BAZELON, Circuit Judge.

At the request of the Greater Baltimore Committee, the Civil Aeronautics Board conducted an adequacy of service investigation to determine “whether carriers currently authorized to serve Washington * * * should be authorized to serve Baltimore as well * * * through [Baltimore’s] Friendship International *713 Airport * * * ” which is approximately thirty miles from Washington. 1 As a result of this proceeding, the Board ordered National Airlines, Inc., to operate one through (single-plane) round trip a day between Baltimore, Maryland, and Jacksonville, Tampa/St. Petersburg and Miami, Florida, with no more than two intermediate stops between Baltimore and each of these Florida cities. Although the order is effective until modified, the Board directed that the proceedings remain open to enable it “to re-examine the situation at Baltimore at the end of one year” and to make “any necessary modifications which may then be indicated.” The order also provided that “if after a reasonable period of good faith compliance * * * [petitioner] should conclude that actual experience demonstrates that further provision of the service ordered is not justified, it can make appropriate application to the Board for relief.”

National brought this petition for review attacking the Board’s determinations respecting (1) inadequacy of existing service, (2) economic feasibility of the service ordered, and (3) the requirement that the additional flights make no more than two intermediate stops between Baltimore and the Florida cities.

(1) During these proceedings, the Board adopted a rough standard to test the adequacy of service at Baltimore. That test was whether through flights were provided to and from each city with which Baltimore exchanged or will exchange ten passengers per day. National contends that this “ten-a-day” standard is not supported by substantial evidence. It was first suggested at the hearing by an expert witness for the Greater Baltimore Committee who testified that airlines should provide through service in markets exchanging ten passengers per day. In its brief, the Bureau of Air Operations urged this test of adequacy upon the trial examiner. He adopted it and sought to support it by a study, based upon data in the record, which showed that two thirds of a sample of city pairs served by through-planes exchanged ten or fewer passengers per day. The Board approved the examiner’s decision.

Whether the examiner’s study was adequate to support the “ten-a-day” standard appears immaterial in the present posture of the case. 2 In response to the carrier’s petition for reconsideration, the Board studied the service provided in 1270 markets exchanging various “graduated volumes of traffic covering the complete range between 5 and 15 passengers daily.” We think that study adequately supported the “ten-a-day” standard. But it appeared for the first time in the Board’s supplemental order denying reconsideration. 3

The question arises whether the Board may invoke this study upon reconsideration as evidence to support the standard applied earlier in the proceeding. 4 The Board’s supplementary study *714 was based upon standard origin and destination surveys and airline schedules which had been stipulated into the record. 5 Its finding that through-plane service is provided in more than half the markets exchanging ten passengers per day is no more than a statistical inference from that data. National did not object to that inference in its motion for stay following the denial of its petition for reconsideration. Nor does it argue here that it was prejudiced by a lack of opportunity to demonstrate error in the Board’s reasoning. Indeed, it asserts none. Moreover, National had ample opportunity both by exception to the trial examiner’s report and on petition for rehearing to proffer evidence and to argue against the “ten-a-day” standard, and to propose an alternative test. Under these circumstances, we think the Board’s supplementary study is sufficient to support the standard of adequacy it adopted. 6

(2) National’s next contention is that the record does not show that additional service to Miami, Jacksonville and Tampa/St. Petersburg would be economically feasible. The parties agree that “manifestly, the economic feasibility of additional or improved service sought under section 404(a) is a relevant and material consideration.” National presented evidence which tended to show that additional service in these markets would be “seriously unprofitable.” That evidence consisted of a record of load factors experienced on the Baltimore-Miami segment of an experimental non-stop Baltimore-Miami service originating in Newark which National offered during the 1956-57 peak traffic winter season. 7 The Board found this evidence “unpersuasive.” It therefore held that National had failed to meet the burden of proof on the issue of economic feasibility which, the Board says, rests upon the carriers. National asserts that it does not have the burden of proof and that, if it does, it has sustained it.

The issues embedded in this assertion are of great importance to both the airlines and the public. Careful consideration of economic feasibility is essential to avoid serious losses to the carriers. 8 On *715 the other hand, cities petitioning for adequate service and the Board would be unduly hampered by any requirement for overly detailed profit and loss projections to establish the economic feasibility of adequate service. This could not only intolerably protract adequacy of service proceedings, 9 but might also create an insuperable barrier to petitioning civic groups lacking both the relevant operating data and the assistance of experts. 10

We think the Board’s practice of keeping adequacy proceedings open to permit either party to attack the remedial order after a trial period and of reviewing its determinations after one year is sound. 11 Where an order of that kind is issued after the Board has established a need for additional service, we think it need not affirmatively find economic feasibility but may properly place the burden of producing evidence of infeasibility upon the carriers. 12 We turn, therefore, to the question whether National sustained that burden.

On its face, National’s evidence would appear substantial on the question whether additional service to Florida would be economically feasible, since it shows that when such service was instituted during the peak winter season, it proved unprofitable. National argues that the Board’s own figures show that Miami is Baltimore’s largest market and that a fortiori additional service to Jacksonville and Tampa/St. Petersburg would be economically infeasible.

But the Board rejected this data because the drawing power of the additional flight was inevitably reduced by the poor over-all pattern of service at Baltimore and the consequent diversion of traffic to Washington, and because a substantial net increase in Miami traffic accompanied National’s experimental flight. 13 We think these facts lessen the predictive value of National’s experience to such an extent that they preclude us from disturbing the Board’s conclusion that the experience was not probative evidence of economic infeasibility. We do not decide whether National’s burden required it to produce detailed cost and revenue projections showing that additional service could only be provided at a loss. 14 Nor need we consider, as the Board urges in its brief, whether National’s burden also *716 required it to show that such service would have a seriously deleterious effect on its system-wide earnings. 15

(3) National’s final assignment of error is that there is no basis in the record for the Board’s requirement that the additional flights ordered make no more than two intermediate stops. Its argument is that the Board adduced no evidence that three or more intermediate stops rendered through-plane service inadequate. We think it was not required to do so.

Service at Baltimore was inadequate, inter alia, because many passengers had to change planes en route, thus increasing their travel time. This inadequacy could not be remedied unless, in addition to ordering single-plane service, the Board limited the number of intermediate stops, since those stops could cause as much delay as plane changes and thereby deprive Baltimore of the benefit of the Board’s remedial order. In fashioning its remedy, the Board has wide discretion and is entitled to rely upon its cumulative experience as well as the record before it. See National Labor Relations Board v. Seven-Up Bottling Co., 344 U.S. 344, 349, 73 S.Ct. 287, 97 L.Ed. 377 (1953); Jacob Siegel Co. v. Federal Trade Comm., 327 U.S. 608, 611-613, 66 S.Ct. 758, 90 L.Ed. 888 (1956) (dictum). The Board noted that “with the telescoping of flight time associated with the growing use of modern equipment, the time required for stops represents an increasingly higher proportion of the total flight time than heretofore.” It therefore concluded that the additional flights it ordered would be adequate only if intermediate stops were limited to two. We think this judgment, based upon the Board’s experience, falls within the zone of reasonableness and may not be disturbed.

So ordered.

Notes

. Adequacy of service investigations are conducted pursuant to Federal Aviation Act, § 404(a) 72 Stat. 760 (1958), 49 U.S.C.A. § 1374(a), which requires airlines “to provide safe and adequate service * *

See generally Comment, Adequacy of Domestic Airline Service, 68 Yale L.J. 1199 (1959).

. Underlying that study was the assumption that service requiring a change of planes between Baltimore and a city with which it exchanges a given number of passengers would be inadequate if other city pairs exchanging a like number of passengers ordinarily received through-plane service. To show the traffic level at which single-plane service is ordinarily available, the examiner studied the traffic flow between city pairs receiving such service. We note that the sample contained no cities connected by flights requiring a change of planes, and thus predetermined the conclusion that cities exchanging ten passengers per day ordinarily receive through-plane service.

. A tabular summary was printed in an appendix to the order.

. This case is unlike Michigan Consolidated Gas Co. v. Federal Power Comm., 108 U.S.App.D.C. 409, 283 F.2d 204, cert. denied, 364 U.S. 913, 81 S.Ct. 276, 5 L.Ed.2d 227 (1960), where we implicitly disapproved the Commission’s reliance *714 upon a standard invoked for the first time in its decision denying reconsideration. Here the ten-a-day standard was announced in the trial examiner’s opinion and the parties were thus able to attack its validity before the Board.

. The Board has announced its decision to take official notice of these sources in the future. 26 Fed.Reg. 8450-51 (1961).

. Cf. American Airlines v. Civil Aeronautics Board, 89 U.S.App.D.C. 365, 369-370, 192 F.2d 417, 421-422 (1951); Delta Airlines v. Civil Aeronautics Board, 247 F.2d 327, 331-332 (5th Cir. 1957). See also National Labor Relations Board v. Seven-Up Bottling Co., 344 U.S. 344, 348-349, 73 S.Ct. 287, 97 L.Ed. 377 (1953); 2 Davis, Administrative Law § 15.03 (1958). Compare Michigan Consolidated Gas Co. v. Federal Power Comm., 108 U.S.App.D.C. 409, 283 F.2d 204, cert. denied, 364 U.S. 913, 81 S.Ct. 276, 5 L.Ed.2d 227 (1960).

The survey examined separately the service provided city pairs 500, 501 to 750, and 751 to 1,000 miles apart. This breakdown was intended to meet the carrier’s contention that distance is an important consideration when determining whether through flights are required. The figures show that through service is available in a majority of the markets comparable in traffic flow and distance to Baltimore-Miami and Baltimore-Jacksonville. Only 43.5 percent of the markets comparable to Baltimore-Tampa/St. Petersburg receive such service, however. But the Board found that through-plane service would be provided in such a market if one terminal were a resort, as is Tampa, and the other a major city, as is Baltimore. We think this adjustment of the figures in the study justified by the Board’s expertise. Of. Southeastern Aviation, Inc. v. Civil Aeronautics Board, 108 U.S.App.D.C. 394, 398, 283 F.2d 189, 193 (1960).

. The load factor is the percentage of available seats occupied by revenue passengers. The factors here ranged from 35 percent to 46 percent. National says “the load factors did not even approach minimum standards necessary to support off-season operations much less peak season operations.” The Board does not challenge this characterization.

. See Toledo Adequacy of Service Investigation, C.A.B. Order No. E-16236 (Jan. 10, 1961); Note, 36 Notre Dame Lawyer 539, 543 (1961).

. The adequacy investigation under review was initiated on January 23, 1957.

. The Board argues it is also at a disadvantage in preparing cost and revenue projections because the carriers alone has access to operating figures. But see 14 C.F.R. §§ 231, 241 (1961).

. For examples of this approach, see Capital Airlines, Inc. v. Civil Aeronautics Board, 108 U.S.App.D.C. 215, 281 F.2d 48 (1960); Service to Ely, Nevada, Av. L.Rep. (1954-57 Cab.Cas.) H 21812 (1955).

. Petitioner argues that Administrative Procedure Act, § 7(c), 60 Stat. 241 (1946), 5 U.S.C.A. § 1006(c), which places the burden of proof upon the proponent of an order, prohibits the Board from placing the burden of proving economic infeasibility upon the carriers. We think that once the Board has established that existing service is inadequate to meet the needs of a community, it has discharged its burden of proof. Economic infeasibility is more like an affirmative defense than an element of the Board’s case. Accordingly we think that the Board properly placed the burden of producing evidence of economic infeasibility upon petitioner. See Specht v. Civil Aeronautics Board, 254 F.2d 905, 914-915, 78 A.L.R.2d 1135 (8th Cir. 1958), and cases cited therein. Cf. Law v. National Labor Relations Board, 192 F,2d 236 (10th Cir. 1951); Montgomery Ward & Co. v. National Labor Relations Board, 107 F.2d 555, 560 (7th Cir. 1939).

. National’s own witness agreed that the poor over-all service adversely affected the results of its efforts to sell the flight. He also pointed out that National intended voluntarily to continue the service, though the plane could be more profitably used elsewhere.

. See Toledo Adequacy of Service Investigation, C.A.B. Order No. E-14629, p. 8, n. 13 (Nov. 10, 1959). Cf. North Carolina v. United States, 325 U.S. 507, 519-520, 65 S.Ct. 1260, 89 L.Ed. 1760 (1945). The Board indicated that National’s evidence was insufficient in part because cost figures were not provided.

. Cf. Fort Smith Light & Traction Co. v. Bourland, 267 U.S. 330, 332-333, 45 S.Ct. 249, 69 L.Ed. 631 (1925); Saint Louis & S. F. R. v. Gill, 156 U.S. 649, 665, 15 S.Ct. 484, 39 L.Ed. 567 (1895).