The opinion of the court was delivered by: OBERDORFER
LOUIS F. OBERDORFER, UNITED STATES DISTRICT JUDGE
Plaintiff National Security Archive is a nonprofit research institute and library whose main function is to discover, analyze, and index documents relating to national security matters. It also organizes these documents into sets which it makes available to libraries, scholars, journalists, Members of Congress, and the general public.
In July 1987, the Archive filed a request under the Freedom of Information Act ("FOIA"), see 5 U.S.C. § 552(b) (1988), with defendant Federal Bureau of Investigation for documents concerning the FBI's "Library Awareness Program." The FBI instituted that program in response to perceived activity by foreign agents in technical libraries across the country.
The FBI suspects that Soviet agents have been attempting to gain access to sophisticated scientific and technical information through the National Technical Information Service. Although that Service is open to the general public, an Executive Order bars Soviet nationals from access to it. The FBI also suspects Soviet agents of using libraries to recruit both librarians and students. According to the FBI, the Library Awareness Program, which was implemented by its New York office, responds to these threats to national security. For the most part, its purpose is simply to alert librarians in twenty-one technical libraries in the New York metropolitan area of suspected Soviet activities and to inform them how to report such activities. Other libraries and librarians have been contacted in response to specific reports of Soviet activity. The FBI, however, does not consider these investigations to be part of the Library Awareness Program.
On July 10, 1987, the Archive requested information concerning the Library Awareness Program from FBI Headquarters. Two months later, it submitted an additional request to the New York office. The FBI initially denied these requests on the grounds that there were no responsive records. In October 1987, the FBI corrected itself and informed the Archive that its earlier determination had been in error and that it would be processing the Archive's requests. By June 1988, however, the FBI had not released any documents. Accordingly, on June 2, 1988, the Archive filed this suit. Later that month, the FBI produced 22 documents. After the Archive challenged the adequacy of the FBI's search and a hearing on the subject was held, the parties entered into a stipulation. Following that stipulation, which was filed May 1, 1989, the FBI reviewed more than 2,000 pages of records. Invoking Exemptions 1, 2, 5, and 7 of the FOIA, the FBI withheld many of these documents entirely and redacted large portions of others. For approximately two hundred of these documents, the FBI filed Vaughn indexes describing the information withheld.
Based upon those declarations, the parties filed cross-motions for summary judgment. A hearing on those motions was held on November 14, 1990. On November 15, 1990, an Order informed the parties that the matter remained under advisement because the record did not provide a sufficient basis for a judgment. The Order also required the FBI to file in camera affidavits concerning most of the information withheld under Exemption 1 and a copy of the briefing book that Geer used to prepare for his testimony before Congress. Finally, the Order required the FBI to file publicly a copy of Geer's testimony and an affidavit concerning whether or not the names of policymaking officials had been withheld under Exemption 7(C). This Memorandum explains and modifies that Order.
Under Exemption 1, an agency need not disclose materials that are
(A) specifically authorized under criteria established by an Executive order to be kept secret in the interest of national defense or foreign policy and (B) are in fact properly classified pursuant to such Executive order. . . .
5 U.S.C. § 552(b)(1) (1988). To properly invoke this exemption, the FBI must show that it has complied with the classification procedures established by the relevant Executive Order and that it has withheld only material properly classified under that order's substantive criteria for classification. See, e.g., King v. United States Dep't of Justice, 265 U.S. App. D.C. 62, 830 F.2d 210, 214 (D.C. Cir. 1987). The Archive does not challenge the validity of the procedures followed by the FBI. Moreover, it does not directly challenge the FBI's application of the relevant criteria for classification. Instead, the Archive criticizes the sufficiency of the description in the FBI's Vaughn index of those justifications and of the passages withheld.
The Archive's focus upon the sufficiency of the FBI's Vaughn indexes reflects the nature of the review of Exemption 1 claims. Because judges "lack the expertise necessary to second guess  agency opinions in the typical national security FOIA case," Halperin v. CIA, 203 U.S. App. D.C. 110, 629 F.2d 144, 148 (D.C. Cir. 1980), courts accord substantial weight to the determination of Executive Branch officials that information is properly classified. See S. Rep. No. 93-1200, 93d, Cong., 2d Sess. 12 (1974); Abbotts v. Nuclear Regulatory Comm'n, 247 U.S. App. D.C. 114, 766 F.2d 604, 606 (D.C. Cir. 1985). As a consequence, even though the Government has the burden of proving de novo that any information it has withheld fits under one of the exemptions to the FOIA, see 5 U.S.C. § 552(a)(4)(B) (1988),
in the national security context that burden is relatively light. An agency's declarations need merely:
describe the withheld information with reasonable specificity, demonstrating a logical connection between the information and the claimed exemption, and . . . evidence neither bad faith on the part of the agency nor a conflict with the rest of the record. . . .
Salisbury v. United States, 223 U.S. App. D.C. 243, 690 F.2d 966, 970 (D.C. Cir. 1982) (citations omitted). Thus, the primary focus of any challenge to a decision to withhold information as classified is normally upon the sufficiency of the description of that decision, rather than upon its reasoning.
In the seminal case of Vaughn v. Rosen, 157 U.S. App. D.C. 340, 484 F.2d 820 (D.C. Cir. 1973), cert. denied, 415 U.S. 977, 39 L. Ed. 2d 873, 94 S. Ct. 1564 (1974), our Court of Appeals set forth the standards for, absent evidence of bad faith, judging the sufficiency of such declarations. In the first place, the declaration must contain a detailed justification for the invocation of each exemption. See id. at 826-27. The declaration must also "specify in detail which portions are disclosable and which are allegedly exempt." Id. at 827. This is usually accomplished "by formulating a system of itemizing and indexing that would correlate statements made in the Government's refusal justification with the actual portions of the document." Id. (footnote omitted).
Upon initial inspection, the Government's Vaughn indexes seem to be satisfactory. In support of its invocation of Exemption 1, the FBI submitted two declarations by a top secret classification and declassification authority. See Declaration of Susan Delia Pruet (filed April 12, 1990) [hereinafter, "Pruet II"]; Declaration of Susan Delia Pruet (filed February 12, 1990) [hereinafter, "Pruet I"]. These declarations meet Vaughn's requirement of specificity. Each document is described separately. The description begins with a listing of basic information: document number or numbers, file references, date, number of pages, and a brief indication of the contents such as "question 36 from the briefing book" or "airtel from New York." Pruet I paras. 35, 52. The location of the segments withheld on the page is, if the page was disclosed,
identified. Finally, the contents of each segment withheld is described, and the justification for withholding is set forth in the following paragraph or in a referenced paragraph. Thus, the FBI's Vaughn index seems to adequately describe, separate, and index each segment of information withheld.
The justifications in that Vaughn index also seem to be adequately detailed. There are thirteen or so of them, normally of ten to twenty lines. See Pruet I paras. 20, 22, 31, 34, 41, 48, 54, 56, 58, 60, 61, 63, 65. Although these rationales are often couched in general terms such as "disclosure of a particular intelligence activity would permit hostile entities to assess" FBI counterintelligence operations, id. para. 34, that level of generality is often necessary to avoid "compromising the secret nature of the information." Hayden v. National Security Agency/Central Security Serv., 197 U.S. App. D.C. 224, 608 F.2d 1381, 1384-85 (D.C. Cir. 1979), cert. denied, 446 U.S. 937, 64 L. Ed. 2d 790, 100 S. Ct. 2156 (1980), quoting Vaughn v. Rosen, 484 F.2d at 826. The fact that justifications are adopted by succeeding paragraphs is not a problem either. See Keys v. United States Dep't of Justice, 265 U.S. App. D.C. 189, 830 F.2d 337, 349 (D.C. Cir. 1987) (holding that a Vaughn index need not contain "phony individualization" or "meaningless variations of language at each invocation of a specific exemption").
Although the FBI's Vaughn index seems on the surface to be adequate, the Archive contends that it in fact suffers from pervasive inadequacies. In particular, the Archive contends that the descriptions of the withheld documents are often too vague, that the justifications for withholding are not thorough enough, and indeed that at points the index is simply unintelligible. These defects, the Archive contends, require rejection of the Pruet Declarations. Because most of these criticisms are off the mark, they do not require rejection of the FBI's Vaughn indexes.
In another instance, the Archive argues that the description of documents 020007-10 is misleading because, while the description of the information withheld refers to specific operations and suggests that the whole document refers to a specific investigation, the disclosed portion of the document indicates a more general purpose of proposing "investigative action in an attempt to counter this KGB activity." No reasonable reader would, however, fail to look first at the disclosed portion, which appears at the beginning of the document. Once this is done, the contents of even this heavily redacted document becomes clear: The document proposes a response to identified KGB activity, and in the course of doing so it discusses the evidence of that activity.
The Archive's criticism of the justifications in the Vaughn index are similarly unpersuasive. For example, the FBI has justified its decision to withhold code names and designator symbols on the grounds that they may be used in a mosaic analysis by hostile intelligence officers. The Archive vigorously argues that the FBI cannot raise the specter of mosaic analysis to justify withholding any information; the FBI must offer "some more concrete reason to consider a particular piece of information to be a potentially sensitive part of a 'mosaic.'" Plaintiff's Motion for Summary Judgment at 15 n. 3. Such a reason is, however, plain from the FBI's explanation of mosaic analysis:
Mosaic analysis is an elimination of common denominator deductive process . . . whereby a trained hostile analyst, through review of innocuous type pieces of information within a document or a series of documents, will draw accurate conclusions . . . and determine the identity of a specific intelligence source, method or activity.
Pruet Declaration I para. 65. If a hostile analyst has code names or designator phrases, he or she has a "common denominator" for all the information concerning a particular source, method, or activity. With that common denominator, that analyst can place seemingly unrelated information into a picture of a particular aspect of the FBI's counterintelligence operations. Without such a common denominator, the analyst may not even be able to begin constructing the mosaic. The information is therefore properly withheld because it would contribute to mosaic analysis of other information.
At some points, the Archive seems to be attacking the use of abstract terms in Vaughn indexes. Our Court of Appeals has, however, found that abstract terms may be specific enough. See, e.g., Goland v. CIA, 197 U.S. App. D.C. 25, 607 F.2d 339, 351 (D.C. Cir. 1978) (finding "intelligence collection and operational devices" and "intelligence methodologies of a friendly foreign government" to be sufficiently detailed descriptions of withheld materials). Moreover, the Archive's criticism of specific passages is not persuasive. For instance, the Archive contends that the description "the identity of a particular type of intelligence target" is too vague. The FBI persuasively counters that it must withhold the identity of the type of target that it pursues because, if a hostile intelligence service knew what type of targets the FBI investigated, that service could "determine the criteria utilized by the FBI to decide what actions by a specific individual warrant the commencement of an investigation." Pruet I para. 41. The description in the Pruet Declarations is more than specific enough to demonstrate the applicability of the rationale offered.
Similarly, King v. United States Dep't of Justice lends little support to the Archive's claim that the Pruet Declarations are too vague. In that case, the Court of Appeals stated that "the availability of categorization does not . . . supplant the demand for particularity." King, 830 F.2d at 224 (footnote omitted). It did so, however, in the context of an index that did not describe each withheld segment, but rather placed beside each withheld segment a code referencing a justification for classification in the affidavit. See id. at 220. The Court of Appeals therefore stressed the need for a particular description of each segment withheld. See id. at 224. However, in this case the Pruet Declarations describe each segment withheld. See supra [Slip. op.] 5, 12.
Underlying the Archive's criticisms of the Pruet Declarations seems to be the suspicion that the FBI is not acting in good faith. Abstract statements such as the ones the Archive has criticized are sufficiently detailed to demonstrate the reasoning employed by the government; they are not sufficiently detailed to allow the Court or the plaintiff to verify that the documents actually say what the government says they do. However, courts must presume that a reasonably specific affidavit is accurate and truthful, unless the plaintiff presents evidence of inconsistencies in the affidavits or other evidence that the agency has acted in bad faith. See, e.g., Gardels v. CIA, 223 U.S. App. D.C. 88, 689 F.2d 1100, 1104 (D.C. Cir. 1982); Hayden, 608 F.2d at 1387; Weissman v. CIA, 184 U.S. App. D.C. 117, 565 F.2d 692, ...