(dissenting):
I respectfully dissent.
While there may have been an overreaction on the part of some to the threats from this “little” reactor, there are two points which seem to me to require reversal — -the absence of AEC safety standards relating to teaching or research reactors and the absence of an Environmental Impact Statement (EIS) under NEPA.
Regarding safety standards, the AEC Safety and Licensing Board itself denied a license because it would not answer questions about the reactor’s possible effect on public health and safety absent applicable substantive criteria and convincing objective standards concerning *240evaluation of possible nuclear accidents. 2 CCH Atom.En.L.Rep. ¶ 11,595 (May 18, 1972). One must certainly agree with the Safety Licensing Board that the “fission products release fraction under postulated accident conditions for such a standard product as a Triga reactor ought to be a matter of proven determination and unquestionable validity.” Moreover, apparently the regulatory staff of AEC has shifted the basis of its safety evaulation of accident conditions from one Triga case to another. The Appeal Board did not really disagree with the need for standard regulations — in fact it strongly recommended adoption of them — but went ahead and granted a license here anyway. While the petitioners’ postulated accidents seem quite remote to me,1 it may be that there is some other postulated accident that should be taken into account that was not in evidence in this proceeding, and a rule-making procedure would seem to be the preferable way to develop this. Through a rule-making procedure the Commission may “educate itself,” cf. Lewis-Mota v. Secretary of Labor, 469 F.2d 478, 482 (2d Cir. 1972); Texaco, Inc. v. FPC, 412 F.2d 740, 744 (3rd Cir. 1969), and formulate all of the necessary standards and requirements based upon appropriate scientific data, and that would make it easier for everyone to know the Commission’s regulations on an issue of substantial importance; apparently Triga and other teaching reactors are widely used and may be even more widely used in the future. Courts, I believe, have the power to remand for rule-making proceedings in cases of this type, where the agency obviously should have formulated a generalized policy but has failed to do so. Cf. Bell Aerospace Co. Division of Textron, Inc. v. NLRB, 475 F.2d 485, 494-496 (2d Cir. 1973) (Friendly, C. J.) (change in Labor Board’s “managerial employees” exception to coverage of National Labor Relations Act made through adjudication remanded to Board for rule-making proceeding).
A rule-making proceeding here is particularly appropriate for another reason —it might serve as a measure of protection against the agency’s own interests influencing its disposition of this particular case. I have already mentioned the Commission’s variable “standards” for licensing Triga reactors.
I continue2 to express concern that the AEC is charged with the dual duty *241of passing on licenses on the one hand but promoting the use of atomic power on the other. The AEC also has an interest in seeing this reactor licensed to promote its Nuclear Education Training Program; to this end it has entered into an agreement with Columbia to provide post-license funds for the operation of the reactor and to waive charges for Commission-owned special nuclear material involved in its operation. My concern is so much the greater where, as here, the independent Safety and Licensing Board has denied the license, but was overruled by the three-member Appeal Board, consisting of two AEC staff members. Moreover,, the safety tests relied upon by the Appeal Board here were conducted by the most interested party, the manufacturer of the reactor, following an exchange of correspondence between “Ralph” (Mr. Peters of Gulf Oil) and “Pete” (Dr. Peter A. Morris, Director of the AEC Division of Reactor Licensing) in which Gulf advised the Commission that “We hope to have you or members of your staff participate fully in these experiments so that they will be deemed to have been done ‘under the auspices of the Regulatory Staff.’ ” Following this Dr. Morris advised Gulf that he would attend the tests “as an observer.” In holding that the FPC could not adopt as its own an EIS drawn up by a state agency-applicant for a power license this court said in Greene County Planning Board v. FPC, 455 F.2d 412, 420 (2d Cir.), cert. denied, 409 U.S. 849, 93 S.Ct. 56, 34 L.Ed.2d 90 (1972): “The danger of this procedure [the one followed by the FPC], and [¿is] obvious shortcoming, is the potential, if not likelihood, that the applicant’s statement will be based upon self-serving assumptions.” (Emphasis supplied.) While Greene County dealt with the requirements of NEPA I cannot believe that the AEC should be held to any lesser standard in its safety licensing procedures. The Commission should not be permitted to abdicate its responsibility for ensuring public safety to a party with a large financial stake in the outcome of the tests it is conducting. One good way to insure it does not do so is to require it to draw up safety standards generally applicable to all teaching reactors, not just to the one in this particular case.
With regard to the necessity for the preparation and filing of an EIS under NEPA, 42 U.S.C. §§ 4321-4347 (1970), the threshold question is whether, or to what extent, the agency’s own determination that a given agency action is not a “major Federal action significantly affecting the quality of the human environment,” 42 U.S.C. § 4332(2) (C), is binding upon us. This court has previously held in Hanly v. Kleindienst, 471 F.2d 823, 829 (2d Cir. 1972), that “the appropriate criterion ... is the ‘arbitrary, capricious’ standard established by the Administrative Procedure Act, since the meaning of the term ‘significantly’ as used in § 102(2) (C) of “NEPA can be isolated as a question of law.” The two relevant factors for consideration, Hanly v. Kleindienst went on to say, 471 F.2d at 830-831, were: “(1) the extent to which the action will cause adverse environmental effects in excess of those created by existing uses in the area affected by it, and (2) the absolute quantitative adverse environmental effects of the action itself, including the cumulative harm that results from its contribution to existing adverse conditions or uses in the affected area.” I do not think that it would be inconsistent with, or stretching, this language to say that the true first test is the extent to which the action has the potential of causing .adverse environmental effects. Cf. Scientists Institute v. AEC, 481 F.2d 1079 (D.C.Cir., 1973) (NEPA requires EIS for AEC’s breeder reactor program even though program is presently only in research and development stage in part because of potential future environmental effects of program). For Hanly articulated a threshold test, and if its questions are answered in the negative then an EIS, through which the full scope of the potential environmental effects of a “major Federal action” are *242actually brought out, would, not be required. If the potential is substantial then the impact statement must be forthcoming; NEPA speaks of making an initial determination “as to whether the proposal would have a significant effect,” not whether it “does have” such an effect.
By upholding the AEC’s determination that there would not be any such significant potential effect here, the majority apparently adopts the “rational basis” standard of review rejected by the Hanly court. See 471 F.2d at 829. The effect of the majority’s decision is to give the agencies a loophole big enough to make NEPA meaningless, or in Judge Skelly Wright’s words a “paper tiger.” See Calvert Cliffs’ Coordinating Committee, Inc. v. AEC, 146 U.S.App.D.C. 33, 449 F.2d 1109, 1114 (1971). The potential effect of a nuclear accident on the environment of New York City would be pretty substantial, if not catastrophic, and even though the Safety and Licensing Board and the Appeal Board did not accept the petitioners’ postulated accidents, they each considered safety standards that involved hypothetical “accidents.” The fact that the former found there were no generalized standards and that the latter found that under the regulatory staff’s postulated accident situation “the health and safety of the public will not be endangered” do not take away from the proposition that in the event of an accident the damage would be great. The fact that the Appeal Board in its decision did “not consider it desirable to use the standards of 10 CFR Part 20 for evaluating the effects of a postulated accident in a research reactor” even while it “strongly recommend [ed] that specific standards for the evaluation of an accident situation in a research reactor should be formulated” in and of itself indicates to me that there is here involved a sufficient “potential” of environmental effect to •require an impact statement.
Moreover, ten pages of the Appeal Board decision is devoted to answering the testimony of appellant’s expert, a Dr. Sternglass, relative to infant mortality rates in the vicinity of Triga reactors at Penn State College, the University of Illinois and in San Diego where the Gulf-General Atomic Corporation has its plant. It may be that Dr. Sternglass’s determinations do not have any “valid scientific foundation,” as the Appeal Board held. (I suppose that many thought that Rachel Carson’s thoughts and fears expressed in Silent Spring had none either.) But does that mean that there is not a potential environmental effect that at least bears examination through the preparation of an EIS? I should suppose that even though this is a “little” research reactor its potential environmental impact is at least as significant as that of a federal jail, Hanly v. Kleindienst, supra, or the abandonment of a 1.8 mile stretch of railroad, City of New York v. United States, 337 F.Supp. 150, 158-160 (E.D.N.Y.1972) (three-judge court). See generally Developments in Environmental Law, 3 E.L.R. 50001, 50005 (1973). The finding here by the Commission that no impact statement is required in my view was arbitrary and capricious.
Were an impact statement to be required the AEC could not argue that the Appeal Board’s decision amounted to one. This is so because (A) the Board relied quite heavily on data furnished by the manufacturer contrary to the requirements of Greene County Planning Board v. FPC, supra, and because (B) there has not been consultation with other concerned federal agencies such as EPA or HEW, as required by 42 U.S.C. § 4332(C).
Because I think that the AEC did not meet the requirements of its own Act in reversing the Safety and Licensing Board and because I think enough potential environmental effect was demonstrated to require the preparation of an EIS, I dissent.
. Petitioners’ expert postulated three hypothetical accidents: (1) sabotage, causing the loss of shielding water plus rupturing of one of several fuel elements; (2) any circumstance causing a fuel element to reach 1000° centigrade, followed by a rupture of the fuel elements in air; (3) an aircraft accident.
Even through sabotage at Columbia University is not exactly an unheard-of-event, there was substantial evidence that the juxtaposition of so many relatively sophisticated operations would apparently be required to make the sabotage effective that the “Mission Impossible” team would be required to effectuate it. There is some evidence to indicate that 1000° centigrade cannot be attained in a Triga. While I cannot agree with the AEC view that it is a one in 10 billion chance that aii airplane would accidentally strike the Columbia Triga, because I don’t think such an event is susceptible to meaningful mathematical computation, there is some evidence, however, that it would take a vertically dropping plane penetrating four feet of earth, two feet of reinforced concrete and a steel beam crane at a specific time when fission product accumulation was high to create a possibility of serious radioactive release.
. See Developments in Environmental Law, 3 E.L.R. 50001, 50008 (1973):
. for the most part the agencies which must do the “full good faith” balancing of economic and social costs against environmental costs are generally structured to be advocates for economic expansion. As long as agencies are left to do the balancing, and as long as they have a dual mandate of environmental protection and economic development in their particular field— for example, power growth for the FPC, nuclear development for the AEC, or flood containment for the Army Corps of Engineers — is not the environment bound to come out on the short end?
(Emphasis original.)