Opinion for the Court filed by Senior Circuit Judge McGOWAN.
McGOWAN, Senior Circuit Judge:This is Harold Weisberg’s fourth — and final — appearance before this court in connection with the Freedom of Information *256Act suit he initiated in 1970 to uncover documents bearing on the assassination of President Kennedy. We find that the government has finally proven the adequacy of its search for all documents Weisberg has requested, and accordingly we affirm the District Court’s grant of summary judgment.
I
In 1970 Weisberg brought suit in District Court to compel the Federal Bureau of Investigation (FBI) to release spectrographic analyses of several items of evidence from the Kennedy assassination. See Weisberg v. U.S. Department of Justice, 489 F.2d 1195, 1196-97 (D.C.Cir.1973) (en banc) [“ Weisberg I”], cert. denied, 416 U.S. 993, 94 S.Ct. 2405, 40 L.Ed.2d 772 (1974). The FBI claimed that it had properly denied Weisberg’s request for these analyses under exemption 7 of the Freedom of Information Act (“FOIA” or “the Act”), a provision protecting investigatory files compiled for law enforcement purposes. See Act of June 5, 1967, Pub.L. No. 90-23, 81 Stat. 54, 55. Sitting en banc in 1973, this court upheld the FBI’s exemption claim. Weisberg I. The following year, however, Congress amended the Act and narrowed the scope of exemption 7. Act of Nov. 21, 1974, Pub.L. No. 93-502, § 2, 88 Stat. 1561, 1563-64 (codified at 5 U.S.C. § 552(b)(7) (1976)).
In 1975, on the first day the new amendment took effect, Weisberg brought suit to speed up compliance with a renewed request, broadened to seek analyses of certain Kennedy assassination evidence by means not only of spectrographic testing but of neutron activation testing as well.1 In his request to the Department of Justice, Weisberg demanded not to be misunderstood: he sought “only the final scientific reports on these tests. Not raw materials, not laboratory work, only the conclusions as embodied in the full report, or the report itself.” Appendix (“App.”) 342 (Nov. 27, 1974 letter from Weisberg to Deputy Attorney General). Noting that much of Weisberg’s request as so limited was available simply through the National Archives, the Attorney General wrote Weisberg six days after he filed suit to see if he would be interested in obtaining any of the FBI’s materials on the assassination beyond the “ ‘final reports’ ” Weisberg requested. R. 12 (attachment B). To discuss the Attorney General’s offer, Weisberg and his attorney attended a meeting with FBI officials on March 14, 1975; Weisberg said he would indeed be interested in the raw data pertaining to the spectrographic testing and the neutron activation analysis. R. 12 (Weisberg affidavit ¶¶ 23-24). To reduce his copying expenses, Weisberg asked not to receive a specified subset of the materials, ineluding the photographic plates from the spectrographic examinations. As for the records Weisberg did want, the FBI furnished several installments over the next few weeks. With an installment sent April 10,1975, the FBI told Weisberg it had fully responded to his FOIA request as supplemented by his requests at the March 14 meeting. R. 17 (attachment to exhibit 1).
Weisberg insisted that the FBI was still withholding relevant documents, a fact he attempted to establish through interrogatories. The District Court, however, quashed the interrogatories as “oppressive,” found that the government had “complied sub*257stantially” with Weisberg’s demands, and accordingly dismissed the case as moot. Transcript of May 21, 1975, Hearing at 22; Transcript of July 15, 1975, Hearing at 19. Noting that Weisberg had been denied the chance to establish his case through interrogatories or depositions, we reversed and remanded for further discovery. Weisberg v. U.S. Department of Justice, 543 F.2d 308 (D.C.Cir.1976) [“Weisberg II”]. Without attempting “to unravel the conflicting claims, assertions, or responses,” we found a material factual dispute in the question whether relevant but unreleased documents still existed. Id. at 310-11.
In remanding the case, we chided Weisberg for addressing his questions only to the present custodians of the files and suggested that he attempt to question those with firsthand knowledge of the tests conducted and records generated. Id. at 311. Thus, on remand, Weisberg deposed four FBI agents involved with testing of Kennedy assassination evidence. He also received answers to interrogatories and to requests for production of documents from the Department of Justice and the Energy Research and Development Administration. Finally, nine months after our remand, Weisberg attempted to depose FBI Special Agent John W. Kilty, who had executed two affidavits before the remand claiming that his searches of FBI files revealed no documents within Weisberg’s request that the agency had not already furnished. The District Court, however, found that the Kilty deposition would be “an unnecessary burden,” R. 39, and awarded the government summary judgment in an opinion that discussed in detail the results of the discovery to date and concluded that the FBI had released all relevant documents still in its possession. Weisberg v. U.S. Department of Justice, 438 F.Supp. 492 (D.D.C.1977) [“Post-Weisberg-II remand”].
On the third appeal to this court, the government offered two arguments in support of the District Court’s summary judgment award. First, it said that the two Kilty affidavits attested to thorough searches for all relevant documents. Second, even apart from the searches, it argued that the District Court opinion had shown that the records Weisberg said the agency had still not released either were no longer in the FBI’s possession or had never been created in the first place. We found these arguments insufficient and remanded for further proceedings. Weisberg v. United States Department of Justice, 627 F.2d 365 (D.C.Cir.1980) [“ Weisberg III ”]. In the first place, we found the Kilty search affidavits too conclusory to justify summary judgment: they “do not denote which files were searched or by whom, do not reflect any systematic approach to document location, and do not provide information specific enough to enable Weisberg to challenge the procedures utilized.” Id. at 371. We also explained that our suggestion in Weisberg II that Weisberg question those with direct knowledge of the tests conducted was not intended to limit the discovery on remand to those persons, id. at 367 n. 13, and that the Kilty deposition was therefore fully appropriate. As for the government’s second claim, we held that the evidence cited by the District Court, in the absence of an adequately documented search, was insufficient to show that no material factual disputes remained concerning whether the agency had released all relevant documents.
On this last point, we found the government unable to carry its burden of showing it had provided all relevant documents with respect to three specific categories. First, Weisberg pointed out that the FBI had not given him a copy of the spectrographic plate from a test of a lead smear found on a curbstone in Dealey Plaza; the test was apparently designed to determine whether the smear might have been caused by a bullet involved in the assassination.2 The *258government admitted that the plate once existed, but pointed to a statement by an FBI agent that the plate must have been discarded in one of the laboratory’s periodic housecleanings. See Weisberg III, 627 F.2d at 369. Second, Weisberg said he had not received the computer printouts from neutron activation analysis performed on specimen Q3, a bullet fragment found on the right front seat of the presidential limousine, and on specimen Q15, scrapings of residue from the car’s windshield. The District Court speculated that these printouts might not have been kept because they duplicated information that agents who analyzed the printouts had recorded on worksheets, copies of which had already been provided to Weisberg. Post-Weisberg-II remand, 438 F.Supp. at 503. Finally, FBI Special Agent Robert A. Frazier testified in a deposition after our remand in Weisberg II that he had asked another agent, possibly Paul Stombaugh, to examine the shirt President Kennedy wore the day of the shooting to determine whether two ragged holes near where the shirt buttoned together at the collar — one on the band of material bearing the buttons and one on the overlapping buttonhole band — coincided when the shirt was buttoned. Weisberg said that the results of an examination by Stombaugh, if indeed they were written down, had not been provided. The District Court noted that Frazier had testified before the Warren Commission that he had conducted this examination himself. Thus, the District Court was persuaded that Frazier’s recent deposition testimony, more than a dozen years after the events in question, was simply mistaken, and that Stombaugh had not performed any such test at all. Id. at 502-OS. We found the explanations with respect to all three areas of evidence to be plausible, but held that in the absence of a fully documented search they were not sufficient to satisfy the government’s burden on summary judgment of showing that no material disputed facts remained. We remanded for the government to provide further documentation about the extent of its search for relevant documents.
Our most recent remand led to significant activity in the District Court. On July 24, 1980, Weisberg requested the production of a wide variety of documents relevant to the agency’s search and to its general record-keeping policy, amounting to approximately 7,000 pages. R. 58 (plaintiff’s document request); R. 65 (opposition to plaintiff's motion to compel). At least in part on account of Weisberg’s poor health, the District Court ordered copies of the documents provided free of charge and not simply made available for inspection at the FBI. R. 69. On December 24, 1980, Weisberg moved to compel release of “all spectrographic plates of any item spectrographically tested” in connection with the Kennedy assassination investigation, R. 67, plates he had been offered at his initial meeting with the FBI in March 1975. The FBI provided copies as requested,3 R. 68, and later provided ‘identifying information to aid in the interpretation of the plates, R. 75 (exhibit C to de*259fendant’s response to interrogatories). By letter dated March 18, 1981, Weisberg inquired as to why some materials sought in his document request had not been provided and made new requests for production of documents. R. 73 (attachment to Lesar affidavit). On April 22, 1981, the FBI responded to the inquiries in detail, R. 89 (exhibit 3), and on May 27, 1981, sent fifty-two pages of the additional documents requested. Id. (exhibit 4).4
On April 6, 1981, Weisberg served interrogatories on the Justice Department, •which the latter answered on May 7, 1981. R. 72, 75. On May 26, 1981, over a year after the remand in Weisberg III, Weisberg noticed the deposition of Special Agent Kilty and scheduled it to take place at Weisberg’s house in Frederick, Maryland. R. 76. Kilty objected because Frederick was.more than forty miles from the place Kilty was served. See Fed.R.Civ.P. 45(d)(2). Although the District Court found that Kilty would normally be “perfectly entitled not to show up,” it ruled that the deposition could take place in Frederick. Transcript of June 5, 1981, Hearing at 2, 16. At that time, Weisberg’s attorney stated that he would like to depose Weisberg, id. at 12 (a request he later dropped and one that he had made and dropped before, see Post-Weisberg-II remand, 438 F.Supp. at 495 n. 2), and that “the whole matter would then be before the Court.”5 Transcript of June 5, 1981, Hearing at 13.
After the Kilty deposition, therefore, discovery was at long last complete. On September 8, 1981, the Justice Department moved for summary judgment, R. 89, and Weisberg filed a motion for an order requiring the FBI to make a more thorough search and to conduct certain new tests if the search failed to uncover specified records, R. 88. Weisberg opposed the motion for summary judgment on October 9, 1981. R. 93. Finally, on November 18, 1981, the District Court granted summary judgment for the defendants and denied Weisberg’s motion. App. 521.
II
On appeal Weisberg makes three main contentions: (1) the FBI has not conducted a search sufficiently thorough to uncover all relevant documents; (2) the agency’s representations that it has conducted such a search cannot be trusted because it has exhibited bad faith in its handling of Weisberg’s requests and in its handling of the Kennedy investigation generally; and (3) the FBI should be required to perform certain new tests if it does not locate the results of tests supposedly conducted earlier. Although the mass of Weisberg’s subsidiary arguments and the vast reams of his supporting documents call for a reasonably extended discussion,6 we have little trouble rejecting all three contentions and affirming the District Court’s grant of summary judgment.
The standard governing a grant of summary judgment in favor of an agency’s claim that it has fully discharged its disclosure obligations under FOIA is well-established. As we mentioned in Weisberg III, the agency bears the burden of showing that there is no genuine issue of material fact, even when the underlying facts are viewed in the light most favorable to the requester. Weisberg III, 627 F.2d at . 368 *260(quoting Founding Church of Scientology v. NSA, 610 F.2d 824, 836 (D.C.Cir.1979)). What the agency must show beyond material doubt is that it has conducted a search reasonably calculated to uncover all relevant documents. As we have recently made clear, “[t]he issue is not whether any further documents might conceivably exist but rather whether the government’s search for responsive documents was adequate.” Perry v. Block, 684 F.2d 121, 128 (D.C.Cir.1982) (per curiam) (emphasis in original); see Goland v. CIA, 607 F.2d 339, 367, 369 (D.C.Cir.1979) (per curiam on motion to vacate and petition for rehearing), cert. denied, 445 U.S. 927, 100 S.Ct. 1312, 63 L.Ed.2d 759 (1980). The adequacy of an agency’s search is measured by a “standard of reasonableness,” McGehee v. CIA, 697 F.2d 1095, 1100-01 (D.C.Cir.1983), and is “dependent upon the circumstances of the case,” Founding Church of Scientology, 610 F.2d at 834. The government may rely upon affidavits to show it has conducted a reasonable search, as long as they are “ ‘relatively detailed’ and nonconclusory and ... submitted in good faith.” Goland v. CIA, 607 F.2d at 352 (quoting Vaughn v. Rosen, 484 F.2d 820, 826 (D.C.Cir.1973), cert. denied, 415 U.S. 977, 94 S.Ct. 1564, 39 L.Ed.2d 873 (1974)); see Weisberg III, 627 F.2d at 370.
Before we consider the first of Weisberg’s claims that the District Court’s summary judgment award failed to satisfy these requirements — his attack on the thoroughness of Kilty’s searches — it will be helpful to put into perspective the current relevance of Weisberg’s allegations that the FBI created specific documents it has not released. At the time of Weisberg III, the FBI had not documented to our satisfaction that it had taken all reasonable steps to find materials responsive to Weisberg’s request. Therefore, to avert summary judgment Weisberg merely had to offer some reason to think the FBI might find a responsive document if it made the effort to conduct a thorough search. Or to put it another way, in the absence of a fully-documented, adequate search, the only way the agency could have justified the grant of summary judgment would have been to show beyond any material doubt that the documents Weisberg had isolated either (1) were in fact never created, or (2) had somehow been disposed of prior to Weisberg’s request. It was clear to us that the District Court’s theories about the fate of the spectrographic plate and the neutron activation printouts, and the nonexistence of the “Stombaugh report,” were not conclusive enough to satisfy this burden in,the absence of evidence that the FBI really tried to locate them, so we remanded for further discovery.
The tables turn, however, if the FBI can provide otherwise convincing evidence that it has conducted a thorough search. It then no longer needs to rely on these theories — that the plate might have been discarded, that Stombaugh might never have examined the President’s shirt — to show its compliance with the Act, for its documentation of a thorough search is enough by itself. Of course, evidence that relevant records have not been released may shed light on whether the agency’s search was indeed adequate; our next task will be to clarify what Weisberg now claims the FBI has still not located, so we can ensure that the agency’s search was especially geared to recover at least those documents. And there may be times when an agency’s inability to retrieve documents known or thought to be in its files is inherently unbelievable.7 But when we set out in the margins a brief summary of the known facts about the documents Weisberg claims the FBI still has, see infra notes 9-11, it will become quite clear that Weisberg’s allegations are not of this sort. The FBI’s explanations about the possible fate or dubious existence of these documents, though not *261alone sufficient to relieve the agency of its obligation to prove it has vigorously looked for them, are more than adequate to defend an otherwise reasonable search against the charge that a good faith thorough effort would necessarily have uncovered them.
Since our remand in Weisberg III, the government has given Weisberg one of the three sets of materials Weisberg claims the agency should still furnish him: the computer printouts containing data from the irradiation of specimens Q3 and Q15 for neutron activation analysis. Weisberg makes a feeble claim that the government’s failure to provide these materials at the outset is circumstantial evidence that the agency’s initial search was inadequate. See Goland v. CIA, 607 F.2d at 370 (per curiam on motion to vacate and petition for rehearing) (“[T]he discovery of additional documents is more probative that the search was not thorough than if no other documents were found to exist.”). The claim is feeble because it became clear on the last remand that Kilty located these computer printouts and showed them to Weisberg at their very first meeting in March 1975. Weisberg’s response was that he “did not want these items.” App. 138 (Kilty deposition). Kilty explained: “With regard to the computer tapes and notebooks with data in them, these were shown to [Weisberg] in a folder ... and he said something about the fact that I can’t make heads nor tails out of those things; I don’t want those things.” Id. Weisberg does not deny this fact, so the timing of the agency’s release of the printouts has nothing to do with the adequacy of the search, but only with the late date at which Weisberg indicated he might be interested in them.8
Thus, only two of the three sets of items discussed in Weisberg III are still in controversy, and it is primarily on these that Weisberg rests his claim that unreleased documents are still in FBI files. When we consider the details of the FBI’s searches, *262these are the documents we will want to be most certain the agency has endeavored to locate. First is the spectographic plate from testing of a lead smear on a Dallas curbstone, a plate the FBI concedes once existed but that an FBI agent speculates was discarded.9 With regard to this spectrographic test, Weisberg also complains that he has not received any notes the examiner might have made. The FBI has insisted from the start that it has no notes from this test aside from a worksheet it gave Weisberg long ago.10 The second record allegedly still in the FBI’s possession is a report by Stombaugh about whether two holes in the President’s shirt overlap when the shirt is buttoned.11 The government has continued to search for this document, *263even though it could have argued at the time of our last remand that the report, if it ever existed, is outside of Weisberg’s request.12
*264In this appeal, Weisberg makes a passing reference to several other tests whose results he says the FBI has not furnished him. Brief for Plaintiff-Appellant at 14. In all the years of this litigation, the first and only discussion of these tests appears in Weisberg’s affidavit filed on September 8, 1981, at the time the government moved for the summary judgment award appealed here. App. 278-86. Weisberg’s tardiness has ensured that there would be some uncertainty whether the government searched for them.13 All but one of these records pertain to stray bullets found near the scene of the assassination some time after the shooting occurred.14 The other concerns a scraping from a sidewalk mark, not to be confused with the curbstone lead smear whose spectrographic examination yielded a plate now missing.15 The results of tests conducted on these specimens, however, are plainly outside the scope of Weisberg’s request. See generally supra note 12. First, nothing in the documents Weisberg has produced indicates that any of these specimens was tested by spectrographic or neutron activation analysis.16 The documents seem to show that the bullets were sent to the lab to determine their caliber, not their composition, and although the scraping was sent for “examination purposes and possible identification,” there is no indication what tests were used. App. 451. Second, Weisberg’s complaint sought tests conducted “for the Warren Commission,” R. 1 (complaint ¶¶ 6, 10), yet each of the recently mentioned specimens about which Weisberg has furnished supporting documents was sent for testing after the Warren Commission issued its final report.17 We cannot *265rule out that the government, which in the past has been generous in searching for new items that Weisberg has added to his request, has already searched for any tests relating to these specimens and found none. On the basis of the evidence produced before the award of summary judgment, however, we cannot say that the government searched for these documents,18 so we leave proof concerning the scope of the government’s search in this respect for such time as Weisberg might file an appropriate request for them. If at that time the agency can show it has already looked for all documents Weisberg then seeks, it of course is under no obligation to look further.
Having now clarified the items Weisberg claims the FBI should still be able to find, we proceed to consider Weisberg’s major contentions: (1) that the FBI’s searches were inadequate, (2) that the FBI has responded to Weisberg’s request in bad faith, and (3) that the FBI should be ordered to conduct certain new tests on Kennedy assassination evidence.
Ill
Since our last remand there have been two developments concerning the FBI’s efforts to locate relevant documents. First, Kilty has conducted yet another search, bringing the total number to at least three, two in 1975 and the most recent one in 1981. Second, Weisberg deposed Kilty at length about the searches he conducted. This deposition provides a detailed explanation of the extent of Kilty’s searches, going far beyond the two 1975 affidavits we found on our last remand to be conclusory and incomplete. We find that the Kilty deposition demonstrates that the FBI’s searches have been more than adequate and that there is no longer any material doubt that the agency has taken all reasonable steps to locate the documents Weisberg seeks.
Kilty first attempted to recall the specifics of his searches in 1975. He testified that he “talked to a number of Agents— anyone who had any familiarity with this case at all — ” and asked them to give him any materials they had concerning the case.19 App. 54. He located two file cabinets in the FBI lab which he knew to contain relevant information based on his work with Special Agent Gallagher and on a discussion with Special Agent Frazier. App. 47 — 49, 52. Kilty said the sections of the file cabinets were organized in some fashion, though he could not then remember precisely how. App. 50. He went through relevant file sections of both “the Kennedy case and the Oswald case,” searching “tremendous numbers” of sections in “cart after cart after cart” of files. App. 53. He also looked through numerous sections of the headquarters’ central files. App. 61-62, 106.
In his 1981 search, Kilty says he looked in every place where documents of the kind Weisberg sought could be found. App. 140. Kilty relied not only on his own knowledge of where the lab stored certain items; he announced at a meeting of laboratory unit chiefs that he was conducting the search and asked about where “anything associated with the Kennedy Assassination case” might be located. Id. Kilty endeavored to find “anything we had that had any connec*266tion with [Weisberg’s request] at all,” and looked specifically for the items we discussed in our last remand. App. 139. He searched again the same file cabinets he searched in 1975 and then searched files containing old neutron activation and spectrographic data that were not specifically related to the Kennedy assassination, on the chance that the plates might be there. App. 144. Kilty specified the room numbers containing files and plate drawers that he searched. App. 140. All told Kilty devoted parts of ten days, and possibly more, to his search. App. 139.
With regard to the curbstone plate, Kilty affirmed that he consulted three spectrographic examiners to determine where Kennedy plates might be. App. 106-07. In his 1981 discussion with the unit chiefs, he called special attention to the missing plate. App. 140. Despite his search of all places where the plate or accompanying notes might be kept, Kilty says he found nothing he did not furnish. App. 129. Kilty also testified that he found no notes produced by Stombaugh concerning an examination of the President’s shirt, App. 131, even though he looked for “anything that was produced by Stombaugh that could be related to” the overlap of holes in that shirt and though he looked through “all the documents produced by Stombaugh.” App. 134.
Taking a different tack from the one we followed in Weisberg III, Weisberg does not fault these representations for being vague or conclusory. Instead, he directly challenges the adequacy of the search described, his main objection being that Kilty did not search the Dallas field office but only FBI headquarters in Washington. Kilty confirmed that, as the office of origin in the Kennedy case, the Dallas office received final lab reports on specimens it had forwarded to Washington for testing.20 App. 57-58; see App. 360 (FBI internal memorandum) (policy applicable to evidence received by lab after Feb. 10, 1964). FBI headquarters kept its own copies of these reports, however, App. 60, 147, so Dallas would not have any reports not retrievable by a thorough search of headquarters.21 Moreover, most of what Weisberg says the FBI has still not released are not final reports but a spectrographic plate and notes from lab examinations, and nothing in the record suggests that material of this sort ever found its way to Dallas.22 To the contrary, Kilty testified that spectrographic plates are kept in drawers in the FBI lab, App. 99, and that an examiner’s notes from a test at the FBI laboratory would not in his “wildest imagination” be found in Dallas. App. 135. All of this supports the FBI’s contention that it would be fruitless and unreasonable to search the Dallas field office for the items Weisberg still seeks.23
*267The same problems plague Weisberg’s assertion that Kilty should have searched other divisions in FBI headquarters and a special file room. Weisberg says these places were occasionally sent copies of outgoing lab reports or other unspecified materials. Kilty, however, has already searched where the original lab reports are kept, and Weisberg has not shown that the type of materials he now mainly stresses — plates and examiners’ notes — were stored in these other locations. Kilty has testified that he consulted with his colleagues who had personal knowledge of the Kennedy investigation, or the type of tests used in that investigation, to determine all places where relevant records might be kept; Weisberg has produced no evidence raising material doubts about the accuracy or completeness of the information Kilty received. His other criticisms of the search are entirely without merit,24 so we have little trouble finding that the FBI has carried its burden of showing it has taken all reasonable steps to find the materials Weisberg has requested.
IV
What Weisberg is unable to show by a direct attack on the search itself he attempts to show indirectly by extensive allegations that the FBI has treated his request in bad faith. This history of bad faith, Weisberg says, renders suspect the agency’s representations about the thoroughness of its search and makes necessary a full trial to establish whether the agency has indeed discharged its FOIA obligations in full. We find nothing in Weisberg’s allegations, however, to suggest that the agency has been anything less than forthright and cooperative in its handling of Weisberg’s request. Indeed, although there is some evidence to suggest that the agency balked at some of Weisberg’s earliest inquiries, once Congress amended FOIA in 1974 and. Weisberg submitted his present request, the FBI has in many instances been unquestionably generous in its efforts to assist Weisberg’s investigation. It should be remembered, for example, that Weisberg showed no interest in the raw materials of lab examinations, and indeed explicitly said he did not want them, until the Attorney General offered to have the FBI supply him with materials he had not requested but might nonetheless find useful. See supra p. 1347. It is therefore somewhat ironic that Weisberg’s objections that he has not received certain raw materials have since formed the mainstay of his efforts to avert summary judgment and prolong discovery. Moreover, much of this discovery, which the agency has patiently endured, has borne only the slightest relation to whether the FBI has failed to release pertinent documents and more closely resembled a private inquiry into the findings of the Warren Commission. In any event, Weisberg’s attacks on the agency’s good faith, though voluminous, are entirely unpersuasive. We consider his major contentions in turn.
*268Weisberg’s first allegation is that the government has given him false information about the examination of Q3 and Q15 by neutron activation analysis. To explain Weisberg’s claim, we must clarify the extent to which these specimens were examined. Although the FBI admitted before our last remand that Q3 and Q15 had been irradiated in a nuclear reactor, the virtually blank worksheets the agency provided Weisberg showed that the results of the irradiations yielded no useful information for analysis. After seeing the computer printouts from the irradiation of these specimens, the examiners aborted the testing of Q3 and Q15 and did not produce the detailed worksheets, the tables, or the reports that accompany an examination by neutron activation analysis. In the terminology of the scientists familiar with neutron activation analysis, this meant that Q3 and Q15 were not “examined” at all. As Kilty explained, “examination” by neutron activation analysis “is the total analysis and handling of a specimen which produces some kind of a report or final comment or final opinion regarding the totality of all the tests and material that you went through on that specimen.” App. 96-97.
On June 23, 1975, Kilty executed an affidavit clarifying that only emission spectroscopy was used on a small set of specimens including Q15; “NAA was not used in examining” those items.25 App. 167-88. Later, Weisberg asked by interrogatory whether “neutron activation testing [was] done” on items other than a specified set of specimens that did not include Q3 or Q15. The- Justice Department answered “No.” App. 194 (answer to interrogatory no. 19). Weisberg says that Kilty and the Justice Department were lying, because Q3 and Q15 were irradiated to yield the computer printouts Weisberg now has. He suggests that the definition Kilty elaborated in his deposition, which excludes the stillborn examination of Q3 and Q15 from the realm of “neutron activation analysis” or “neutron activation testing,” is somehow disingenuous, yet we cannot agree. It is hard to imagine why the agency would gerrymander a definition to conceal the existence of the printouts for Q3 and Q15, when it offered those printouts to Weisberg at the very outset of this litigation. Weisberg now insists that the specimens were at least subjected to neutron activation “testing” if not to neutron activation “analysis.” But we see nothing odd or evasive in the FBI’s assumption, apparently shared by the scientists familiar with these procedures, that the salient aspect in any “testing” to determine a specimen’s elemental composition by this method is the analysis of how the specimen behaves after irradiation, and that if this analysis cannot take place the specimen has not been “tested.” This assumption was reasonable enough when it led Weisberg explicitly to define his initial request for “tests” to include only the examiners’ analysis, and not-the raw data, R. 1 (exhibit D), so we cannot see why it was unreasonable when it underlay the FBI’s answers to questions related to that very request.
Weisberg mounts a second attack on Kilty’s veracity in an effort to show the agency’s bad faith. Kilty testified that, in both his 1975 and 1981 searches, he looked for responsive documents in file cabinets located in the FBI laboratory. See supra pp. 1356-1357. Weisberg claims that the admission that the FBI lab has “files” shows that Kilty lied, not in this case, but in a Weisberg FOIA suit for materials on the assassination of Martin Luther King, in which Weisberg says Kilty denied that the FBI *269lab kept its own “files.” See Kilty Deposition of Oct. 12, 1979, at 7, 15, 20, Weisberg v. U.S. Department of Justice, C.A. No. 75-1996 (D.D.C. orders entered Dec. 1,1981, & Jan. 5,1981), cross appeals docketed, Nos. 82-1229 & 82-1274 (D.C.Cir. Mar. 4,1982, & Mar. 15,1982). There is absolutely nothing to this argument. The most casual glance at that deposition reveals that Kilty made perfectly clear that the FBI kept some records, such as neutron activation analysis raw data, within the lab. Id. at 13, 19-21. His statements that the lab did not keep “files” reflect no more than his stated preference in that deposition to limit the use of that term to the collections of information in the headquarters’ central filing system, which contained the bulk of materials pertaining to the King case, including the final results of FBI lab examinations. See id. at 19-21. Regardless whether Kilty referred to the places in the lab that he searched in this case as “files,” “storage locations,” or anything else, nothing in his revelation that the lab kept documents that might have been responsive to Weisberg’s request conflicts with his testimony in the King case.
Weisberg’s third attack on the FBI’s good faith is that the agency attempted in 1980 to revoke the general fee waiver it granted Weisberg in 1978 to enable him to receive without charge all Kennedy assassination materials it released thereafter. After our last remand, Weisberg requested the production of documents totaling over 7,000 pages, many of which the government said it had given Weisberg before. R. 65. The agency said it thought Weisberg had gone too far, and that it would make the documents available but would not copy them free of charge. Id. Weisberg responded by limiting his request to documents he had not previously been given. R. 66. Once the request was so limited, the District Court ordered the Justice Department to provide those materials for free in the interest of “the just, speedy, and inexpensive termination of this long pending action.” R. 69.
Nothing in the government’s belief that Weisberg was abusing his fee waiver strikes us as evidence of its bad faith. It was certainly reasonable for the government to insist that it not pay to copy documents it had given Weisberg before. And though the government did not volunteer to reinstate the fee waiver once Weisberg narrowed his request to exclude those documents Weisberg already had, its brief attempt to exert some restraint on Weisberg’s discovery efforts is certainly understandable: we have already stated our view that Weisberg in many instances exceeded the scope of reasonable discovery. See supra p. 1358. It is beside the point whether, as Weisberg asserts, the FBI provided some of these same documents to other litigants free of charge. The government could reasonably have thought Weisberg was abusing the fee waiver in ways that other litigants presumably were not, so its attempt to treat Weisberg differently was, if not unquestionably laudatory, entirely understandable.
Weisberg’s fourth allegation of bad faith is directed not to the FBI’s handling of his FOIA request but more generally to its handling of the Kennedy investigation. He charges that the FBI willfully concealed a relevant fact from the Warren Commission and that this past concealment creates a present motive for the agency to withhold the spectrographic plate from the examination of a lead smear on a Dealey Plaza curbstone. We find that Weisberg’s theory is simply too attenuated and unsubstantiated to raise a genuine issue of material fact concerning the FBI’s compliance with its FOIA obligations. Weisberg points to the affidavit of a bystander at the assassination who noticed a “mark” in the sidewalk near where he was standing, R. 47 (affidavit of James T. Tague ¶ 5), described in his interview with the FBI as a “chip,” id. (Tague affidavit exhibit C). When the bystander returned to the spot some six months later, it did not appear to him that the mark was still present. Id. (Tague affidavit exhibit I at 69). An FBI agent who viewed the curbstone in August 1964 also said that no “mark or nick” was visible. App. 415 (Aug. 5,1964, memorandum). Closer examination *270revealed, however, that the curbstone did bear a “mark,” to which adhered “[s]mall foreign metal smears.” App. 411 (Aug. 12, 1964, letter from Hoover to Rankin). The FBI dug up the curbstone, brought it to the FBI lab, and spectrographically tested the metal smears, which were determined to be “essentially lead with a trace of antimony.” Id.
The first link in Weisberg’s chain of argument is that, because the “chip” was less visible six months after the bystander first observed it (so much so that by August 1964 it could be called no more than a “mark” or a “smear”), it must have been patched before the FBI removed the curbstone for testing. Weisberg does not discuss the possibility that a small chip on a public sidewalk might be subject to some erosion over a six-month period. Moreover, he does not explain whether he thinks the lead smear adorned a patch consisting of some other material or whether the lead smear was the patch.26 In any event, Weisberg’s second link in the argument is that the FBI discovered that the curbstone had been altered but concealed this fact from the Warren Commission. Weisberg adduces no evidence to support this point, and he does not make clear whether the FBI would have discovered the alleged patch on the curbstone by its spectrographic examination of the metal smear, or by some other examination it might have conducted on the curbstone. The latter would seem the only remotely plausible theory; it is hard to imagine how the results of testing on metal smears would lead the FBI to “realize” that the curbstone had been patched. Yet Weisberg seems to rely on the former theory when he asserts the third and final link in his argument: that the FBI is withholding the spectrographic plate to conceal that it knew the curbstone was patched. Especially when one realizes that all three links in the chain are necessary to bring Weisberg’s allegations within the scope of this lawsuit, it is plain that his theory is too unsubstantiated and incredible to create any material doubt whether the FBI has furnished all relevant records within its possession.
Weisberg’s fifth allegation of the agency’s bad faith relies on a document that Weisberg says he first discovered on January 12, 1982, when the FBI filed it in separate litigation, and that he therefore could not present to the District Court before it granted summary judgment on November 18, 1981.27 We have said before that “[fjactfinding and the creation of a record are the functions of the district court,” and that the “proper procedure for dealing with newly discovered evidence is for the party to move for relief from the judgment in the district court under rule 60(b) of the Federal Rules of Civil Procedure.” Goland v. CIA, 607 F.2d at 371 (per curiam on motion to vacate and petition for rehearing). Although the one year in which litigants may file rule 60(b) motions based on newly-discovered evidence — which runs even during the pendency of an appeal, Greater Boston Television Corp. v. FCC, 463 F.2d 268, 280 & n. 22 (D.C.Cir.1971), cert. denied, 406 U.S. 950, 92 S.Ct. 2042, 32 L.Ed.2d 338 (1972) — has by now elapsed, Weisberg discovered this document in time to make such a motion before the District Court. Nonetheless, he declined to follow the preferred approach and chose instead to suggest that we either remand for the District Court to consider the evidence, as part of our general power to remand for further proceedings “as may be just under the circumstances,” 28 U.S.C. § 2106 (1976), or take judicial notice of the document and consider it ourselves, see Landy v. Federal Deposit Insurance Corp., 486 F.2d 139, 150-51 (3d Cir.1973), cert. denied, 416 U.S. 960, 94 S.Ct. 1979, 40 L.Ed.2d 312 (1974). We deferred until now any decision on how we should treat this *271evidence, though we did give Weisberg leave to refer to the nonrecord document in his reply brief, in the event we decided that the document’s purported significance would affect how we treated it on appeal.28 It is now apparent that the new evidence in no way casts doubt on the appropriateness of summary judgment. Therefore, we reject Weisberg’s attack on the agency’s good faith without assessing the potential significance of his refusal to follow the procedure the federal rules establish for consideration of newly-discovered evidence or defining with exactitude the reach of our judicial notice powers.
In this document, the so-called “Shea memorandum,” the director of the Justice Department’s Office of Privacy and Information Appeals, in opposing a withdrawal of a fee waiver for Weisberg’s administrative FOIA requests, expresses concern that the FBI had interpreted too narrowly the scope of Weisberg’s administrative request for documents relevant to the Martin Luther King assassination. Shea contends that Weisberg's administrative request, as opposed to his lawsuit, was for any documents related to any aspect of the King assassination' or investigation, but that the agency failed to look in files that, though not organized by subject matter in the King section, might still contain documents that are “related” to the King assassination. Obviously, the central point of the Shea memorandum is not directly relevant to the request underlying the present lawsuit, and indeed, even the District Court adjudicating the lawsuit to enforce a request for King documents found the Shea memorandum unpersuasive:
Mr. Shea clearly did not share the FBI’s interpretation of the scope of plaintiff’s numerous administrative requests. But his comments do not indicate disagreement with the scope of this action. Neither do they indicate that the FBI deliberately deceived plaintiff, the Court or Congress by withholding information. Mr. Shea made these comments in opposing the withdrawal of a fee waiver by the FBI for plaintiff’s administrative requests under the Freedom of Information Act.29
Weisberg v. U.S. Department of Justice, C.A. No. 75-1996 (D.D.C. June 22, 1982) (memorandum opinion denying motion to amend orders of Dec. 1, 1981, & Jan. 5, 1982) (reprinted in Appellees’ Supplemental Brief, exhibit A), cross appeals docketed, Nos. 82-1229 & 82-1274 (D.C.Cir. Mar. 4, 1982, & Mar. 12, 1982) (appeal from Dec. 1, 1981, & Jan. 5, 1982, orders). Although Shea does assert that some documents that are “factually, logically and historically relevant to the King and Kennedy cases” might not have been located, Reply Brief for Plaintiff-Appellant, Addendum 7, at 2, it is apparent that the remark in no way concerns the Kennedy request underlying the present suit, which is not for any materials generally related to the assassination, but for specified scientific tests. Locating particular scientific tests from an investigation is obviously a more straightforward task than locating documents “related” to an investigation in some broad sense. All told, therefore, we cannot say that the Shea memorandum, by itself or in conjunction with the other attacks Weisberg has made, casts sufficient doubt on the believability of the government’s representations to create a genuine issue of material fact about whether the government has carried its burden of showing full compliance with Weisberg’s request.
V
Having disposed of Weisberg’s attacks on the agency’s search and on its good *272faith in the handling of his request, we now turn to his claim that the District Court should have ordered the FBI to conduct certain tests on Kennedy assassination evidence. Specifically, he requested that, “in the event that a thorough search fails to uncover allegedly missing records pertaining to the spectrographic testing of the Dealey Plaza curbstone allegedly struck by bullet, and FBI Special Agent Paul Stombaugh’s examination of President Kennedy’s shirt collar,” the court order the FBI “(1) to perform tests to determine whether said curbstone was patched, and (2) to conduct an examination of President Kennedy’s shirt collar to determine whether the alleged bullet holes in it coincide.” R. 88 (motion at 1). The Supreme Court, however, has made clear that FOIA puts an agency under no obligation to create documents, but “only requires disclosure of certain documents which the law requires the agency to prepare or which the agency has decided for its own reasons to create.” NLRB v. Sears, Roebuck & Co., 421 U.S. 132, 162, 95 S.Ct. 1504, 1522, 44 L.Ed.2d 29 (1975); accord Krohn v. Department of Justice, 628 F.2d 195, 197-98 (D.C.Cir.1980). Moreover, even if an agency creates a document, FOIA requires disclosure only of records “for which [agencies] have chosen to retain possession or control.” Kissinger v. Reporters Committee, 445 U.S. 136, 151-52, 100 S.Ct. 960, 968-69, 63 L.Ed.2d 267 (1980). Weisberg has cited one case in which a district court ordered an agency to reconstruct certain information relevant to the defense in a criminal trial when the agency “inadvertently,” but undeniably, destroyed requested documents after the request was made and the litigation begun. Levine v. United States, 34 Ad.L.Rep.2d (P & F) 633 (S.D.Fla. Mar. 22, 1974) (correct date from copy of original document; reporter gives date as Mar. 2, 1974). Nothing suggests, however, that results from the tests he seeks to have the agency perform, if they ever existed, were within the agency’s possession at the time of the request. See Kissinger, 445 U.S. at 155, 100 S.Ct. at 971 (State Department did not have “possession or control of the documents at the time the requests were received”). Thus, without saying whether we would go as far as or further than the case Weisberg cites, we do say that Weisberg has offered no reason to depart from the general rule that FOIA requires an agency to do no more than disclose nonexempt documents within its possession that are retrievable by a reasonable search.
As for the' first tests Weisberg seeks, nothing in the record even suggests they were ever conducted. He seeks to have the FBI perform “[w]hatever tests are needed” to determine whether the Dealey Plaza curbstone was patched. R. 88 (memorandum of points and authorities at 14). Although the FBI once conducted a spectrographic examination of a lead smear on that curbstone, Weisberg has not shown that this test was relevant to whether the curbstone was patched; its apparent purpose was to determine whether the metal deposit could have been left by a bullet. Indeed, Weisberg himself alludes to Kilty’s testimony that x-ray fluorescence would be the appropriate method for determining whether the sidewalk was patched, id.; see App. 110-11 (Kilty deposition), yet there is no evidence that the FBI ever performed such a test. As for the second test he seeks, Weisberg has apparently already received the results of such an examination, though they are contained in a report by Frazier, not Stombaugh. See supra note 11. A recreation of any test Stombaugh might have conducted would still not be the “Stombaugh report,” and we fail to see what purpose would be served by ordering the FBI to conduct a test to generate information Weisberg has already received. Moreover, the FBI’s most recent search has cast serious doubt on whether Stombaugh ever performed such a test, and if he did, Weisberg has raised no suspicion that the agency disposed of the results under circumstances that would warrant the extraordinary relief Weisberg seeks.
VI
This court has bent over backwards to ensure that Weisberg have a reasonable *273opportunity to develop his case against the FBI. We have strictly held the government to the level of proof it is required to satisfy in cases of this type. We have sorted through the mass of Weisberg’s contentions, in- many instances with more care than he has demonstrated in making them. At long last, we are convinced that the government has carried its burden and that the District Court’s most recent decision to grant summary judgment was fully appropriate. We therefore bring this long and difficult litigation to a close by affirming that decision.
It is so ordered.
. Spectrographic and neutron activation analyses are designed to determine the composition of small samples of materials. In spectrographic analysis, samples are sparked or burned to produce a spectrum of light that is exposed to a photographic plate; the plate may be analyzed to measure elements present in the sample. In neutron activation analysis, samples are bombarded in a nuclear reactor; the energy the samples then emit may be measured for the same purpose.
Weisberg initially directed his 1975 request to the FBI and to the Energy Research and Development Administration (ERDA), the successor agency to the Atomic Energy Commission (AEC). The AEC granted the FBI access to a nuclear reactor as necessary for the latter’s neutron activation analysis. The AEC, however, did not retain results of tests conducted at its facilities, so ERDA no longer figures in this litigation. See Record (“R.”) 29-30 (ERDA responses to request for production of documents and to interrogatories); Weisberg v. U.S. Department of Justice, 438 F.Supp. 492, 493 n. 1 (D.D.C.1977).
. As noted earlier, at the March 14, 1975, meeting Weisberg specifically asked not to receive copies of the spectrographic plates themselves. Although Weisberg now claims that he would have requested the plates had the FBI not overstated the costs of copying them, App. 215 (Weisberg affidavit ¶ 53), he admits that he declined to have any plates copied at that time. *258It was apparently not until shortly before our decision in Weisberg III, in fact, that Weisberg again mentioned one of the plates: he noted in his opposition to the government’s 1977 motion for summary judgment that, according to an FBI internal memorandum, the FBI’s search had turned up neither notes from the curbstone spectrographic testing nor the plate itself. R. 47 (opposition at 10). The District Court seems to have assumed that Weisberg was objecting to having received neither the notes nor the plate; it endeavored to explain that an FBI agent had speculated that the plate had been discarded and that this accounted for the search’s failure to locate the plate. See PostWeisberg-II remand, 438 F.Supp. at 503-04. Emboldened on his appeal to this court, Weisberg referred to the spectrographic plate as an example of “materials which are within the scope of Weisberg’s request but which have not been given to him.” Brief for Plaintiff-Appellant at 33, Weisberg III. Rather than object that Weisberg was now complaining about failing to receive an item that fell within a category of materials he had originally asked not to be given, the government rested instead on its representation that its search had failed to turn up anything relating to the curbstone spectrographic testing beyond what it had already given Weisberg.
. Still missing was the curbstone spectrographic plate we discussed in Weisberg III, which the agency has maintained from the outset it has been unable to locate.
. On April 6, 1981, Weisberg filed another motion to compel pertaining to four items discussed in Weisberg’s March 18 letter. R. 73. The government opposed the motion, on grounds similar to those explained in its April 22, 1981, letter. R. 74. By June 5, 1981, Weisberg’s attorney had decided not to press the matter further, Transcript of June 5, 1981, Hearing at 4, 12, and the District Court consequently denied the motion to compel, id at 17.
. In March 1981 Weisberg expressed an interest in serving interrogatories upon Special Agents Stombaugh and Heilman. Transcript of Mar. 6, 1981, Hearing at 12. The court directed Weisberg’s counsel to do so within 30 days. He responded, “I certainly will,” id. at 13, but neither did so nor asked for more time to do so.
.Indeed, Weisberg has received so many documents from the FBI — in voluntary releases and in this and other litigation — that he says his own searches of his own files have not necessarily produced all materials in his possession that are relevant to his contentions. See R. 50 (Weisberg affidavit ¶ 10).
. For instance, we noted in one case: “[S]ince NSA’s prime mission is to acquire and disseminate information to the intelligence community, it seems odd that it is without some mechanism enabling location of materials of the type appellant asked for, particularly with identifying details as extensive as those furnished.” Founding Church of Scientology, 610 F.2d at 835. This combined with several other factors to create a triable issue of fact in that case.
. Because Weisberg explicitly said at the March 1975 meeting that he did not want the printouts, Kilty necessarily understood those materials to fall outside Weisberg’s request at that meeting for the “available material relating to examination of the windshield of the President’s automobile [e.g., Q15], and examination regarding metal fragments from the President’s automobile [e.g., Q3],” see App. 348 (Mar. 24, 1975, FBI internal memorandum). Weisberg apparently did not mention the printouts again until his 1977 opposition to the government’s motion for summary judgment, when he asserted he had not been “given” any computer printouts. R. 47 (opposition at 9). This was the same memorandum in which Weisberg made the first mention of a spectrographic plate in over two years, see supra note 2, and was the last filing by either party before the District Court granted the summary judgment we reversed in Weisberg III. Therefore, the government did not become aware that Weisberg was interested in the printouts until the District Court was deciding the summary judgment motion. Because the court granted the motion, it was not until we reversed and remanded in Weisberg III that the government was obliged to check whether printouts for Q3 and Q15 were among the mass of pages it had previously shown Weisberg. It duly provided these documents on remand, so nothing in the timing of its release reflects on the agency’s thoroughness or good faith.
It is also worth noting that it was the District Court, not the government, which speculated that the printouts might have been discarded because they were “duplicative” of the data contained in the worksheets the FBI had already provided. See Post-Weisberg-II remand, 438 F.Supp. at 503 (citing Gallagher deposition). Agent Gallagher, who conducted the tests, testified that the FBI “[p]robably” still had the printouts, “unless they were judged to be worthless and not kept.” Gallagher Deposition at 92. He never said he discarded them or that he looked for them. When Weisberg belatedly indicated that he wanted copies of the printouts after all, the District Court used Gallagher’s testimony as evidence that those particular printouts had been discarded, an inference we found insufficiently established to survive summary judgment. The government, however, had no chance to discuss the printouts before the District Court made its finding and never represented that it did not have them. Indeed, on appeal it brought to our attention the fact that it had in its possession “continuous, folded tapes — similar to stenographic tapes — containing NAA raw data, that this material was shown to both appellant and his. counsel by Kilty in early 1975, and that appellant said then that he did not want copies of this material.” Brief for Appellees at 21 n. 17, Weisberg III. But because it waited to search that large collection of documents until it was obliged to do so, it could not tell us definitively at the time of Weisberg III whether the Q3 and Q15 printouts were among them.
. William R. Heilman, a former FBI agent, told Kilty that if the curbstone spectrographic plate was not with the other Kennedy assassination plates it must have been discarded. App. 100 (Kilty deposition). When asked if it was unusual that only one plate be discarded, Kilty responded: “Well, this was done completely at a different time and by a different examiner that [sic] did all the other work in this case and he may not have attached his plate to where the other plates were.” App. 101. Weisberg’s brief inexplicably cites this part of Kilty’s deposition as evidence that he “conceded” it was unusual that only one plate was destroyed. Brief for Plaintiff-Appellant at 13. Weisberg’s affidavit cites another portion of the deposition to show that “[f]inally, [Kilty] did agree ‘that it would be unusual to have one plate destroyed.’ ” App. 264 (Weisberg affidavit ¶ 235). The quotation is grossly out of context. Kilty said the curbstone spectrographic plate “obviously” was not filed with the others, though he had no personal knowledge of when it was tested. App. 103. The deposition proceeded:
Q. Okay. Now, if it were filed with the other plates, it would be most unusual if that plate, alone, were destroyed, would it not?
MR. COLE: You’re assuming the hypothetical, Mr. Lesar?
MR. LESAR: Yes.
A. I agree that it would be unusual to have one (1) plate destroyed.
Q. Almost sinister.
MR. COLE: Objection, Mr. Lesar.
App. 104 (emphasis added). Thus, the answer was given to an explicitly hypothetical question at variance with Kilty’s understanding of the facts.
. In support of his claim that the FBI is knowingly withholding notes from this test, Weisberg points to a June 16, 1975, FBI internal memorandum stating that an “exhaustive search of pertinent files, and storage locations has not turned up the spectrographic plates [from testing of the curbstone lead smears] nor the notes made therefrom.” App. 182. Someone has circled the word “notes” and drawn a line to a scribbled marginal notation reading “block diagram [with] symbols and relative concentrations.” Id We simply cannot conclude that this creates a material disputed fact about whether the FBI is knowingly withholding notes. The inference Weisberg would apparently have us draw is that the marginal notation described particular notes the writer had seen rather than the general format notes of that type would take if indeed they existed. It must be remembered that the only unambiguous message of the memorandum is that an exhaustive search did not reveal any notes. Although we are bound to give Weisberg every fair inference from the facts adduced, we believe he asks us to draw an inference of duplicity far beyond what the notation and surrounding evidence concerning the agency’s good faith could conceivably warrant.
.Unlike the spectrographic plate, there is serious doubt whether this document ever existed. Weisberg asserts it does based on (1) Frazier’s testimony in 1977 that he thought he had asked Stombaugh to conduct such an examination, Frazier Deposition at 60-62; and (2) the fact that Stombaugh was qualified to conduct “Microscopic Analysis (Fibers)” but Frazier was not, App. 188 (Justice Department’s answer to interrogatory no. 5(c)). Weisberg, however, has never taken Stombaugh’s testimony (and thus never run the risk of establishing for certain that the document does not exist), and there is much to indicate that the government’s search for a “Stombaugh report” has been a wild goose chase. When he appeared before the Warren Commission in 1964 —at a time much closer to the events in question — Frazier testified that he had conducted this examination himself. See Post-WeisbergII remand 438 F.Supp. at 502. The FBI has produced a document prepared by Frazier which describes the hole near the President’s shirt collar and says, “[t]his hole is through both the button and buttonhole portions of the shirt due to the overlap.” R. 89 (exhibit 6). Thus, Frazier might well have examined the shirt himself, on the theory that buttoning a shirt and looking to see if two holes lined up did not require a fibers expert; when he said Stombaugh examined the President’s shirt, he might have had in mind instead Stombaugh’s examination of the shirt worn by Oswald. See Post-Weisberg-II remand, 438 F.Supp. at 502-OS. Moreover, although Frazier testified that Stombaugh might have made a report to him, he did not specify whether it was in writing, and there is evidence elsewhere in the record *263that tests forming ingredients of Frazier’s reports were sometimes delivered orally, see Gallagher Deposition at 83, 100. Although we held in Weisberg III that the evidence suggesting that the “Stombaugh report” never existed did not excuse the FBI from conducting a thorough search, it should be clear that Weisberg’s own evidence suggesting the existence of such a document is not sufficient to raise material suspicions once the government’s thorough search has failed to locate such a report.
. Obviously, it does little good for the government to note that a document falls outside of the original request if Weisberg can simply initiate another request and thereby force it to search for the new document. It appears that the government has generally thought it best simply to search for anything Weisberg • requested along the way in order to end the matter once and for all. Nonetheless, Weisberg’s ever-expanding view of the scope of his original request has prolonged this litigation and must have been a source of continuing irritation for the government.
Weisberg’s complaint sought to force compliance with the requests contained in two letters. See R. 1 (complaint ¶¶ 6, 10). First was a letter of September 19, 1974, to the AEC requesting final reports of “any” tests that agency conducted in connection with the Kennedy assassination. R. 1 (exhibit D). The AEC, however, apparently did no more than provide facilities to enable the FBI to perform neutron activation analysis. See supra note 1. Second was a letter of November 27, 1974, to the Justice Department requesting final reports from spectrographic and neutron activation analyses. App. 342. The complaint refers to this letter as a request for spectrographic analysis and “other scientific tests conducted for the Warren Commission.” R. 1 (complaint ¶ 6). The letter makes clear that the request was not for any scientific testing in addition to spectrographic analysis, but only for additional tests of a specific type, neutron activation analysis. Weisberg’s January 15, 1975, letter appealing the “denial” of his request refers to a letter of December 6, 1974, in which Weisberg supposedly requested scientific tests in addition to spectrographic and neutron activation analysis. R. 1 (exhibit B). No December 6, 1974, letter appears in the record, however, and it is evident that Weisberg was simply exaggerating the scope of his November 27 request, which he mistakenly thought was dated December 6, the date the FBI director said he received the November 27 request. See R. 1 (exhibit A); App. 153.
At the March 14, 1975, meeting held after Weisberg initiated suit, he expanded his request to include raw data, but apparently continued to express interest only in spectrographic and neutron activation analysis. An FBI internal memorandum dated March 24, 1975, lists what Weisberg requested at the meeting. The only requests not explicitly limited to spectrographic or neutron activation analysis were requests for (1) “available material relating to examination of the windshield of the President’s automobile [e.g., Q15], and examination regarding metal fragments from the President’s automobile [e.g., Q3],” and (2) “laboratory examination data” concerning testing of the curbstone. App. 348. Subsequent discussion of these items has been limited to spectrographic and neutron activation analyses. Moreover, Weisberg’s affidavit filed June 3, 1975, says he sought at the March 14 meeting to examine “all the spectrographic and neutron activation materials and select which documents I wanted copied.” R. 12 (Weisberg affidavit ¶ 23).
As the litigation progressed, however, Weisberg became more casual and expansive in describing his request. Although the specific items he claimed he had not been furnished were usually limited to spectrographic or neutron activation analysis (with occasional mention of microscopic analysis), he began to say his request was for all reports or data on all items of evidence “scientifically]” tested in connection with the assassination. See Transcript of May 2, 1975, Hearing at 4; R. 12 (Weisberg affidavit ¶ 16). After our remand in Weisberg II, when Frazier mentioned that Stombaugh might have examined the holes in the President’s shirt, Weisberg objected that he had not received any such report, R. 47 (opposition at 10), though the examination was not by spectrographic or neutron activation analysis. The District Court dismissed this fact not by saying the document fell outside the request, but by explaining why no such document existed. Post-Weisberg-II remand, 438 F.Supp. at 502-03. Thereafter, the government was willing to treat the “Stombaugh report” as falling within Weisberg’s request, though it insisted it could not locate any such document. We therefore had no reason to do otherwise in Weisberg III.
The Stombaugh report is not Weisberg’s only attempt to avert a grant of summary judgment by pointing to documents he has not received and falsely implying that they were within the scope of his original request. For instance, after the District Court awarded judgment following our remand in Weisberg II, Weisberg moved for reconsideration and claimed — for the first time, as far as we can tell — that he not only wanted copies of certain tests but all copies of those tests. R. 50 (Weisberg affidavit ¶ 18). This enabled him to claim that the FBI must search not only its Washington headquar*264ters but its Dallas field office, to unearth extra copies that might be kept there.
. Although it is possible that Weisberg discovered the existence of a few of these tests only at this date, through releases from one or another of his FOIA cases, he knew of at least some as early as 1970, for Weisberg himself sent the FBI bullets for testing at that time. See App. 285-86 (Weisberg affidavit ¶¶ 314-18). Another bullet supposedly received wide publicity at the time of the House investigation into the Kennedy assassination, App. 284 (Weisberg affidavit ¶ 312); the Select Committee on Assassinations issued its final report in early 1979, see H.R.Rep. No. 1828, 95th Cong., 2d Sess. (1979). And Weisberg says it was one of his books that called the Aldredge sidewalk scar, see infra note 15, to the FBI lab’s attention. App. 280 (Weisberg affidavit ¶ 293).
. First, in the order Weisberg describes the specimens, was a bullet one Rex M. Oliver is said to have found while working on a road project near the scene of the crime; it was sent to the lab in early 1969 (Q629/C329). App. 278-79, 434-35. Second was a bullet or cartridge that William A. Barbee is said to have found embedded in the roof of a Dallas building; it was sent for testing in late 1967 (Q614/C327). App. 282-84, 465. Third was a bullet said to have been dug up by one Richard Lester at some railroad tracks. App. 284. Fourth was an unfired bullet said to have been found by Melvin Gray and William Koye near a grassy knoll at Dealey Plaza (Q628/C328). App. 284. Last were bullets Weisberg sent to the FBI for testing in late 1970: one supposedly found in a planter near the Kennedy memorial (thought by Weisberg to be a “grim hoax,” App. 276) and at least one other bullet purposely doctored to look like another Kennedy assassination bullet. App. 285-86, 468, 470-73.
. According to Weisberg, Eugene P. Aldredge noticed a sidewalk scar at the time of the assassination and brought it to the FBI’s attention after the Warren Commission issued its report. App. 279-81. Aldredge said that upon reexamination of the scar, he found it had been patched with some sort of plastic material. App. 451. (According to the FBI, Aldredge charged that “the Soviets have infiltrated the FBI and that [tjhe hole was patched up by the FBI in order ‘to protect the Soviets’ and hide the scar,” id.) Ifie FBI took scrapings of the material and sent them for unspecified “examination” in November 1964. Id.
. It is not clear whether Weisberg expects to receive results from any examinations of these specimens by means other than spectrographic or neutron activation analysis. At one point he says, “even if these examinations [of the Bar-bee bullet] did not include compositional analysis, I should have received them because I was to have been given everything.” App. 283-84 (Weisberg affidavit ¶ 311). In discussing the Aldredge sidewalk scar one page earlier, however, he says, “[m]y request is for compositional analyses.” App. 282 (Weisberg affidavit ¶ 305).
. The Commission transmitted its final report with a letter dated September 24, 1964. Report of the President’s Commission on the Assassination of President Kennedy vii (1964). November 1964 was the earliest any specimen about which Weisberg provides relevant information was sent to the lab for testing. See App. 451 *265(Aldredge sidewalk scar); see generally supra notes 14-15.
. Weisberg’s lateness in showing any interest in these specimens is the main reason to doubt the FBI looked for any tests it might have performed on them. Moreover, none of the recently mentioned specimens appeared in the FBI’s 1976 listing of Kennedy assassination-evidence. See App. 185-86 (Justice Department’s answer to interrogatory no. 5). (Presumably the tests the FBI performed concluded that none of the specimens was related to the shooting, and indeed Weisberg has never objected to the exclusion of those specimens from the list, though he has known for many years that at least some of them were sent for testing.) That the FBI does not consider these items to be Kennedy assassination evidence — let alone that they were not tested “for the Warren Commission” — is further reason to suspect that it never occurred to the FBI that Weisberg wanted results from any tests it might have performed on them.
. This included Agents Cunningham, Heilman, Heiberger, and Gallagher. App. 54.
. Weisberg himself has adduced evidence that at least some aspects of standard practices concerning the routing of reports through Dallas were not followed with the very records he seeks. The FBI lab would often supply the Warren Commission with its results indirectly by sending a report to the Dallas office which would in turn incorporate that report into a report to the Commission. See Post-WeisbergII remand, 438 F.Supp. at 496-97. Weisberg, however, does recall “two times that the reporting of Lab work to the Commission was not via Dallas. One was the results of the NAA’s, the other the results of the curbstone testing.” App. 236 (Weisberg affidavit ¶ 133).
. Weisberg has made general assertions that the Dallas office has some documents not found in FBI headquarters but has not refuted the FBI’s evidence that headquarters kept copies of outgoing lab reports.
. The only document Weisberg now claims he has not been given that would fall within the category of reports that might have been sent to Dallas is the “Stombaugh report.” It would be one thing, had Kilty’s search turned up notes or some other evidence that a final report of this type might exist, to ask the FBI to search the Dallas office in the faint hope that the report was sent there, but that the lab copy is missing. It is a faint hope based on a speculation, however, to ask the FBI to do so when Kilty’s now-documented thorough search through Stombaugh’s materials revealed nothing even related to such an examination. See generally supra note 11.
. Weisberg’s final attempt to force a search of the Dallas field office is his request for all copies the FBI has of any reports he seeks. This, however, plainly exceeds the scope of his original request, see supra note 12, though it would appear to fall within another Weisberg *267request, now being litigated, for all records in the Dallas and New Orleans field offices pertaining to the Kennedy assassination, Weisberg v. FBI, C.A. No. 78-420 (D.D.C. filed Mar. 10, 1978). See Brief for Defendants-Appellees at 27 n. 5.
. Aside from his criticism of the timing of the government’s release of neutron activation printouts, see supra p. 1352, Weisberg’s only other challenge to the search that bears mentioning is his complaint that Kilty failed to supply him with copies of lab documents that Weisberg admits were misfiled and that the FBI has previously released to him. Apparently, the lab set of some curbstone spectrographic notes is missing some pages which were misfiled in the main Oswald file behind an unrelated document pertaining to a German article about Oswald. See App. 265-68 (Weisberg affidavit ¶¶ 240-51). Weisberg claims that the pages the lab did have bore the number of the file containing the misfiled, missing pages, though it is not clear how Weisberg knows this, since all parties agree that Kilty deleted the file numbers from the copies he provided. It is also not clear why the lab copy would bear the number of a file in which materials were erroneously kept. Even if it did, however, Weisberg has not explained why Kitty should have tracked down this particular Oswald file if he had no reason to think that the lab copy he did locate was not itself complete. More important, because the FBI has already given Weisberg the contents of the Oswald file (in a release unrelated to this lawsuit), Kilty was under no additional obligation to search those files.
. This corrected a statement he made in his earlier affidavit: that “[n]eutron activation analysis and emission spectroscopy were used to determine the elemental composition of the borders and edges of holes in clothing and metallic smears present on a windshield [Q15] and a curbstone.” App. 159 (affidavit of May 13, 1975). Weisberg now says that this was the correct version and the later “correction” a lie. We explain in text why there is nothing evasive in his corrected response. And we can derive no significance from the fact that Kilty was initially mistaken, especially when that mistake overstated the records the FBI would be bound to provide. It is apparent that Kilty initially thought Q15 had been analyzed because it had been sent to the nuclear reactor for that purpose. When he later found out that no analyses existed because the test was aborted, he corrected his earlier error.
. Nor does he venture a guess why lead would play any role in a person’s effort to patch a bullet hole to conceal its existence.
. Weisberg also presented documents to us in Weisberg III that he had discovered too late to present to the District Court. See Weisberg III, 627 F.2d at 371 n. 49. We did not decide how to treat that evidence because we found sufficient cause for remand in the record compiled before the District Court.
. We gave the government leave to file a short • supplemental brief addressing the significance of the document. It did so while continuing to maintain that we should not consider the new evidence on appeal.
. Portions of these proceedings are. now on appeal, and our citation of this opinion is in no way an endorsement of it. Even if the portion of the opinion we have quoted should be affected on appeal, our main point remains: that the Shea memorandum is directly relevant to the scope of searches for certain King documents, not to the searches now on review.