COFFIN, Chief Judge.

This is our second look at a controversy surrounding a housing project undertaken by a private developer with assistance from the Department of Housing and Urban Development (HUD). The facts of the case are to be found in Silva v. Romney, 473 F.2d 287 (1st Cir. 1973). See also 342 F. Supp. 783 (D.Mass. 1972). Appellants, owners of land abutting the proposed Forest Glen Project in Stoughton, Massachusetts, argue that the district court wrongly dissolved the injunction against HUD and the developer because it erroneously found that the environmental impact statement (EIS) submitted by HUD complied fully with the National Environmental Policy Act of 1969 (NEPA), 42 U.S.C. § 4332(2)(C). While we agree with the district court that determining good faith compliance with § 4332(2)(C) should not be a vehicle for a court to "interject itself within the area of discretion of the executive", Natural Resources Defense Council, Inc. v. Morton, 148 U.S.App.D.C. 5, 458 F.2d 827, 838 (1972), we have concluded that the EIS in this case was insufficient on its face measured by the restrained but not meaningless applicable judicial review standards.

When HUD withdrew its initial appeal in this case in July, 1972, and agreed to the preparation of an EIS, there was every expectation that, although such an exercise involving a project of this scope was novel, HUD would be able to conduct a thorough analysis within a reasonable time. A draft EIS was sent out for comment on August 18, 1972 and replies collected from numerous federal, state, and local agencies and others, including appellants who presented objections to various aspects of the plan and the plan as a whole. Appellants contend that the final EIS, issued six and one half months later, on March 5, 1973, is deficient both in failing to discuss adequately the objections put forward and in not sufficiently explaining HUD's resolution of various environmental problems, as required by § 4332(2)(C). Thus, appellants argue, HUD is thereby legally precluded from giving final approval to the development.

In identifying the source of review power courts have, as did the district court here, drawn on NEPA itself, 42 U.S.C. § 4331 et seq., and the general "federal question" power, 28 U.S.C. § 1331. We think an alternative source, equally applicable, is the Administrative Procedure Act, 5 U.S.C. §§ 702 and 706. In any event, the judicial inquiries are whether the agency's findings and conclusions in the EIS are arbitrary, capricious, an abuse of discretion, or otherwise not in accordance with law, and whether the agency followed the procedures required by law. In this case we face an underlying procedural matter which has much to do with a court's ability to perform its function of reviewing challenges to environmental impact statements.

In reviewing the final EIS and HUD's decision to proceed with the plan as described therein, the district court considered only the final statement, the draft statement and comments filed thereto, certain affidavits and testimony taken in court. It refused appellants' requests that the administrative record be produced. This record contains the more detailed studies and background of deliberation which form the basis of the final EIS. We think that the law requires production of the entire administrative record. See Citizens to Preserve Overton Park v. Volpe, 401 U.S. 402, 419-420, 91 S.Ct. 814, 28 L.Ed.2d 136 (1971); Camp v. Pitts, 411 U.S. 138, 93 S.Ct. 1241, 36 L.Ed.2d 106 (1973); see 5 U.S.C. § 706; cf. 28 U.S.C. § 2112.

Aside from the fact that this prerequisite to judicial review seems legally required, it makes good practical sense. Thus, even assuming that the final EIS might seem fully adequate standing alone, uncertainty must attend a court's approval of a statement based on an unknown "record of expert views and opinions, the technological data and other relevant material . . . on which the [agency] acted" but now refuses to supply. Appalachian Power Co. v. E. P. A., 477 F.2d 495, 507 (4th Cir. 1973). Moreover, to the extent that there is more than a perfunctory exchange of views incorporated in the record, i. e., a reflection of comments not merely on a preliminary draft but on new approaches or evidence subsequently developed, the need for taking additional evidence in court diminishes. The objective should be to develop an administrative record which is self-sufficient for adequate judicial review. Overton Park, supra, 401 U.S. at 420-421, 91 S.Ct. at 825-826. Finally, full disclosure is both a spur to reasoned decision making and a protection against criticism unfairly sought to be made after the agency's processes have concluded. Cf. General Transportation Co. v. United States, 65 F. Supp. 981 (D.Mass.), aff'd, 329 U.S. 668, 67 S.Ct. 75, 91 L.Ed. 590 (1946).

Once the complete record is before the court, it must determine the adequacy of the EIS. The relevant legal standards are those of NEPA which specifies procedures to be used in preparing an EIS for projects subject to that act; requires that "presently unquantifiable environmental amenities and values . . . be given appropriate consideration in decision-making along with economic and technical considerations", 42 U.S.C. § 4332(2)(B); and mandates the agency involved to "study, develop and describe appropriate alternatives to recommended courses of action in any proposal which involves unresolved conflicts concerning alternative uses of available resources". 42 U.S.C. § 4332(2) (D).

The "detailed statement" required by § 4332(2)(C) serves at least three purposes. First, it permits the court to ascertain whether the agency has made a good faith effort to take into account the values NEPA seeks to safeguard. To that end it must "explicate fully its course of inquiry, its analysis and its reasoning." Ely v. Velde, 451 F.2d 1130, 1139 (4th Cir. 1971); Appalachian Power Co., supra, 477 F.2d at 507. See also Natural Resources Defense Council v. E. P. A., 478 F.2d 873 (1st Cir. 1973); Environmental Defense Fund v. Ruckelshaus, 142 U.S.App.D.C. 74, 439 F.2d 584 (1971). Second, it serves as an environmental full disclosure law, providing information which Congress thought the public should have concerning the particular environmental costs involved in a project. To that end, it "must be written in language that is understandable to nontechnical minds and yet contain enough scientific reasoning to alert specialists to particular problems within the field of their expertise." Environmental Defense Fund v. Corps of Engineers, 348 F. Supp. 916, 933 (W.D.Miss. 1972). It cannot be composed of statements "too vague, too general and too conclusory." Environmental Defense Fund v. Froehlke, 473 F.2d 346, 348 (8th Cir. 1972). Finally, and perhaps most substantively, the requirement of a detailed statement helps insure the integrity of the process of decision by precluding stubborn problems or serious criticism from being swept under the rug. A conclusory statement "unsupported by empirical or experimental data, scientific authorities, or explanatory information of any kind" not only fails to crystallize issues, Natural Resources Defense Council v. Grant, 355 F. Supp. 280, 287 (E.D.N.C. 1973), but "affords no basis for a comparison of the problems involved with the proposed project and the difficulties involved in the alternatives." Monroe County Conservation Council v. Volpe, 472 F.2d 693, 697 (2d Cir. 1972). Moreover, where comments from responsible experts or sister agencies disclose new or conflicting data or opinions that cause concern that the agency may not have fully evaluated the project and its alternatives, these comments may not simply be ignored. There must be good faith, reasoned analysis in response.

The final statement presented here falls far short of what is required and fails to serve the several purposes we have noted. Appellants' first and major challenge is to the statement's treatment of the drainage problem. The project's site contains a low wetland portion, in and near an area where the water table is high. Adjacent lower lying areas have historically experienced chronic flooding. This is plainly a major problem. We think it is not too much to ask that the problem be fully depicted, that HUD describe the approach that was taken, and the reasons why the particular mode of control was chosen in preference to others.

The draft statement circulated to agencies contained a four line description of a third drainage plan, agreed upon by HUD and the developer, describing the diameter of two pipes. This was inserted in the draft statement, together with a prior statement that a second plan, involving the construction of a series of holding tanks, was deemed "workable", and that the maintenance costs for the second plan could be provided for in the mortgage. The draft plan not only contained this inconsistency on its face but another which appears from external sources. While the draft plan, circulated as of August 18, 1972, stated that HUD had restudied the drainage problem with the developer and agreed to the third plan, the government affidavit, filed after hearing in the spring of 1973, stated that HUD had made its request for an in-house review to provide it with an engineering evaluation of the third drainage plan (dated June 27, 1972) in January of 1973 to be used for the preparation of the Final Environmental Impact Statement. If HUD had in fact "restudied" the plan before August 18, 1972, the sole reference to a January, 1973, review is not easily understood.

In any event, the draft plan's presentation of HUD's approach to drainage, internally consistent or not, drew heavy fire, as being wholly inadequate, from the federal Departments of Agriculture and of Commerce, and the Environmental Protection Agency, agencies with expertise as to drainage problems equal to or greater than that of HUD. HUD's response to these comments in the final statement was a bland reference to the section on drainage. Apart from repeating the pipe diameter descriptions of the draft statement, the final statement contained nine lines of evaluation.

HUD's first point made was that discharge "to the low area beyond Washington Street would reduce or eliminate the drainage problem . . . at Britton Avenue by providing adequate drainage and loweing [sic] the water table in the area of the 42 [sic] inch drain." No explanation is given for the lowering of the water table or for the scope of "the area" where lowering would take place. While Britton Avenue at the edge of the project seems to be taken care of, nothing is said about the effects of discharge on "the low area beyond Washington Street" where flooding problems have been endemic. A second point is possibly contained in the statement: "With the implementation of Plan Number 3, a determination was made to the addition of the proposed project." This is one of several statements which, through omission or other error, are simply incomprehensible to us. The references to "the plan" help us no more, for all we find is a pre-plan 1971 letter on run-off estimates from the engineering company which later proposed the plan and a sketchy blueprint which merely indicates several locations for pipes. The statement concludes with the twin pronouncements that the plan "for an off-site drainage system appears to be the best method for draining the site and would certainly lower the water table of those acres surrounding the site"; and that the "proposed project . . . does not affect any [areas facing present or potential flooding problems]."

The statement does not recognize the scope of the flooding problem as described by others, or, if they are inaccurate, does not say why. It does not describe alternative approaches. It does not say why the plan agreed upon was chosen in preference to others. It does not pretend to estimate with any detail what the agreed-upon plan will accomplish, and for what area. And it is myopic in the sense of focusing on the project site itself, with no supported judgment as to what will happen in other areas nearby. Even the presence of the full administrative record would have done little to remedy the clear violation of the mandate under NEPA for a "detailed statement" of the drainage problem.

A second objection of appellants is that the statement inadequately discussed alternatives as to numbers and placement of housing units. Four alternatives were considered: no action, in which event housing of greater density and under less restriction might be constructed; recreation or conservation, which would require a public entity, not foreseen, purchasing the property; utilizing another site, there being assertedly no better available sites; and changing the configuration of buildings on the existing site. This last was said to be the real area of choice, allowing (1) high-rise buildings with more vacant land, (2) the same number of low-rise buildings crowding a smaller portion of Forest Glen, or (3) a smaller number of buildings, leaving more of the land in a natural state. As to this last possibility, perhaps the likeliest alternative, the statement said that this "would be to substantially reduce the number of units, an economically unsound alternative."

This conclusion may very well be supportable. But the agency must go beyond mere assertions and indicate its basis for them. The size of the project began at 186 units. It was reduced once to 166 and then to 138. We echo the comment that the Department of the Interior made to the statement, that the discussion of alternatives should be "expanded to identify costs associated with appropriate alternative actions so as not to prematurely foreclose options which might have less detrimental effects." It is not too much to require an explanation why a lesser aggregation of housing units on the property would be either economically unsound or, as a policy matter, would represent less of a benefit in terms of low and middle cost housing needs than would be justified by the costs associated with the disadvantages to the environment. Here as with the drainage problem, what the courts look for is an informed and adequately explained judgment. When courts can say that such a judgment has been made, their mission is at an end.

The EIS and administrative process used by HUD to treat comments falls far short of what is required. Review of the statement and the ultimate decision to proceed on the project has been reduced to a game of "blind man's bluff", Appalachian Power Co., supra, 477 F.2d at 507. The improvement of the regulatory system administered by each agency has been attained in this case, only minimally, if at all; and the full public disclosure required by NEPA is inadequate. An EIS as defective as the one before us cannot serve to fulfill the mandate that environmental factors "be given appropriate consideration in decision-making along with economic and technical considerations", 42 U.S.C. § 4332(2)(B), and a decision to proceed with a project based on that statement is not in accordance with the law.

Upon remand, the question remains: what should be done at this juncture? HUD has had nearly a year to prepare a satisfactory EIS on a relatively small project. Perhaps, because it was small and involved a partnership arrangement with a private developer, the task was novel. Nevertheless, unless we abdicate all responsibility, we could not conscientiously say that the statement, either in spirit or letter, conformed to the requirements of NEPA. In the meantime, the developer, since earlier argument before us, has stayed his hand. Now, with costs of construction rising, and costs of delay accumulating, he faces further delay. And the appellants, sometimes excessive in their demands, have nevertheless served the public purpose in pursuing this cause.

We remand to the district court with instructions that it return the final EIS to HUD, requesting it to supply more detailed reasons supporting its conclusions as to the third drainage plan — if it is of this opinion still — and, with the same caveat, as to the economic unsoundness of a smaller number of housing units, and sufficient data to serve as a basis for these reasons. We would not contemplate that the district court, in the absence of special circumstances which we cannot now foresee, would have the need to take more evidence. But we do not attempt to prejudge. See, e. g., Allison v. Froehlke, 470 F.2d 1123 (5th Cir. 1972). Upon the presentation of an amended Final Statement, together with the administrative record to date, which we think is required, we would hope that the court could come to a fairly rapid conclusion as to the validity of the statement.

Reversed and remanded.