceptions, and Replies to Exceptions) and before the Commission on de novo review (Opening Briefs and Replies, oral argument, Petitions for Reconsideration and responsive pleadings, two Motions to Reopen the Record and responsive pleadings, and sundry other motions), the opportunity to respond provided by the Commission appears entirely adequate under applicable legal standards. CMS vigorously took full advantage of these opportunities; indeed, it was successful in persuading the Commission to vacate many of the ALJ's key findings which were allegedly derived from damaging rebuttal testimony.

In sum, an observation from another proceeding seems entirely a propos to the case at hand: challenges such as the various procedural claims advanced here "reflect little more than the boundless litigiousness of disappointed claimants." Christian Broadcasting Network, Inc. v. Copyright Royalty Tribunal, 720 F.2d 1295, 1319 (D.C.Cir.1983).

In addition to its quarrels over the lack of opportunity for cross-examination and surrebuttal, CMS claims that it was sandbagged by the award of any preference for the relative reliability of the applicants' demand studies. CMS protests vehemently of a lack of notice that this matter would be the subject of comparative evaluation.

Obviously, it is a fundamental tenent of the administrative process that parties to comparative proceedings are entitled to notice of the decisional criteria and burdens of proof. As this court has observed:

It is beyond dispute that an applicant should not be placed in the position of going forward with an application without knowledge of requirements established by the Commission, and elementary fairness requires clarity of standards sufficient to apprise an applicant of what is expected.

Bamford v. FCC, 535 F.2d 78, 82 (D.C.Cir.), cert. denied, 429 U.S. 895, 97 S. Ct. 255, 50 L. Ed. 2d 178 (1976). The question is whether that statutory principle was trodden asunder in the proceedings in this case.

* To answer that question, we are obliged to return briefly to the beginning of this controversy. The Designation Order in this case, issued in December 1982, trifurcated comparative Issue (a) into separate sub-issues. The ALJ was first to determine and compare the geographic area and population which each applicant proposed to serve; second, the ALJ was "to determine and compare the relative demand for the services proposed in said area," and third, the ALJ was to determine and compare the applicants' respective abilities to accommodate anticipated demand. Pittsburgh Designation Order, 52 Rad.Reg.2d at 1110. While this order indisputably gives notice that the parties' demand determinations were to be compared separately, it was issued five months after the direct cases were filed. CMS contends that the Designation Order departs from the comparative standards previously announced in Cellular Rulemaking by making demand studies the subject of a separate preference and by basing the evaluation of an applicant's "ability to accommodate demand" on the extent to which system planning reflected demand. CMS Brief at 51-52. CMS suggests that the full extent of this "newly announced" burden was unknown until the Commission affirmed the Initial Decision. Id. at 54-55.

Contrary to CMS's adamant assertions, however, the Report and Order eliminated demonstration of public need only from the basic qualifications to be determined prior to the comparative proceeding. Cellular Rulemaking, 86 F.C.C.2d at 502. Instead, "the comparative criteria ... include a consideration of the public need each applicant is likely to serve." Id. (emphasis added). The Report and Order plainly included among "other significant factors" to be considered in comparing proposed service areas "indications of a substantial public need for the services proposed, including the results of public need surveys." Id. (emphasis added). That same document made it clear that "the public need indicated by subscriber surveys" would be relevant. Id. at 503.

In the face of what appears to us to be abundantly clear language, CMS argues that these references are to be understood as treating demand studies as just one kind of evidence which would be entertained by the Commission. CMS Brief at 52. Specifically, CMS maintains that the "most reasonable reading ... is that an applicant could demonstrate need entirely by other means, without recourse to surveys at all." Id. at 53 n. 51.

At the same time, CMS admits that the rulemaking gave ample notice that applicants were to demonstrate the relative efficacy of their proposed service areas to meet demand. CMS Brief at 52. As we have seen, the second major criterion established by the rulemaking (Issue (b)) was the applicants' prospective abilities to expand their systems "in a coordinated manner ... in order to serve an increasing number of local subscribers and roamers as demand warrants." Cellular Rulemaking, 86 F.C.C.2d at 502-03. This criterion obviously incorporates the requirement that a system design was to be based on demand. Thus, CMS had notice--and appears to acknowledge as much--that demand was indeed to be determined by the parties and utilized, at the very least, in comparing service areas and expansion ability.46

But if there were any lingering doubt as to CMS's being adequately provided with notice in this respect, that doubt is dissipated by the veritable floodtide of evidence in CMS's own pleadings and exhibits; that evidence powerfully suggests that CMS both invited and fully expected comparison of demand methodologies and an examination of the nexus between the parties' demand surveys and their respective designs. By its own account, CMS invested "considerable resources" in no less than six demand surveys. Its voluminous treatment of demand methodology and analysis comprised over 1,000 pages in Volumes II and III of its direct case exhibits. In one exhibit, CMS introduced its "extensive" demand methodology and analysis by noting that the key factors involved in the determination of demand cannot "be addressed with simplistic analytic approaches." CMS Direct Exh. II, at 12. For this reason, CMS presented several demand-determination methods.

Equally important, CMS did not challenge the relevancy of opponents' exhibits which detailed their respective demand methodologies. Quite to the contrary, CMS vigorously objected to Celcom's demand exhibits on substantive grounds. CMS then proposed to cross-examine Celcom's witness on the methodology employed to design, conduct, and analyze Celcom's demand survey. CMS Objections to Direct Exhibits and Requests for Cross-Examination, at 4, 6, 16-17 (Jan. 11, 1983).

Finally, and tellingly, CMS's Rebuttal Exhibit V expressly proclaimed in its opening sentence that "[t]he Commission has indicated that subscriber surveys evidencing public need for cellular service are a potentially important factor in a comparative consideration of applicants' cellular proposals." The 22-page exhibit was devoted to a detailed comparative analysis of Celcom's and MCI's demand methodologies. CMS's words in its opening paragraph are starkly revealing in light of its present attack: "[d]eficiencies in these two applicants' methodologies and in the results generated thereby will be highlighted. As will be developed, essential aspects of these applicants' demand surveys, demographic analyses, and calculations of demand are subject to serious question." CMS Rebuttal Exh. V, at 1 (Jan. 27, 1983). CMS then articulated a long litany of asserted errors and omissions, such as Celcom's sampling methods and assumptions. Id. at 1-16. Nor was Celcom alone in drawing CMS's fire; to the contrary, CMS sharply criticized MCI's asserted failures to screen survey respondents; to provide quantitative support for its "Intent Translation Scale"; to provide a quantifiable basis for its subscriber projections; and to explain the methodology used in applying various demand-influencing variables. Id. at 16-21. After subjecting its competitors' demand studies to the foregoing fusillade, CMS concluded with the criticism that neither of its worthy competitors presented a statistically valid methodology. The suggestion was that the unreliability of Celcom's and MCI's demand estimates cast doubt upon the ability of their respective system designs to meet likely demand. Id. at 21-22.47 Given the sheer volume and vigor of CMS's submissions in this respect, the explanation that its own inclusion of this information was merely "fortuitous" is entirely farfetched. See CMS Reply Brief at 17.

Our resolution of CMS's various procedural attacks brings us to its second major contention, namely that the FCC's application of its comparative standards resulted in a decision that is arbitrary, capricious and unsupported by substantial evidence. It is well-settled, of course, that the essential elements of a valid comparative choice are that the bases for the agency's final conclusion be clearly stated; that the findings be sufficient rationally to support the conclusion reached; that the findings be supported by substantial evidence; that findings cover all substantial differences between the parties; and that the final decision be based upon a composite consideration of all of the findings as to each applicant. Greater Boston Television Corp. v. FCC, 444 F.2d 841, 851 (D.C.Cir.1970), cert. denied, 403 U.S. 923, 91 S. Ct. 2233, 29 L. Ed. 2d 701 (1971); Johnston Broadcasting Co. v. FCC, 175 F.2d 351, 357 (D.C.Cir.1949). Of special pertinence to our case, in circumstances where competing proposals are so similar as to create a standoff, a slight difference with respect to a single factor of substantial import may fairly determine the outcome. Greater Boston Television Corp., 444 F.2d at 862-63.

Importantly, CMS has not challenged--either on procedural or substantive grounds--the FCC's primary determination under Issue (a) that MCI's proposal entailed greater geographic and population coverage.48 This factor was repeatedly emphasized by the Commission to be the predominant criterion, the "key" to the Commission's regulatory structure for cellular services. See, e.g., Pittsburgh, 96 F.C.C.2d at 1018; Cellular Rulemaking, 86 F.C.C.2d at 502, 509. Indeed, as we alluded to before, virtually any other aspect of a cellular applicant's facilities or services may be modified without substantial regulatory oversight; a change in the boundaries of coverage, however, is deemed so significant as to be considered a "major application" equivalent to an entirely new application. Cellular Rulemaking, 86 F.C.C.2d at 509. From both the language of Cellular Rulemaking and the Designation Order, it would appear that the coverage factor outranks not only the second and third sub-issues under Issue (a), but surpasses Issues (b) and (c) in importance as well. Indeed, a cursory examination of subsequent cellular proceedings reveals that, in all save one, the applicant which proposed greater geographic coverage was awarded the license.49

In this proceeding, CMS does not argue either that the coverage factor was overvalued or that the Commission erred in determining that MCI's proposal proposed greater coverage. Moreover, the Commission did not limit its determination to a simple comparison of numbers, but as we have previously observed, the FCC examined the other factors articulated by Cellular Rulemaking as significant--the presence of densely-populated regions, highways, and areas likely to have high mobile usage--as reflected in the applicants' non-common areas or pockets. Instead, CMS's challenge focuses upon the second and third sub-issues of Issue (a), and the second major issue (Issue (b))--expansion of the system.

* The first of CMS's principal substantive attacks on the FCC's decision relates to the determination of demand, the second sub-issue under Issue (a). CMS's challenge is twofold: (1) that the Commission erred in finding that CMS had failed adequately to explain its demand methodology, and (2) that the Commission's findings on this issue conflict with the findings in the subsequently rendered Portland decision.50 As to the first issue, CMS complains that it was denied a preference not because its demand estimate was erroneous, but because its proposal contained insufficient detail. This, it claims, is improper. CMS Brief at 54-55. We find this attack singularly unpersuasive. Sophisticated applicants such as CMS would be rather naive to believe that the FCC would meekly accept its demand studies as valid without close scrutiny of the methodology employed. Indeed, the FCC could realistically undertake this entire exercise only by examining assumptions and methodology to determine whether an applicant's demand forecast was likely to be valid. Absent an appropriate methodological demonstration, the applicant's proposed demand would likely be "unrealistic." See Houston Mobilfone, Inc., 65 F.C.C.2d 848 (Rev.Bd.1977), review denied, FCC 78-477 (July 11, 1978), remanded sub nom. Mobilfone Service, Inc. v. FCC, 605 F.2d 572 (D.C.Cir.1979), aff'd on remand, Houston Mobilfone, Inc., 78 F.C.C.2d 1067 (1980). Moreover, CMS's demand predictions were significantly higher than those of MCI and Celcom; in fact, for each of the first five years, CMS optimistically forecasted more than twice the demand prophesied by its competitors. See supra note 20. Given this enormous discrepancy, it was reasonable for the Commission to look closely at all of the parties' respective analyses and methodologies in evaluating which applicant was not only most capable of reliably determining demand but which had actually done so.

Specifically, as we have seen, the Commission rejected CMS's bedrock assumption that paging is a precursor to cellular radio, on the ground that this assumption curtailed, without adequate explanation, the sample group. While the Commission acknowledged that CMS had advanced various reasons for this assumption, CMS had nonetheless "failed to establish a direct link" between paging demand and cellular demand. Pittsburgh, 96 F.C.C.2d at 1023.

We find that the record evidence now highlighted by CMS provides precious little support for its assumption. See CMS Brief at 33-34 n. 30. For example, Exhibit II, upon which CMS relies, simply sets forth the assumption, justifying it with the statement that "paging service has the necessary market history to serve as a precursor" because, unlike conventional mobile radio, paging service is driven by demand and is not constrained by fixed capacity. CMS Direct Exh. II, at 5, 12, 14. Data from the Westat survey is likewise invoked by CMS as empirical support for the claimed high correlation between cellular demand and paging demand. Upon analysis, however, that data appears equivocal; CMS's own non-paging-limited research suggests that almost half the demand for cellular radio in Pittsburgh (and over half the demand in all CMS markets) would come from businesses which do not utilize paging services. MCI Rebuttal Exh. 3, at 16; CMS Direct Exh. II, Vol. 3, Tables 3-7A to -8C. According to its own evidence, CMS's demand estimates were based wholly on a "need factor" derived from the Claritas survey, which analyzed paging demand exclusively, and Graphic Scanning Corporation's paging experience. While it is true that two of CMS's studies were addressed to the general business population, not simply to paging users, these surveys were admittedly used only to "verify" the modeling results. Only the Claritas paging survey appears to have affected CMS's final projections of demand. CMS may have surveyed the general business population, but it did not include them in its effective sample group.51

The most critical finding, also undisputed by CMS, was the FCC's rejection of CMS's "need factor" or "cellular penetration factor." CMS projected a demand of 13,584 cellular units. CMS Direct Exh. II, at 14. Yet, its Claritas study projected a demand of 20,525 pagers in the Pittsburgh SMSA; notwithstanding this discrepancy, nowhere is there forthcoming a direct explanation of how demand for 20,525 pagers translates into demand for 13,584 cellular units. CMS did not obtain its demand prediction from cellular demand surveys aimed at the Pittsburgh market. Instead, CMS used the same model utilized in all of the major markets in which it applied. Although assigned a need factor of 6 units per 1,000 population, Pittsburgh's actual need could have been anywhere from 5.25 to 6.75 units per 1,000 population, a range which when multiplied by the Pittsburgh population would yield demand estimates varying from 11,886 units to 15,282 units. The choice of a need factor of 0.006, although undoubtedly convenient to CMS, thus appears to have been wholly arbitrary.52

In sum, we cannot quarrel with the Commission's finding of insufficient explanation in the record as to how CMS derived its demand determination or what relevance its volumes of survey data had to its determination. Nor can the FCC be gainsaid for questioning whether CMS's resulting demand figures were sufficiently reliable to provide a reasonable basis for shaping a CGSA and designing a cellular system which could satisfy demand for its cellular services in that market.

Moving to a different tack, CMS contends that the Commission imposed different standards of proof on the applicants. Upon careful review, however, we can divine no disparity in treatment of the competing candidates. The only standard applied by the FCC to assumptions underlying the parties' surveys and their respective determinations of five-year anticipated demand was that of reasonableness. MCI satisfied this standard, the FCC concluded, by a clear explication of assumptions found upon analysis to be reasonable. CMS's assumptions were, in contrast, determined by the agency to be unreasonable; had those assumptions been properly supported, however, they might well have satisfied the applicable standard.53

CMS also devotes considerable argument to the alleged inconsistency among the various cellular proceedings, suggesting that the findings regarding CMS's assumptions in the subsequent Portland proceeding do not comport with the findings in the instant proceeding. This challenge, however, puts the cart before the horse. Whether the Pittsburgh and Chicago cases are inconsistent with subsequent cellular proceedings, as CMS claims, is in the end a question for another day. The Commission has selected the Pittsburgh and Chicago decisions as its standard-bearers. Pittsburgh, 96 F.C.C.2d at 1017 n. 4; MCI Cellular Telephone Co., FCC 83-580, at 2 (Dec. 9, 1983) (Order). They must stand or fall on their own merits; it is the subsequent cases which must measure up to the Pittsburgh and Chicago precedents. The Commission will obviously be held to these precedents in subsequent cases, barring of course an adequate explanation of departures from binding precedent. Should elements of subsequent decisions be found to lack fidelity to these two decisions, and if that the infidelity is decisionally significant, then those decisions will naturally be subject to challenge.54

Moreover, without adversarial briefing and the full records of the other proceedings before us, this court is singularly ill-equipped to invalidate such highly factbound determinations on grounds of inconsistency with unreviewed agency decisions. CMS does not in reality argue that the entire pattern of FCC decisions in the nonwireline proceedings is a crazy quilt; instead, CMS argues that in some of the cases the FCC has applied different criteria, standards, and burdens of proof to different applicants. While this is a serious charge, it is one that is difficult to evaluate in this first, precedent-setting proceeding, especially given the Commission's effort to maintain consonance among the various proceedings, as by including specific textual discussions of these lead cases in later cellular decisions.55 See, e.g., Denver Cellular Telephone Co., FCC 85-18, at 12 n. 24, 26 (Jan. 31, 1985), reconsideration pending, FCC No. 83-620 (Mar. 4, 1985), appeal docketed sub nom. Denver Cellular Telephone Co. v. FCC, No. 85-1137 (D.C.Cir. Mar. 4, 1985); Portland, 56 Rad.Reg.2d at 1371-73 & n. 24; Cf. Airmark Corp v. FAA, 758 F.2d 685 (D.C.Cir.1985). Instead, we must evaluate whether substantial evidence in this record supports the agency's decision. Indeed, that is all we can do given that not all of the subsequent records are before this court. If such support exists here, it is not relevant that the FCC might have adopted a different conclusion at a later time. Cf. Consolo v. Federal Maritime Commission, 383 U.S. 607, 620, 86 S. Ct. 1018, 1026, 16 L. Ed. 2d 131 (1966) (drawing two inconsistent conclusions from the evidence does not vitiate substantial evidence support).

CMS' second major challenge is directed at the Commission's award of a slight preference to MCI (and no preference to CMS) under the third sub-issue of the coverage criterion, namely "to determine and compare the ability of each applicant's cellular system to accommodate the anticipated demand for both local and roamer service." Pittsburgh Designation Order, 52 Rad.Reg.2d at 1110. As we have seen, CMS has already met with substantial success in persuading the Commission to reverse many of the ALJ's adverse findings, but before us, CMS argues that the Commission failed to go far enough.

To recap briefly, Judge Ehrig's reasons for awarding this preference to MCI were twofold: first, only MCI's proposal reflected an understanding of the anticipated demand in the Pittsburgh SMSA; second, CMS's proposed "depth of coverage" system violated what the ALJ deemed to be fundamental principles of cell spacing design, creating the potential for co-channel interference and requiring significant re-tuning over time. The ALJ, third, awarded a slight preference to MCI for its proposal to accommodate roamer traffic. The Commission, it will be recalled, sustained the ALJ's first holding, but rejected outright the ALJ's other two conclusions.

CMS attacks this outcome on four grounds: (1) that the finding regarding MCI was based on factual error; (2) that the Commission failed to compare the applicants' respective abilities to accommodate demand; (3) that the failure to award a preference to CMS for its superior design was arbitrary and capricious; and (4) once again, that the Pittsburgh outcome was inconsistent with that in Portland.

CMS first contends that the FCC erred in finding that MCI's demand forecasts were relied on in the successful applicant's design and planning. The Commission expressly found that MCI designed its system, distributed demand, and determined its channelization plans based on its demand forecasts. Pittsburgh Reconsideration Order, 56 Rad.Reg.2d at 939. CMS, however, cites MCI's own evidence to the effect that the latter distributed demand among cells solely according to population density as proof that MCI's design was not in truth based on its demand determination. MCI Applic.Exh. 5, at 7; see also Pittsburgh, 96 F.C.C.2d at 1022.

This attack is unavailing, as it asks us to ignore the larger picture presented by the record. MCI's demand determination included many variables--such as demographic and geographic analysis and need surveys--in addition to population. MCI Direct Exhs. 28 & 29. MCI then admittedly distributed the demand among cells according to population density, MCI Applic.Exh. 5, at 7, but based upon a correlation, confirmed by its research, between population density and predicted traffic volume. MCI Direct Exh. 28, at 19. Thus, MCI's channelization plans were based on total subscriber forecast, derived from market research, which was then allocated among cells according to population density; the channel assignments were determined on the basis of the number of subscribers allocated, the anticipated peak channel load, and the desired grade of service. Pittsburgh, 96 F.C.C.2d at 1022; Pittsburgh Initial Decision, 96 F.C.C.2d at 1047-48; MCI Applic.Exh. 5, at 1-7. As we see it, the Commission committed no error in this respect. The record adequately supports the finding that MCI had specifically planned and designed a system to accommodate the demand it had forecast.

In contrast, the record is devoid of evidence that CMS's choice of service area and system design bore any relationship whatever to the perceived demand for its services. Pittsburgh, 96 F.C.C.2d at 1026; Pittsburgh Initial Decision, 96 F.C.C.2d at 1063. Instead, CMS employed a broad modeling approach irrespective of actual demand or population density. The CMS plan merely divided the city into two major areas, see Pittsburgh Initial Decision, 96 F.C.C.2d at 1050, a central core city area of seven to nine cells presumed to be high density, and an aureole of five cells presumed to be low density. The intermediate zone was comprised of two to four cells. CMS then assigned the same number of subscribers and the same number of channels to every cell within each area, varying the assignments only in the small intermediate zone. CMS Direct Exh. IV, at 3, 12-13. Cells 1 through 7 were all presumed to have 762 subscribers in the third year and were correspondingly assigned 31 channels each. CMS Direct Exh. IV, at 13. It thus appears that CMS not only developed an anticipated demand figure independent of and unrelated to its survey results, but also proceeded to design its system without regard to the demand identified through its surveys, to the demand derived from its model, or to the actual population in its chosen cells.

Apparently recognizing its inability to demonstrate that it designed a system aimed at accommodating anticipated demand, CMS now urges that the real issue is whether either system was better able to accommodate demand. CMS Brief at 39-40. Additionally, CMS objects that the Commission failed to make necessary supporting findings that CMS (1) had proposed an inferior system; (2) had inaccurately projected demand; or (3) had failed reasonably to tailor its geographic distribution to areas of likely demand. Id. at 54-55. If by "inferior system" CMS means system design, the Commission made its position quite explicit that design in itself would not be preferred or disfavored. CMS's own lack of support for its demand projection prevented the FCC from evaluating whether CMS's demand was accurate or not. As to the last objection, the Commission's entire justification for its award of a preference to MCI on this sub-issue was its finding that only MCI had tailored its system to the demand it had specifically identified in Pittsburgh.

CMS argues that no showing was made that the parties, in fact, proposed different channel distributions. This claim, however, is belied by specific evidence in the record, in both the Initial and Final Decisions, which reveals very different channel distributions, Pittsburgh, 96 F.C.C.2d at 1024-26; Pittsburgh Initial Decision, 96 F.C.C.2d at 1047-50, 1062-63; CMS Direct Exhs. II & IV; MCI Applic.Exh. 5; MCI Direct Exhs. 28 & 29, a difference on which CMS itself tries to capitalize in its claim for a preference for its purportedly superior "depth of coverage."

CMS next contends that, under the language of the Designation Order, a proper comparison of the relative abilities to accommodate demand must involve a Commission determination of the actual demand for cellular radio in Pittsburgh. CMS Brief at 38. In oral argument, counsel for CMS flagged the agency's failure to measure the systems against a single demand figure (or "true" demand) as the "essential failing" of this proceeding. CMS adds that no showing has been made that CMS's design would be incapable of satisfying either its own or MCI's demand forecast. CMS Brief at 39.56

While this argument is initially appealing, a fair reading of the Designation Order and Cellular Rulemaking reveals its error. The rulemaking Report and Order expressly included demand determinations among the comparative criteria, which were to include "a consideration of the public need each applicant is likely to serve," that is, the public need for the services proposed by that applicant. 86 F.C.C.2d at 502 (emphasis added). The Designation Order is no less clear in characterizing the issue as a comparison of "the relative demand for the services proposed." 52 Rad.Reg.2d at 1110 (emphasis added). Moreover, it would appear well nigh impossible objectively to determine a single common demand in a comparative context, inasmuch as the demand for an applicant's cellular radio service is dependent upon the chosen CGSA, the cost of service, and the nature and quality of the services offered. Such a determination would fly in the face of one of the Commission's avowed purposes, namely to encourage diverse approaches and to maximize service areas where justified by the level of demand. The Commission deliberately chose not to establish the service boundaries for all applications, nor to determine demand for each metropolitan area, nor to impose rigid design requirements; instead, the FCC placed the burden on the parties to present their respective proposals as to the best way to serve the Pittsburgh area, as well as the Commission's larger goals of nationwide compatible service and rapid implementation.

CMS again argues that the Pittsburgh decision must be remanded due to a fatal inconsistency with Portland, an issue we have already addressed generally supra pp. 53-54. Here, CMS claims that both MCI and CMS (and its Portland affiliate) each presented nearly identical proposals in the Pittsburgh and Portland proceedings. MCI was awarded a slight preference in Pittsburgh, but no preference was awarded in Portland. This presents an apparent factual inconsistency, but not necessarily an inconsistent application of standards. As we have discussed, see supra p. 54, all we can ask of the agency is that it apply standards consistently and that its conclusion under each standard be supported by the record in this proceeding. Here the Commission meets this requirement. In both Portland and Pittsburgh, the Commission required all applicants to demonstrate a nexus between demand assessment and design; in both cases the Commission found that CMS had failed to demonstrate this "critical factor" and denied a preference. Portland, 56 Rad.Reg. at 1375; Pittsburgh, 96 F.C.C.2d at 1026. Similarly, in both cases the FCC expected the applicants to provide a detailed showing of how demand was distributed among cells. In Portland, MCI apparently failed to make this showing; in contrast, in Pittsburgh, MCI's showing was found to be adequate. Portland, 56 Rad.Reg.2d at 1376; Pittsburgh, 96 F.C.C.2d at 1026. We have found that the FCC's conclusion in this case was supported by substantial evidence. See supra pp. 56-58. This is all we can and should do in this appeal.

Finally, CMS objects here, and again under Issue (b), to the denial of a preference for its purportedly superior "depth of coverage" design. From the first page of its description of the proposed Pittsburgh system, CMS Direct Exh. IV, to this appeal, CMS has pertinaciously argued that its high degree of cell overlap was not only advantageous (in the event of loss of a cell or saturation of a cell to capacity) but also resulted in superior service to "in-building" portable units. CMS argues that the Commission's decision is devoid of specific independent comparison of the applicants' respective portable service, charging that the Commission instead "summarily dismissed" the matter. We cannot agree. What appears to CMS to be a summary dismissal of a "critical factor" appears to us to be a summation of an extensive and well-reasoned examination of this issue, first by the ALJ and then by the Commission. The ALJ recognized the relevance of CMS's "depth of coverage" evidence, but she explicitly stated that it was relevant to the "ability to accommodate demand." Transcript at 32, 52. In all three decisions are to be found lengthy discussions of the pros and cons of cell overlap and cell spacing design; those discussions set forth the claims of CMS in this respect, the expression of doubts harbored by the ALJ over CMS's creativity in design, and the detailed findings and conclusions of both the ALJ and the Commission. Pittsburgh Reconsideration Order, 56 Rad.Reg.2d at 938; Pittsburgh, 96 F.C.C.2d at 1025-26, 1028-1030; Pittsburgh Initial Decision, 96 F.C.C.2d at 1050, 1054-55 & n. 13, 1062-63. When all was said and done, the Commission simply declined to favor any design absent the requisite comparative showing that superior coverage would result.

Substantial evidence in the record supports this conclusion. CMS's rebuttal evidence fails to indicate any specific areas which would receive inferior portable coverage under MCI's proposal and superior coverage under its own. CMS simply asserts that more cell overlap is better, pointing to its areas of greater cell overlap as justification for the preference. However, we cannot conclude that the FCC erred in deeming CMS's rebuttal evidence as less probative than CMS would have it.

Given its stated desire not to dictate more than minimal technical design standards and to encourage technical diversity while favoring flexibility, Cellular Rulemaking, 86 F.C.C.2d at 478, the FCC repeatedly and, in our view, properly declined to compare cellular design under the designated issues.57 Moreover, the obviously highly technical nature of this issue cautions the court to tread lightly in this respect; particularly in this specific area, it is appropriately the province of the FCC to weigh the public interest and decide in light of that polestar whether a particular system characteristic is to be garlanded with a preference. FCC v. WNCN Listeners Guild, 450 U.S. 582, 595-96, 602-03, 101 S. Ct. 1266, 1268, 1274-75, 1278, 67 L. Ed. 2d 521 (1981). On this record, it cannot reasonably be said that the Commission's rejection of CMS's inadequately supported "depth of coverage" claims was unsupported by substantial evidence or was otherwise arbitrary or capricious.

Issue (b) presents quite a different question for review, for the Commission chose not to award a preference to any of the three applicants for their respective expansion proposals. CMS, of course, has already succeeded in persuading the Commission to vacate the preference awarded by the ALJ to MCI under Issue (b). This issue, it will be remembered, calls for a comparison of the competing proposals to expand cellular system capacity in a coordinated manner within the CGSA in order to serve an increasing number of subscribers "as demand warrants." Cellular Rulemaking, 86 F.C.C.2d at 502-03. Preference is to be given to system designs characterized by efficient frequency use, which entails not only plans to split or add cells and to coordinate channels with neighboring systems, but also the degree to which the system is capable of frequency reuse. Id. at 503.

As under Issue (a), the ALJ criticized CMS's proposal for its failure to reflect the principles of a bona fide cellular system and expansion plan. In fact, virtually every ALJ reviewing the expansion plans of CMS affiliates in other proceedings discerned the same faults with the CMS "expansion cell" or "depth of coverage" approach. As we have seen, the Commission has repeatedly renounced the unfavorable treatment accorded CMS's design,58 concluding that "[a]ll of the applicants have presented bona fide cellular designs." Pittsburgh, 96 F.C.C.2d at 1025-26, 1029-30. In contrast to its administrative law judges, the Commission foreswore any notion of some sacrosanct cellular principle of uniform spacing, calling attention instead to its rejection of rigid standards and its emphasis on fostering flexibility and diversity of cellular designs. Id.; Cellular Rulemaking, 86 F.C.C.2d at 505-09. However, in rejecting CMS's demerits as undeserved, the Commission did not embrace CMS's claims of superiority; that of course would have had the effect of penalizing orthodoxy in design techniques, as reflected in MCI's and Celcom's traditional proposals. But in any event, CMS's design creativity was at bottom found lacking in substantiation. As the FCC well put it, "CMS has not demonstrated that building penetration and service to portables in Pittsburgh would be inadequate without its proposed depth of coverage. Accordingly, its proposal might be viewed simply as different from, not better than, the others." Pittsburgh, 96 F.C.C.2d at 1030.

In essence, the Commission's ultimate holding on Issue (b) was that all of the applicants offered, at minimum, "adequate" proposals. MCI's and Celcom's were said to provide "more flexibility." The flexibility finding is rooted in Judge Ehrig's finding that the MCI and Celcom systems were smaller and more conservative at the outset, "allowing greater flexibility for growth and adjustment as facts about the SMSA become known" and allowing adaptation with a minimum of costly and complex engineering modifications. Pittsburgh Initial Decision, 96 F.C.C.2d at 1064.

Before us, CMS inaccurately places the onus of the Commission's ultimate conclusion in this respect on the finding that CMS offered a "less flexible" proposal. CMS Brief at 62. While its plan was in fact described as less flexible, the plan was, critically, not downgraded as a result of that unflattering characteristic. Pittsburgh, 96 F.C.C.2d at 1031. Although CMS's proposal was "slightly more detailed," this was apparently deemed insufficient to warrant a preference, regardless of the design's relative flexibility or inflexibility; the primary reason advanced for refusing to award a preference was that "none of the expansion proposals is set forth in sufficient detail to warrant a preference." Id.

In reviewing a decision founded on insufficient detail in the competing proposals, we cannot invalidate that determination as arbitrary, capricious, or unsupported if the decision embodied a considered and sufficiently articulated judgment and was founded upon substantial evidence in the record. Even were we invited to sally forth into technical determinations as to sufficiency of detail or flexibility of the various system designs, we would be obliged to decline the invitation, for this highly technical task, once again, necessitates considerable deference to the specialized agency's expertise. MCI Cellular Telephone Co. v. FCC, 738 F.2d 1322, 1333 (D.C.Cir.1984). "The court will not substitute its judgment on highly technical and factual matters for that of the agency charged with the supervision of the industry." Community Nutrition Institute v. Novitch, 773 F.2d 1356, 1363 (D.C.Cir.1985). Our task, rather, is to "assure that the agency has given learned consideration to all the material facts and issues." Greater Boston Television Corp. v. FCC, 444 F.2d at 851. We are entirely satisfied that the Commission has faithfully discharged that bedrock obligation.

Undaunted, and labeling Issue (b) as the "smoking gun," CMS launches another assault on the flanks of the FCC's decision with the cry of "inconsistency" with the subsequent Portland decision.59 CMS claims that MCI and CMS each supplied identical exhibits in the Pittsburgh and Portland proceedings. In the Pittsburgh licensing, as we have seen, no preference was awarded because no applicant was deemed to have supplied sufficient detail. In Portland, however, CMS and all other applicants were granted a preference over MCI because their proposals reflected greater detail.

We have already observed the fundamental impropriety of this argument, since infidelity to Pittsburgh and Chicago, as the lead, precedent-setting cases, is a potentially pertinent inquiry in subsequent review proceedings. While adhering to that position here, we simply note in this respect that in Portland, the Commission expressly addressed its prior decisions, stating:

Our finding here is not inconsistent with our conclusions in Chicago and Pittsburgh, where we found that CMS's design was less flexible than its competitors' designs. Those cases involved much larger initial systems and required significant expansion in the first few years of operations. Chicago Final Decision, at para. 43 and Pittsburgh Final Decision, at para. 45.

Portland, 56 Rad.Reg.2d at 1379 n. 38. The Commission further distinguished Portland from the lead decisions in its subsequent Philadelphia decision, stating:

Our finding here is not inconsistent with our conclusion in Portland where we rejected the judge's finding that CMS' system is less flexible because its cell configuration was based on the fifth year of demand. There, all four applicants had proposed 6-8 cell systems with essentially static cell configurations for the first five years of operation. Here, on the other hand, we are comparing CMS' 23-cell system for the first five years with [a competitor's] dynamic expansion plan from an initial 18-cell system to 60 cells by the fifth year. Under these circumstances, a finding that [the competitor's] plan is more flexible is warranted. See Chicago Final Decision, ... at para. 43.

Philadelphia, FCC 84-649, at 16 n. 29.

It is therefore manifest that the agency has not blatantly refused to recognize the precedential effect of its own prior decisions. These decisions, based on what has been furnished to us, do not--at least on their face--exhibit the unacceptably erratic ebb and flow of administrative inconstancy which vitiates judicial deference to agency decisionmaking. See, e.g., Airmark Corp., 758 F.2d at 685; Local 777, Democratic Union Organizing Committee v. NLRB, 603 F.2d 862 (D.C.Cir.1978); Garrett v. FCC, 513 F.2d 1056 (D.C.Cir.1975).

The ultimate purpose of the Communications Act is to "secure the maximum benefits of [communication channels] to all the people of the United States." National Broadcasting Co. v. U.S., 319 U.S. 190, 217, 63 S. Ct. 997, 1009, 87 L. Ed. 1344 (1943). The statute imposes on the FCC the task of balancing the "competing objectives of quality, quantity, and timeliness of essential communications services" such as cellular radio. Cellular Rulemaking, 86 F.C.C.2d at 642 (statement of Commissioner Fogarty).

As the first of the comparative non-wireline cellular proceedings, this was a close decision for an agency faced with an overwhelming national demand and a mandate to move cellular telephone service into the marketplace. In this proceeding, three applicants were at hand qualified to construct and operate a cellular system to serve Pittsburgh. Faced with a new technology, admittedly imperfect theoretical models, untested criteria, and the goal of promoting technical diversity by exercising regulatory self-restraint in dictating system design choices, the Commission fully anticipated that its proceedings would be plagued with protracted challenges. It sought, reasonably and prudently, to guard against the "boundless litigiousness" of aspiring applicants.

In this proceeding, we conclude that CMS cannot persuasively claim either surprise as to the bases for comparison or the denial of a meaningful opportunity to address the comparative issues. We uphold entirely the Commission's exercise of discretion in this respect. We also hold that the FCC's comparative evaluation of the applicants reflects sound and reasoned decisionmaking and is supported by substantial evidence. That the decision was difficult renders it neither arbitrary nor cavalier. That the decision is close, resting on slight yet carefully drawn differences among qualified applicants, suggests in the end the correctness of our course in affirming the Commission.

Affirmed.