Gloria Kronisch, executrix of the estate of Stanley Milton Glickman (“Glickman” or “plaintiff’), appeals from a judgment of the United States District Court for the Southern District of New York (Kimba M. Wood, Judge) granting summary judgment in favor of defendants. Glickman brought suit against the United States of America and two officials of the Central Intelligence Agency (the “CIA”), Sidney Gottlieb and Richard Helms, alleging that he was one of the victims of the CIA’s program to test the effects of mind-altering drugs, including lysergic acid diethylamide (“LSD”), on unwitting subjects beginning in the early 1950s. Glickman claims that Gottlieb or some other agent of the United States government placed LSD in his drink in a Paris café in October 1952. The district court, adopting in full the conclusions of the Report and Recommendation of Magistrate Judge Naomi Reice Buchwald, granted defendants’ motion for summary judgment on the bases that plaintiff had failed to establish a genuine issue of material fact as to liability, that his claims were time-barred, and that the court lacked personal jurisdiction over Gottlieb and Helms. We affirm in part, vacate in part, and remand.
I.
A. The CIA’s Drug-Testing Programs
The Select Committee to Study Governmental Operations with Respect to Intelligence Activities, chaired by the late Senator Frank Church of Idaho (the “Church Committee”), held hearings in 1975 to investigate various CIA activities, including the testing and use of chemical and biological agents by the intelligence community. The final report of the Church Committee, published in 1976, explained that the CIA was acutely concerned in the late 1940s and early 1950s that the Soviet Union, China, and other Communist countries had used, or were developing the capacity to use, chemical and biological agents and other techniques for purposes of interrogation, brainwashing, and attacks against United States and Allied personnel abroad. See Final Report of the Select Committee to Study Governmental Operations with Respect to Intelligence Activities, 94th Cong., 2d Sess. (Apr. 26, 1976), at 392 (“Church Committee Report”). “Of particu*117lar concern was the drug LSD.”1 Id. As described by the Church Committee Report,
In order to meet the perceived threat to national security, substantial programs for the testing and use of chemical and biological agents — including projects involving the surreptitious administration of LSD to unwitting nonvolunteer subjects at all social levels, high and low, native American and foreign — were conceived, and implemented. These programs resulted in substantial violations of the rights of individuals within the United States.
Id. at 393 (internal quotation marks omitted). In the late 1970s, at the request of the Subcommittee on Health and Scientific Research of the Senate Committee on Human Resources, chaired by Senator Edward M. Kennedy of Massachusetts (the “Kennedy Committee”), the CIA attempted to identify victims of the drug tests. The CIA ultimately identified sixteen unwitting subjects of LSD tests in the United States, but did not identify any victims of overseas testing.
The earliest CIA program involving the use of chemical and biological agents was approved by the Director of Central Intelligence (“DCI”) in 1950. Id. at 387. Known initially as Project BLUEBIRD and renamed ARTICHOKE in August 1951, the objectives of this program were the development of defensive techniques for resisting interrogation as well as “the evaluation of offensive uses of unconventional interrogation techniques, including hypnosis and drugs.” Id. The project included both controlled in-house experiments regarding interrogation techniques, as well as overseas interrogations on foreign nationals using nonhallucinogenic drugs (so-called “truth serums”) that occurred in “safehouses” or other secure locations. See id. at 387-88. The overseas interrogations were conducted upon known or suspected foreign intelligence agents, double agents, and defectors, but, according to the government, also may have included persons not suspected of espionage or wrongdoing. See Declaration of Nancy G. Milburn, Assistant United States Attorney for the Southern District of New York, in Support of Defendants’ Motion for Summary Judgment, Exhibits 4-8 (April 25, 1996) (“Milburn Declaration”). During two overseas trips in 1952, interrogations were conducted by ARTICHOKE teams using “new drugs,” identified as seconal, dexedrine, and marijuana. Id. Ex. 10. The documents contained in the record before us do not indicate the use of hallucinogenic drugs such as LSD as part of the BLUEBIRD/ARTICHOKE Program, nor do the BLUEBIRD/ARTICHOKE documents indicate that interrogations took place in France — where Glickman alleges he was drugged in October 1952 — or that they were performed on Americans overseas. See Kronisch v. United States, No. 83 Civ. 2458, 1997 WL 907994, at *5-*6 (S.D.N.Y. April 14, 1997) (“Kronisch III”).
Based on a proposal to the DCI from defendant Richard Helms — then the CIA’s Assistant Deputy Director for Plans — the CIA’s technical research component, known as the Technical Services Division (“TSD”), initiated Project MKULTRA, the agency’s principal project to study the effects of chemical and biological agents on human behavior. See Church Committee Report at 389-90. Although not formally approved by the DCI until April 13, 1953, id. at 390, Project MKULTRA was “actively underway since the middle of 1952.” Milburn Declaration, Ex. 23. MKULTRA included 149 subpro-jects that the CIA contracted out, by means of a special funding mechanism intended to ensure the highest level of secrecy, to specialists in universities, hospitals, pharmaceutical houses, and public and private institu*118tions. See Church Committee Report at 390-91; Kronisch III, 1997 WL 907994, at *6; Milburn Declaration, Ex. 28. Defendant Sidney Gottlieb, who at the time was Chief of the Chemical Division of TSD, was put in charge of planning and organizing MKUL-TRA in 1952.
As described in the Church Committee Report, LSD was one of the materials tested in the MKULTRA program, and that testing included the “surreptitious administration [of LSD] to unwitting nonvolunteer subjects in normal life settings by undercover officers of the Bureau of Narcotics acting for the CIA.” Church Committee Report at 391. In the view of the CIA, such surreptitious testing was important because the “testing of materials under accepted scientific procedures fails to disclose the full pattern of reactions and attributions that may occur in operational situations.” Id. (internal quotation marks and citation, omitted). The record shows that Bureau of Narcotics Agent George White, who was asked by Gottlieb in 1952 to work on LSD research for the CIA, see Declaration of Sidney Gottlieb ¶4 (April 23, 1996) (“Gottlieb Declaration”), conducted LSD experiments on unsuspecting persons in New York City. White may have given LSD to unwitting friends in his New York City apartment in November 1952, and beginning in June 1953 he administered LSD to unsuspecting persons (typically drug informants and prostitutes) with whom he came into contact in his work as a narcotics agent. See Kronisch III, 1997 WL 907994, at *6 n. 8. He may also have administered LSD to one or more unsuspecting persons at a bar. See Hearings 'Before the Subcommittee on Health and Scientific Research of the Senate Committee on Human Resources on S. 1893, 95th Cong., 1st Sess. 204 (Sept. 20 and 21, 1977) (“Kennedy Committee Hearings”) (testimony of Sidney Gottlieb).
In addition to testing LSD on unwitting subjects in the United States, “[a] special procedure, designated MKDELTA, was established to govern the use of MKULTRA materials abroad.” Church Committee Report at 391. The Church Committee determined “that the use of these materials abroad began in 1953, and possibly as early as 1950.” Id- (emphasis added). Defendants assert that overseas use of LSD was limited to formal interrogation settings, and that these interrogations were performed exclusively on foreign nationals. They claim that Gottlieb traveled overseas to participate in these interrogations on three occasions between late 1953 — a year or more after plaintiffs alleged incident — and 1956. They assert that the first of these trips was to the Far East, and while national security concerns prevent them -from divulging the locations of the other two interrogations, they have represented that neither occurred in France. See Brief of Defendants-Appellees at 8. Other than these three occasions, defendants assert that “neither Gottlieb nor anyone else in TSD participated in any overseas interrogations or operations in which a person was given LSD.” Id.
However, as the Church Committee Report indicates, “[b]ecause MKULTRA records were destroyed, it is impossible to reconstruct the operational use of MKULTRA materials by the CIA overseas.” Church Committee Report at 391. MKULTRA records were destroyed in January 1973 by TSD personnel acting on the orders of Gottlieb, who in turn had obtained approval from Helms. See id. at 389. During hearings into the CIA’s drug-testing activities held by the Kennedy Committee, Gottlieb testified that he had ordered the files destroyed for three reasons: to hide the identities of researchers who had assisted the CIA on express assurances of confidentiality; to prevent the records (many of which were incomplete) from being misunderstood once he and others who were most familiar with the program retired; and as part of a CIA-wide drive to reduce the amount of paper it stored. See Kennedy Committee Hearings at 195 (testimony of Sidney Gottlieb).
The destruction of MKULTRA documents “made it impossible for the [Church] Committee to determine the full range and extent of the largest CIA research program involving chemical and biological agents,” and “prevented the Committee from determining the full extent of the operations which made use of materials developed in. the MKULTRA program.” Church Committee Report at *119404. However, many financial records related to MKULTRA — some of which referred to substantive aspects of the program — did escape destruction, and were the subject of testimony by Gottlieb and others at the Kennedy Committee hearings. Neither that testimony nor any of the recovered financial records contain any reference to plaintiff, or to overseas testing of LSD on unwitting subjects in France, or to any overseas testing in 1952, when Glickman claims to have been drugged. The record also shows, however, that few if any Americans other than those working with the CIA had access to LSD in the early 1950s.
B. Glickman’s Personal History
On appeal from a grant of summary judgment for defendants, we are required to view the facts in the light most favorable to plaintiff. See Lightfoot v. Union Carbide Corp., 110 F.3d 898, 908 (2d Cir.1997). From that standpoint, Glickman’s personal history includes the following:
In October 1952, Glickman was pursuing a promising career as an artist in Paris. He was approached one evening by an acquaintance who asked Glickman to accompany him to the Café Select to meet some American friends. Upon meeting these men at the Café Select, Glickman and the men engaged in several hours of contentious debate on political issues. As Glickman prepared to leave, one of the men offered Glickman a drink as a conciliatory gesture, and Glickman eventually accepted. Rather than call over the waiter, the man walked to the bar to get the drink, at which point Glickman observed that he had a clubfoot. Halfway through the drink, Glickman “began to experience a lengthening of distance and a distortion of [his] perception,” and he observed that “[t]he faces of the gentlemen flushed with excitement as they watched the execution of the drink.” Affidavit of Stanley Milton Glickman ¶ 13 (August 20, 1983). One of the men then “brought the topic of discussion to the working of miracles,” id., and suggested to Gliek-man that surely he would be capable of this power. Glickman left the café and experienced distortions of color and other hallucinations, believing that he had been poisoned. When Glickman awoke the next morning, he was hallucinating intensely.
For approximately two weeks, Glickman “wandered in the pain of madness, delusion and terror,” and then decided to return to the Café Select, where he “consciously closed [his] eyes to wait for ‘someone’ to come and tell [him] what had happened.” Id. ¶ 16. He was then carried from the café, placed on the floor of an automobile, and taken to the American Hospital of Paris, where he was admitted on November 11, 1952. Glickman remained in the hospital for two days, during which time he was examined and given electroshock treatment. He signed himself out of the hospital against the wishes of his attending physician, but returned a day later and remained as a patient for seven days— during which time he believes he was given additional doses of hallucinatory drugs — until a friend arrived and helped him to sign out and to return to his studio. Over the next ten months, Glickman remained mostly in his studio, experiencing “stress, terror, hallucination and difficulty eating,” which “reduced [his]’ body to a feeble quality.” Id. ¶20. When friends of his brother-in-law’s family saw him on the street and observed his condition, they contacted Glickman’s family, who arranged for him to be brought back to the United States in July 1953.
Glickman was treated by a doctor for several weeks, and his physical condition began to improve, but his mental condition did not. He saw psychiatrists on a few occasions, but refused to continue treatment. Over the next twenty-five years, he held various odd jobs but never painted again and never led a normal social life. Glickman died on December 11,1992.
C. Procedural History
In 1977, Glickman’s sister, Gloria Kroniseh, called him and said that she had read an article indicating that the CIA “had experimented with L.S.D. on unsuspecting persons in foreign countries in the 1950’s.” Id. ¶ 28. Kroniseh sent Glickman the article, and Glickman began to watch the Kennedy Committee hearings on television. It was at this time that Glickman first formed a belief that he had been drugged by the CIA in 1952. *120During the latter months of 1977 through the end of January 1978, Glickman expressed this belief in letters to Senator Kennedy, the CIA, and the Attorney General, and at some point he unsuccessfully sought legal representation. After a friend, Dean Corren, traveled to' Washington in April 1981 and obtained additional information about the CIA’s drug-testing programs by examining the CIA files that had earlier been made public pursuant to the Freedom of Information Act (“FOIA”), 5 U.S.C. § 552, Glickman filed an administrative tort claim with the CIA on December 22, 1981. The CIA denied the claim by letter dated December 1, 1982. Glickman filed this suit on March 30, 1983.
Glickman sued defendant United States under the Federal Tort Claims Act, 28 U.S.C. § 2671 et seq. (“FTCA”), raising common-law tort claims of negligence, invasion of privacy, misrepresentation, and intentional infliction of emotional distress. He also sued defendants Gottlieb and Helms in their individual capacities and in their official capacities as former employees of the CIA,2 under Bivens v. Six Unknown Named Agents of Federal Bureau of Narcotics, 403 U.S. 388, 91 S.Ct. 1999, 29 L.Ed.2d 619 (1971), raising claims under the First, Fourth; Fifth, and Eighth Amendments to the United States Constitution. The ease languished as the result of a series of delays, see Kronisch v. United States, No. 83 Civ. 2458, 1994 WL 524992, at *2 (S.D.N.Y. Sept. 27, 1994) (“Kronisch I”) (describing procedural history), and after discovery was completed,3 defendants moved for summary judgment. The Magistrate’s Judge’s Report and Recommendation recommended that the motion be granted on the bases that plaintiffs claims were untimely, that plaintiff had failed to establish a genuine issue of material fact on the merits, and that the court lacked personal jurisdiction over Gottlieb and Helms. Judge Wood adopted the recommendations of Magistrate Judge Buchwald in their entirety, and granted defendants’ motion. See Kronisch III, 1997 WL 907994.
II.
A. Statute of Limitations
We review de novo the district court’s determination that plaintiffs claims are barred by the statute of limitations. See Eagleston v. Guido, 41 F.3d 865, 870 (2d Cir.1994).
1. FTCA Claims
The FTCA provides that “[a] tort claim against the United States shall be for-*121ever barred unless it is presented in writing to the appropriate Federal agency within two years after such claim accrues.” 28 U.S.C. § 2401(b). Inasmuch as plaintiff filed an administrative claim with the CIA on December 22, 1981, his FTCA claim against the government is untimely if it accrued prior to December 22, 1979.
Ordinarily, a plaintiff’s FTCA claim accrues at the time of injury. See Barrett v. United States, 689 F.2d 324, 327 (2d Cir.1982). However, in eases such as this one, where the government conceals the acts giving rise to plaintiffs claim, or where plaintiff would reasonably have had difficulty discerning the fact or cause of injury at the time it was inflicted, the so-called “diligence-discovery rule of accrual” applies.4 Under this rule, “accrual may be postponed until the plaintiff has or with reasonable diligence should have discovered the critical facts of both his injury and its cause.” Id. Discovery of the “critical facts” of injury and causation is not an exacting requirement, but requires only
knowledge of, or knowledge that could lead to, the basic facts of the injury, i.e., knowledge of the injury’s existence and knowledge of its cause or of the person or entity that inflicted it.... [A] plaintiff need not know each and every relevant fact of his injury or even that the injury implicates a cognizable legal claim. Rather, á claim will accrue when the plaintiff knows, or should know, enough of the critical facts of injury and causation to protect himself-by seeking legal advice.
Guccione v. United States, 670 F.Supp. 527, 536 (S.D.N.Y.1987) (citations omitted) (Motley, J.), aff'd on other grounds, 847 F.2d 1031 (2d Cir.1988), cert. denied, 493 U.S. 1020, 110 S.Ct. 719, 107 L.Ed.2d 739 (1990). A claim does not accrue when a person has a mere hunch, hint, suspicion, or rumor of a claim, see Hobson v. Wilson, 737 F.2d 1, 35 (D.C.Cir.1984), cert. denied, 470 U.S. 1084, 105 S.Ct. 1843, 85 L.Ed.2d 142 (1985), but such suspicions do give rise to a duty to inquire into the possible existence of a claim in the exercise of due diligence, see id. at 35 & n. 107.
We agree with the district court that Glick-man was aware of the basic facts of his FTCA claim before December 22, 1979. Glickman first came to believe that he had been drugged by the CIA in 1977, when he was told by his sister about the Kennedy Committee hearings and began to watch the hearings on television. As Glickman testified at his deposition, after watching the Kennedy Committee hearings in 1977 and learning of the CIA’s drug tests on unsuspecting persons in the 1950s, he reached “the logical conclusion after remembering the events of my life in 1952 that I was one of those victims.” On October 3, 1977, Glickman made a FOIA request to the CIA for copies of any documents filed under his name or containing his name, believing that “inasmuch as I was one of [their] victims, they would have my record in my personal file.” Deposition of Stanley Milton Glickman at 283 (October 23, 1987) (“Glickman Deposition”). Upon receiving a response from the CIA advising him that the agency’s only record was of one Stanley Glickman from Illinois who had received employment information from the CIA in 1976, Glickman responded, by letter dated January 30, 1978, that he was not that individual, and added that “as an aid to your search I will specify that Paris[,] France in the 1950’s is the area for you to focus upon.” Glickman also wrote to Senator Kennedy on October 11, 1977, urging him to identify the individual victims of the CIA’s tests. Senator Kennedy responded, by letter dated November 1, 1977, that neither the CIA nor the Kennedy Committee had yet been able to identify victims, “principally due to the fact that the records concerning these tests were destroyed in 1973.” Glickman again wrote to Senator Kennedy on January 28, 1978, asking him for help in obtaining his records from the CIA, and stating that “[m]y continued interest in the subject of CIA drug experimentation on unsuspecting persons in the 1950’s is due to my belief that I AM ONE OF THOSE PERSONS.” Finally, Glickman sought help from Attorney General Griffin Bell, informing him in a letter dated Novem*122ber 9, 1977 that Glickman believed himself to be a victim of a CIA drug experiment, and asking him to expedite Glickman’s receipt of CIA files.
Accordingly, it is clear that Glickman had formed a firm belief prior to December 22, 1979 — and certainly no later than January 28, 1978, the date of his second letter to Senator Kennedy — that he had been an unwitting victim of a CIA drug experiment. Moreover, based on the information he possessed prior to that date, Glickman believed strongly enough in his claim to (unsuccessfully) seek legal representation. See Guccione, 670 F.Supp. at 536 (“[A] claim will accrue when the plaintiff knows, or should know, enough ... to protect himself by seeking legal advice.”) (citing United States v. Kubrick, 444 U.S. 111, 122-25, 100 S.Ct. 352, 62 L.Ed.2d 259 (1979)).5 Glickman’s firmly expressed belief in his claim and his effort to seek legal advice undermine his later contention that he had only a mere “hunch” or “suspicion” of his injury and its cause.
In any event, even if plaintiffs awareness of his injury and its cause prior to December 1979 could only be characterized as a mere “hunch” or “suspicion,” plaintiff would still have been under a duty to diligently investigate his claim. See Hobson, 737 F.2d at 35 & n. 107. However, after his initial inquiries in 1977 and the early part of 1978, plaintiff took no farther steps whatsoever to pursue his claim until April 1981, when his friend Dean Corren went to Washington to examine publicly available CIA files. Accordingly, plaintiffs FTCA claim is untimely both because he was aware of the basic facts of his injury and its cause more than two years before filing his administrative claim, and because he failed to exercise due diligence in pursuing his claim. Although plaintiff contends that his claim did not accrue until Corren’s trip to Washington in 1981, which provided Glickman with more concrete information about the CIA’s drug-testing program, this information was merely a more detailed description of the program of which Glickman became aware as early as 1977, and which had given rise to his firm belief at that time that he was one of the victims of this program. Indeed, information of the sort obtained by Corren was publicly available (among other places, in the Church Committee Report) long before April 1981, and Glickman offers no explanation as to why he did not try to obtain such information sooner.
Plaintiff argues that the responses he received from the CIA and the Justice Department to the letters he sent them were misleading, and constituted a “fresh concealment”6 that should toll the statute of limita*123tions. This argument is without merit.-The CIA merely informed Glickman, in a letter dated February 10,1978, that it had completed the search of its files and had uncovered no record of him. Even assuming arguendo the merits of Glickman’s claim, there is no reason to believe that this response was inaccurate in light of the destruction of MKUL-TRA documents in January 1973. Moreover, at the time he received the CIA’s response, Glickman was already aware, from Senator Kennedy’s letter of November 1, 1977, that the MKULTRA files had been destroyed in 1973. Therefore, the fact that the CIA uncovered no documents containing Glickman’s name should not have come as a surprise to him — much less can it be characterized as an obstructive “concealment” warranting the tolling of the statute of limitations. Indeed, although Glickman took no further actions to investigate his claim until April 1981, he stated at his deposition that, even after being told that the CIA had no records of him, he continued to believe that he was a victim of a CIA experiment. As to the Department of Justice’s response of December 7, 1977, that letter did nothing more than inform Glick-man, who had requested that the Department expedite the CIA’s response .to his FOIA request, that the Department lacked jurisdiction to do so and that the inquiry did not appear to merit a special investigation by the Attorney General. This letter cannot be regarded as a “concealment” that should toll the statute of limitations.
In sum, because plaintiff was aware of the basic facts of his FTCA claim prior to December 22, 1979, and because he failed to exercise reasonable diligence in pursuing his claim between early 1978 and April 1981, his FTCA claim against the United States is untimely under the applicable two-year statute of limitations.
2. Bivens claims against Helms and Gottlieb
Plaintiff alleges Bivens claims against defendants Helms and Gottlieb for violating his constitutional rights by-“devising and executing a program of experimental L.S.D. testing and electro shock treatment, of which the Plaintiff was a subject.” Amended Complaint of Stanley Milton Glickman ¶¶ 44-53. He also alleges that Gottlieb was the actual person who gave him the drink laced with LSD. See id. ¶ 18. We conclude that plaintiffs Bivens claims are untimely except insofar as they allege that Gottlieb was actually the person who gave plaintiff the LSD-laced drink.
Federal courts in New York apply a’ three-year statute of limitations period to Bivens claims. See Chin v. Bowen, 833 F.2d 21, 23-24 (2d Cir.1987). While state law supplies the statute of limitations period, “federal law determines when a federal claim accrues.” Eagleston, 41 F.3d at 871. Under federal law, the limitations period governing plaintiffs Bivens claims, like the period governing his FTCA claims, will be equitably tolled so long as defendants’ concealment of their wrongdoing prevented plaintiff from becoming aware of, or discovering through the exercise of reasonable diligence, his cause of action. See Barrett, 689 F.2d at 327 (“[R]ead into every federal statute of limitations is the equitable doctrine that in case of defendant’s fraud or deliberate concealment of material facts relating to his wrongdoing, time does not begin to run until plaintiff discovers, or by reasonable diligence could have discovered, the basis of the lawsuit.”) (internal punctuation and citation omitted). Accordingly, inasmuch as plaintiff filed his complaint against Gottlieb and Helms on September 13, 1983, his claims against them are untimely if he was aware of these claims, or should have been aware of them through the exercise of reasonable diligence, before September 13, 1980.7
*124Insofar as plaintiff alleges that Helms and Gottlieb are liable for devising and executing a program of covert-CIA drug-testing of which plaintiff was a subject, we agree with the district court that these claims accrued at or about the same time as plaintiffs FTCA claims against the government — in no event later than January 28, 1978, well before September 1980. It is undisputed that, after learning of the Kennedy Committee hearings and watching some of the testimony on television, Glickman was aware in 1977 that Gottlieb headed the CIA’s drug experimentation program in the 1950s. See Original Affidavit of Stanley Milton Glickman ¶ 27 (appended to Affidavit of Stanley Milton Glickman (August 20, 1983)). Moreover, if in the exercise of reasonable diligence plaintiff had made any inquiries into which other CIA officials had authority over the agency’s drug testing program, it is clear that plaintiff would have been able to identify Helms. Among other things, Helms’s role was well documented in the Church Committee Report. Indeed, plaintiff does not argue that he could not have discovered Helms’s role within the limitations period. See Kronisch III, 1997 WL 907994, at *16. Rather, plaintiff argues that the limitations period for his Bivens claims should be tolled for the- same reasons as he advances with respect to his FTCA claims. We reject these arguments for the reasons stated above.
Our analysis differs, however, with respect to plaintiffs claim that Gottlieb himself was the person who laced Glickman’s drink with LSD. Although plaintiff knew in 1977 that Gottlieb headed the CIA’s drug testing program, he only first came to believe that Gottlieb himself was the man who gave him the drink in the Café Select when Cor-ren asked him, upon returning from Washington in April 1981, whether any of the men in the café had a clubfoot. Plaintiff alleges that he then recalled that the man who gave him the drink was indeed clubfooted — a characteristic, Corren informed him, that was shared by Gottlieb. To the extent that plaintiffs Bivens claims are based on the allegation that he was drugged by somebody connected with the CIA — pursuant to a program installed and executed by Helms and Gott-lieb — plaintiff was clearly on notice of these claims. However, it is less clear — and, we believe, a question for the jury — whether plaintiff should have been able, through the exercise of due diligence, to plausibly connect Gottlieb to the man in the café prior to being told by Corren that Gottlieb had a clubfoot. See Guccione, 670 F.Supp. at 537 (“The question whether [a plaintiff] knew or should have known the critical facts of his claim, and the subsidiary question of whether he exercised due diligence to discover them are ordinarily matters for the finder of fact,” except where “it is beyond dispute that plaintiff should have known [or] indeed, actually knew the critical facts of his claim.”).
Defendants contend that “[assuming solely for argument that one could reasonably infer from Gottlieb’s clubfoot that he personally participated in plaintiffs alleged injury, plaintiffs reasonably diligent inquiries between 1978 and 1981 could have alerted him to Gottlieb’s foot condition, which is discussed in [John] Marks’s book, [The Search for the ‘Manchurian Candidate’].”8 Brief of Defendants-Appellees at 35. We believe that this is a question of fact for the jury. It is not so clear that the exercise of reasonable diligence prior to April 1981 would have led Glickman to identify Gottlieb as the clubfooted man who allegedly drugged him in 1952 that we may conclude as a matter of law that Glickman failed to fulfill his duty of due diligence. See Orlikow v. United States, 682 F.Supp. 77, 83-85 (D.D.C.1988) (refusing to *125find that plaintiffs’ claims arising from MKULTRA drug tests were untimely based on plaintiffs’ failure to read Marks’s book or particular articles about the .CIA’s drug testing program, and concluding that these issues of diligence were for the jury).
Similarly, the fact that Glickman may have watched Gottlieb testify during the 1977 Kennedy Committee hearings does not require the conclusion that, as a matter of law, Gliekman should have been able to connect him to the man in the Café Select prior to April 1981. First of all, notwithstanding the district court’s observation that “Glickman has stated that he watched on television as Gottlieb testified about the [CIA’s drug-testing] projects before the United States Senate,” Kronisch III, 1997 WL 907994, at *16, in fact the record contains conflicting evidence as to whether Glickman actually watched Gottlieb testify, or whether he simply was aware of Gottlieb’s role within the agency from other witnesses’ testimony, but did not see Gottlieb himself testify. See Gliekman Deposition at 283 (stating, in response to question as to how Glickman had concluded that the CIA was responsible for drugging him, “I believe that newspaper article or the television program had stated while Mr. Gottlieb was on TV at the time, or his testimony was on, I don’t know that I saw him, but I believe — I don’t remember who testified, but there was information given of the programs and their overseas — of then-use of LSD in the 1950s. I don’t know whose testimony I heard.”). The question of whether Gliekman actually saw Gottlieb testify in 1977 is a question of fact for the jury. Even if Glickman did see Gottlieb testify, the question of whether Glickman should have reasonably been able to connect Gottlieb to the clubfooted man who gave him the LSD-laced drink twenty-five years earlier is also a question for the jury.
In sum, Glickman’s Bivens clhims are untimely insofar as he alleges that Helms and Gottlieb were in charge of the CIA program that led to his being unwittingly drugged with LSD.9 But his Bivens claims are not untimely as a matter of law insofar as he alleges that Gottlieb himself was the man who administered the LSD in the Café Select in 1952. Whether Glickman should have been able, through the exercise of reasonable diligence, to make the connection between . Gottlieb and the man in the café within the limitations period is a question that the jury must decide.10
B. The Merits — Genuine Issues of Material Fact
We review a district court’s grant of summary judgment de novo. See Krumme v. *126Westpoint Stevens Inc., 143 F.3d 71, 83-84 (2d Cir.1998). Summary judgment is appropriate only if, construing the evidence in the light most favorable to the non-moving party, no genuine issue of material fact remains to be resolved by a jury. See Anderson v. Liberty Lobby, Inc., 477 U.S. 242, 247-50, 106 S.Ct. 2505, 91 L.Ed.2d 202 (1986). The non-moving party who will bear the burden of proof at trial must produce specific facts indicating that a genuine factual issue exists. See Scotto v. Arcadio Almenas, 143 F.3d 105, 114 (2d Cir.1998). However, “[w]e do not resolve disputed issues of fact but determine whether genuine issues of fact exist to be resolved at trial.” Krumme, 143 F.3d 71, 83-84.
A discussion of whether Gliek-man has produced sufficient evidence to survive summary judgment on his sole timely claim (his Bivens claim that Gottlieb was the person who laced his drink with LSD) must begin by focusing on the CIA’s destruction of its MKULTRA files in 1973, at the direction of Gottlieb and with the approval of Helms. Although we believe, like the district court, that a jury might be skeptical of plaintiffs claim that he was drugged by Gottlieb, we also believe, contrary to the district court, that a jury should be permitted (but not required) to draw an adverse inference against Gottlieb based on the destruction of MKULTRA documents. Further, we believe that, when combined with the possibility that a jury would choose to draw such an adverse inference, plaintiffs circumstantial evidence that he may have been one of the victims of the CIA’s drug tests was enough — barely enough, but enough nonetheless — to entitle him to proceed to trial.
It is a well-established and longstanding principle of law that a party’s intentional destruction of evidence relevant to proof of an issue at trial can support an inference that the evidence would have been unfavorable to the party responsible for its destruction. See, e.g., Nation-Wide Check Corp. v. Forest Hills Distributors, 692 F.2d 214, 217-18 (1st Cir.1982); 2 John Henry Wigmore, Evidence in Trials at Common Law § 291 (James H. Chadbourn rev.1979).11 This adverse inference rule is supported by evidentiary, prophylactic, punitive, and remedial rationales. The evidentiary rationale derives from the common sense notion that a party’s destruction of evidence which it has reason to believe may be used against it in litigation suggests that the evidence was harmful to the party responsible for its destruction. The prophylactic and punitive rationales are based on the equally commonsensical proposition that the drawing of an adverse inference against parties who destroy evidence will deter such destruction, and will properly “plac[e] the risk of an erroneous judgment on the party that wrongfully created the risk.” Nation-Wide Check, 692 F.2d at 218. Finally, courts have recognized a remedial rationale for the adverse inference — namely, that an adverse inference should serve the function, insofar as possible, of restoring the prejudiced party to the same position he would have been in absent the wrongful destruction of evidence by the opposing party. See Skeete v. McKinsey & Co., No. 91 Civ. 8093, 1993 WL 256659, at *5 (S.D.N.Y. July 7, 1993); Turner v. Hudson Transit Lines, Inc., 142 F.R.D. 68, 74 (S.D.N.Y.1991).
In order for an adverse inference to arise from the destruction of evidence, the party having control over the evidence must have had an obligation to preserve it at the time it was destroyed. This obligation to preserve evidence arises when the party has notice that the evidence is relevant to litigation — most commonly when suit has already been filed, providing the party responsible for the destruction with express notice, but also on occasion in other circumstances, as for example when a party should have known that the evidence may be relevant to future litigation. See Turner, 142 F.R.D. at 72-73; *127Skeete, 1993 WL 256659, at *4. Once a court has concluded that a party was under an obligation to preserve the evidence that it destroyed, it must then consider whether the evidence was intentionally destroyed, and the likely contents of that evidence. See Weinreich v. Sandhaus, 850 F.Supp. 1169, 1181 n. 19 (S.D.N.Y.1994).
The district court properly rejected defendants’ argument that an adverse inference was not warranted because at the time the MKULTRA documents were destroyed in 1973 no litigation, administrative action, or congressional investigation had commenced, and because Helms’s and Gottlieb’s reasons for destroying the evidence allegedly had nothing to do with the fear of future litigation. See Kronisch III, 1997 WL 907994, at *21. Although defendants claimed that the documents were destroyed to preserve the confidential identities of outside participants in the MKULTRA program, to prevent incomplete documents from being misunderstood, and to prevent paper overflow, the district court concluded that “[i]t is somewhat hard to believe that both Gottlieb and Helms ... were concerned only with the effect of disclosure on other persons connected to the drug program, and not with the possible consequences to themselves or to the CIA.” Kronisch III, 1997 WL 907994, at *22. Moreover, the district court observed, Gottlieb’s own expressed fear that the documents might be “misunderstood” could be interpreted in a number of ways, including as a fear that the documents would become the subject of litigation. See id. At the very least, the district court could not rule out the possibility that a reasonable jury would find that Helms and Gottlieb feared the prospect of litigation against them individually, and that this prospect may have played a role in their decision to order the destruction of MKULTRA files. Inasmuch as the veracity of Gottlieb’s stated reasons for destroying the MKULTRA documents “is an issue of credibility best left for trial,” id., the district court presumed for purposes of considering the motion for summary judgment that defendants had an obligation to preserve the files and that the destruction was intentional, see id. Defendants do not challenge this sound approach on appeal.’
Concluding (or, for purposes of summary judgment, assuming) that a party has intentionally destroyed evidence that it had an obligation to preserve is not the end of the story, however. We must also attempt to determine whether there is any likelihood that the destroyed evidence would have been of the nature alleged by the party affected by its destruction. This inquiry is part of our attempt to place the innocent party in the same position he would have been in had the evidence not been destroyed by the offending party. The task is unavoidably imperfect, inasmuch as, in the absence of the destroyed evidence, we can only venture guesses with varying degrees of confidence as to what that missing evidence may have revealed. Nonetheless, before we permit the drawing of an adverse inference, we require some showing indicating that the destroyed evidence would have been relevant to the contested issue. See Stanojev v. Ebasco Services, Inc., 643 F.2d 914, 923-24 (2d Cir.1981) (refusing to draw inference, based on nonproduetion of personnel records, that the records would have substantiated plaintiff’s age discrimination claim, where the nonproduction bore “no logical relationship to a finding of age discrimination” because “the documents [were] from a time period prior to [plaintiffj’s assumption of the position from which he was discharged”); Skeete, 1993 WL 256659, at *7 (refusing to draw adverse inference- where party had “not demonstrated prejudice from the denial of access to the destroyed or lost materials” because it had “fail[ed] to provide any extrinsic evidence that the subject matter of the lost or destroyed materials would have been unfavorable to [the opposing party] or would have been relevant to the issues in this lawsuit”). Wigmore states the rule as follows:
The failure or refusal to produce a relevant document, or the destruction of it, is evidence from which alone its contents may be inferred to be unfavorable to the possessor, provided the opponent, when the identity of the document is disputed, first introduces some evidence tending to show that the document actually destroyed or withheld vs the one as to whose contents it is desired to draw an inference.’
*1282 Wigmore, Evidence in Trials at Common Law § 291, at 228 (emphasis in original).
Wigmore’s description of the proper rule was offered in the context of a hypothetical in which a particular document (e.g., a deed meant to show the conveyance of a certain piece of land from X to Y), known by all to contain critical evidence in the case, was destroyed. Where, as here, a party loses the opportunity to identify such a particular document or documents likely to contain critical evidence because the voluminous files that might contain the document(s) have all been destroyed, the situation becomes more complex — but there can be no doubt that the same basic principle proposed by Wigmore still applies. That is, the prejudiced party may be permitted an inference in his favor so long as he has produced some evidence suggesting that a document or documents relevant to substantiating his claim would have been included among the destroyed files.
Just how much evidence is enough to support an inference about the content of destroyed evidence cannot be precisely defined, and will necessarily vary from case to case, but we remain mindful of Wigmore’s admonition that “care should be taken” not to require too specific a level of proof. Id. at 228. Indeed, holding the prejudiced party to too strict a standard of proof regarding the likely contents of the destroyed evidence would subvert the prophylactic and punitive purposes of the adverse inference, and would allow parties who have intentionally destroyed evidence to profit from that destruction. Certainly, the level of proof that will suffice to support an inference in favor of the innocent party on a particular issue must be less than the amount that would suffice to survive summary judgment on that issue. Otherwise, innocent parties meant to benefit from the adverse inference against offending parties would receive no benefit at all, having been deprived of evidence that may have been crucial to making their case, and yet being held to precisely the same standard of proof before they may present their case to a jury.
We do not suggest that the destruction of evidence, standing alone, is enough to allow a party who has produced no evidence — or utterly inadequate evidence — in support of a given claim to survive summary judgment on that claim. See Nation-Wide Check, 692 F.2d at 218-19 (“The issue before the court was not whether the destruction was sufficient, standing alone, to warrant an adverse inference about the documents’ contents; it was simply whether the destruction was at all relevant to the tracing issue, and if so, whether it was sufficiently probative in conjunction with the other evidence to support the tracing conclusion.”). But at the margin, where the innocent party has produced some (not insubstantial) evidence in support of his claim, the intentional destruction of relevant evidence by the opposing party may push a claim that might not otherwise survive summary judgment over the line. In the absence of such a result, as noted above, the purposes of the adverse inference are eviscerated. Although we emphasize again that there are reasons to be skeptical of plaintiffs claim, we disagree with the district court’s conclusion that plaintiffs evidence is so utterly insubstantial as to render an adverse inference unwarranted as a matter of law. We believe that plaintiff has produced enough circumstantial evidence to support the inference that the destroyed MKULTRA files may have contained documents supporting (or potentially proving) his claim, and that the possibility that a jury would choose to draw such an inference, combined with plaintiffs circumstantial evidence, is enough to entitle plaintiff to a jury trial.
We reiterate that, at the summary judgment stage, we are required to draw all factual inferences in. Glickman’s favor. Moreover, for purposes of this summary judgment motion (only), defendants do not contest that Glickman was unwittingly drugged with LSD in a Paris cafe in 1952. See Kronisch III, 1997 WL 907994, at *20. Although at trial Gottlieb might be able to offer explanations for the events in Glick-man’s life beginning in October 1952, and might be able to discredit the conclusion of plaintiffs expert that these events are explained by Gliekman’s unwitting consumption of LSD, defendants have not yet attempted to do so. Defendants likewise do not contest *129at this stage that Glickman was given the LSD-laeed drink by an American with a clubfoot — a distinguishing characteristic possessed by Gottlieb, the head of the MKUL-TRA1 program. See id. The allegation that the man who gave Glickman the drink was clubfooted is also one that might be attacked by Gottlieb at trial. Glickman claims to have only remembered this distinguishing characteristic some twenty-nine years after the fact when his friend Dean Corren, upon Corren’s return from Washington after viewing CIA files in 1981, asked him whether the man who gave him the drink in 1952 was clubfooted. Glickman’s recollection in response to his friend’s suggestive question will doubtless be challenged at trial. However, whatever doubts Gottlieb might reasonably be able to create in the minds of a jury as to the credibility of this recollection are not doubts that we may now entertain, inasmuch as we are reviewing the grant of a motion for summary judgment.
Accordingly, we must accept as true for purposes of reviewing the district court’s decision on this motion that Glickman was given LSD by an American with a clubfoot in 1952. Plaintiff has produced evidence that the pool of Americans who would have had access to LSD at the relevant time — not to mention the subset of all • such individuals also having a clubfoot — was quite limited, inasmuch as “LSD was known only to a few researchers and government agents” in 1952. Report of Dr. Lester Grinspoon, Associate Professor of Psychiatry, Harvard Medical School, at 8 (May 3, 1988). Moreover, as the district court recognized, “it appears that the CIA, and specifically Gottlieb, was engaged in some form of LSD research in 1952.” Kronisch III, 1997 WL 907994, at *20. In conducting LSD tests on unwitting victims within the United States, the CIA at times would make contact with the subject in a bar or other public place, and administer the drug by slipping it into the person’s drink — a scenario like .the one allegedly experienced See Kronisch III, 1997 WL 907994, at *6, *20. Apart from the domestic testing of LSD on unwitting subjects, MKULTRA materials were used abroad on a number of occasions, “possibly as early as 1950.” Church Committee Report at 391. Although there is no direct evidence that any overseas tests were conducted in Paris or on American subjects in non-interrogational settings, and although Gottlieb testified at his deposition that his only overseas use of LSD involved three interrogations of foreign nationals after 1952 in places other than France, the destruction of MKULTRA files has made it “impossible to reconstruct the operational use of MKULTRA materials by the CIA overseas,” id., and thereby, perhaps to contradict Gottlieb’s assertions. by Glickman.
In sum, while plaintiffs chain of circumstantial evidence may prove to be altogether vulnerable at trial, it is sufficient to suggest the reasonable possibility that, had Gottlieb and Helms not ordered the destruction of the MKULTRA files, these files may have contained evidence helping to substantiate plaintiffs claim that Gottlieb drugged him in Paris in 1952. Accepting, as the current record requires on defendants’ motion for summary judgment, that very few Americans — other than those working with the CIA — knew of or had access to LSD in 1952; that Gottlieb was engaged in some form of LSD research in 1952; that the CIA performed LSD tests on unwitting subjects in the United States at or near the relevant time; that the CIA drug tests were at times performed using the same means alleged here; that the CIA tested LSD abroad, albeit under different conditions and circumstances; and that the man who gave Glickman the drink, like Gottlieb, had the distinguishing trait of a clubfoot, we cannot say that it is so unlikely that the destroyed MKULTRA files would have contained evidence buttressing plaintiffs claim that an adverse inference is unwarranted as a matter of law.12 Although plaintiff has not *130produced direct proof that the CIA conducted LSD tests in Paris, or that the overseas use of LSD involved tests on Americans (rather than foreign nationals) in non-interro-gational settings, the most obvious source of such proof, if it were to exist at all, has been destroyed at Gottlieb’s direction. Under these circumstances, requiring more direct proof than plaintiff has provided before permitting an adverse inference to be drawn against Gottlieb would be at odds with the purposes of the adverse inference rule.
Assuming that a jury were to find that Gottlieb had an obligation to preserve the MKULTRA documents that he ordered to be destroyed, the jury would be entitled to draw an adverse inference against Gottlieb. The possibility that the jury would choose to draw such an inference, along with plaintiffs other circumstantial evidence that he was drugged by the CIA — specifically, by Gottlieb — is enough to entitle plaintiff to a jury trial.
C. Personal Jurisdiction
The district court concluded that it lacked personal jurisdiction over both Helms and Gottlieb. Inasmuch as we have already concluded that the claims, against Helms are untimely, we have no need to consider plaintiffs challenges to the district court’s conclusion with respect to Helms. However, we disagree with the district court’s conclusion that it lacked personal jurisdiction over Gott-lieb.
“Personal jurisdiction of a federal court over a non-resident defendant is governed by the law of the state in which the court sits — subject, of course, to certain constitutional limitations of due process.” Robinson v. Overseas Military Sales Corp., 21 F.3d 502, 510 (2d Cir.1994). Under the relevant portions of New York’s long-arm statute, personal jurisdiction may be asserted over a non-domiciliary who, in person or through an agent, “transacts any business within the state.” N.Y.C.P.L.R. § 302(a)(1) (McKinney’s 1990). Where, as here, the dis-triet court decides a pretrial motion to dismiss for lack of personal jurisdiction on the basis of the written record, without holding an evidentiary hearing, the plaintiff “ ‘need only make a prima facie showing of jurisdiction through its own affidavits and supporting materials’ ” to defeat the motion. Welinsky v. Resort of the World D.N.V., 839 F.2d 928, 930 (2d Cir.1988) (quoting Marine Midland Bank, N.A. v. Miller, 664 F.2d 899, 904 (2d Cir.1981)).
We are persuaded that Glickman has made a prima facie showing of personal jurisdiction over Gottlieb under the “transacting business” provision of N.Y.C.P.L.R. § 302(a)(1). Under this provision, plaintiff must show that Gottlieb “purposefully availed] himself of the privilege of conducting activities within New York” such that bringing Gottlieb before a New York court does not offend due process, and that plaintiffs cause of action “ar[o]se out of’ those activities in New York. CutCo Indus., Inc. v. Naughton, 806 F.2d 361, 365 (2d Cir.1986) (internal punctuation omitted); see Kreutter v. McFadden Oil Corp., 71 N.Y.2d 460, 527 N.Y.S.2d 195, 198-99, 522 N.E.2d 40 (1988); McGowan v. Smith, 52 N.Y.2d 268, 437 N.Y.S.2d 643, 644-45, 419 N.E.2d 321 (1981). In order for a cause of action to “arise out of’ a-party’s activities in New York, there must be “an articulable nexus,” or a “substantial relationship,” between the claim asserted and the actions that occurred in New York. See Kreutter, 527 N.Y.S.2d at 198-99, 522 N.E.2d 40; McGowan, 437 N.Y.S.2d at 645, 419 N.E.2d 321.
Gottlieb admits that he made approximately six trips to New York in 1952, and that he “visited George White on two or three occasions in 1952 to discuss his becoming a consultant for the CIA in LSD research.” Gott-lieb Declaration ¶4. White, a Bureau of Narcotics Agent, conducted LSD tests on unsuspecting persons in New York, including one experiment in which he gave LSD to a group of his friends -in his New York apart*131ment in November 1952. The record also reflects that beginning in June 1953, White covertly administered LSD to people in New York with whom he came into contact in his role as a narcotics agent. Gottlieb testified before the Kennedy Committee that White may have conducted his tests on one or more occasions by administering the drug to an unsuspecting person in a bar.
Apart from his contacts with White, Gott-lieb also declared that he may have visited Dr. Harold Abramson, a physician who later became an MKULTRA consultant, in New York in 1952. See Gottlieb Declaration ¶ 4. According to Gottlieb, “Dr. Abramson may have been engaged in LSD research in 1952 sponsored by TSD, or in which TSD was interested, and if so, I would have had contacts with him about the progress of his research.” Id. White and Abramson, in Gott-lieb’s words, “regularly reported the[] results” of their research to him. Id. Finally, Gottlieb and others within TSD self-administered LSD during 1951-1953, and Gottlieb “believe[s] that one or more of these administrations took place in á New York City hotel room and ... that Dr. Abramson may have been present during one of these administrations,” although he could “not recall specifically if any self-administrations took place in New York in 1952.” Id ¶5.
We believe that these contacts with New York are sufficient to establish a prima facie showing of personal jurisdiction over Gottlieb inasmuch as Gottlieb’s activities in New York were aimed at laying the groundwork for the LSD testing program of which Glickman claims to have been one of the unwitting victims. Apart from Gottlieb’s contacts with Dr. Abramson, and his self-administration of LSD with other TSD officials — both of which bear some relationship to the development of the CIA’s LSD testing program, and hence to Glickman’s allegation that he was a victim of this program — Gottlieb’s New York contacts with George White are particularly significant. White’s project was aimed at the administration of LSD to totally unwitting persons, and may have included, according to Gottlieb’s testimony before the Kennedy Committee, administration to an unwitting person in a New York bar on one or more occasions. In the words of the Church Committee Report, the administration of LSD to “unwitting nonvolunteer subjects in normal life settings by undercover officers of the Bureau of Narcotics” was important to the CIA because the “testing of materials under accepted scientific procedures fails to disclose the full pattern of reactions and attributions that may occur in operational situations.” Church Committee Report at 391 (internal quotation marks and citation omitted). Accordingly, the fact that Gottlieb was working with White in New York in the latter part of 1952 is significant, for purposes of establishing a nexus to Gottlieb’s alleged drugging of Glickman in Paris in October 1952, not only because White was experimenting with LSD, but because White was retained by Gottlieb, in part, for the purpose of performing the very type of “normal life setting” LSD experiment that Glickman claims to have experienced.13
In short, we believe that Gottlieb’s alleged drugging of Glickman in Paris is sufficiently related to Gottlieb’s work in New York to satisfy plaintiffs prima facie showing of long-arm jurisdiction over Gottlieb under N.Y.C.P.L.R. § 302(a)(1).
III.
In sum:
(1) We affirm the dismissal of Glickman’s FTCA claims against the United States as untimely under the applicable two-year statute of limitations;
(2) We affirm the dismissal of Glickman’s Bivens claims against Gottlieb and Helms as untimely under the applicable three-year statute of limitations insofar as these claims allege that Gottlieb and Helms administered a program of LSD testing on unwitting sub*132jects that included Glickman among its victims;
(3) We vacate the dismissal for untimeliness of Glickman’s Bivens claim against Gott-lieb insofar as Glickman alleges that Gottlieb himself was the person who administered the LSD-laced drink;
(4) Assuming that a jury, were to find that Gottlieb had an obligation to preserve the MKULTRA documents that he ordered to be destroyed, the jury would be permitted (but not required) to draw an adverse inference against Gottlieb to the effect that the destroyed documents would have contained evidence supportive of Glickman’s claim. The possibility that a jury would choose to draw such an inference, combined with plaintiffs other circumstantial evidence of liability, is sufficient to allow plaintiff to survive defendants’ motion for summary judgment;
(5) The district court has personal jurisdiction over Gottlieb under N.Y.C.P.L.R. § 302(a)(1);
(6) We have considered all of the parties’ other arguments, and find them to be without merit.
The judgment of the district court is affirmed in part and vacated in part as noted above, and the case is remanded to the district court for further proceedings consistent with this opinion, which proceedings are to be limited to the Bivens claim that Gottlieb himself administered the LSD-laced drink to plaintiff in Paris in October 1952. Each party shall bear its own costs.
. As the Church Committee Report explains:
The CIA had received reports that the Soviet Union was engaged in intensive efforts to produce LSD; and that the Soviet Union had attempted to purchase the world's supply of the chemical. As one CIA officer who was deeply involved in the work with this drug described the climate of the times: "[It] is awfully hard in this day and age to reproduce how frightening all of this was to us at the time, particularly after the drug scene has become as widespread and as knowledgeable in this country as it did. But we were literally terrified, because this was the one material that we had ever been able to locate that really had potential fantastic possibilities if used wrongly.”
Id. at 392-93, 91 S.Ct. 1999 (quoting testimony of CIA officer, 11/21/75) (alteration in original).
. The district court dismissed the complaint against Helms and Gottlieb insofar as it sought money damages from them in their official capacities. See Robinson v. Overseas Military Sales Corp., 21 F.3d 502, 510 (2d Cir.1994) (Bivens claim for money damages can only be brought against féderal official in his individual capacity). Appellant does not challenge this disposition on appeal.
. "In the course of discovery, defendants produced approximately thirty thousand pages of documents, many of which were heavily redacted by the CIA on the basis of the statutory privileges established by the National Security Act of 1947, 50 U.S.C: § 403-3(c)(5), and the Central Intelligence Agency Act of 1949, 50 U.S.C. § 403g, and the state secrets privilege. In addition, a relatively small amount of information was redacted pursuant to the Privacy Act of 1974, 5 U.S.C. § 552a.” Kronisch I, 1994 WL 524992, at *1 (footnote omitted). Plaintiff sought leave to move to compel disclosure of the redacted material. To determine whether to permit plaintiff to make this motion, Magistrate Judge Buchwald requested that plaintiff bhoose fifty documents that he believed would be most advantageous to his case, which the court then reviewed in camera in unredacted form to determine whether they contained any relevant information and whether the privilege assertions were well-founded. The parties submitted additional materials to aid the court in making this determination. See id. The CIA was unable to locate an unre-dacted copy of one of the requested documents, and instead submitted in its place an unredacted copy of a different document requested by plaintiff. Magistrate Judge Buchwald sustained the United States' assertions of privilege, found the unredacted documents to be irrelevant to plaintiff's claim, and denied plaintiff leave to move to compel the production of unredacted material in the remaining documents. See 1994 WL 524992. Judge Wood subsequently affirmed Magistrate Judge Buchwald's discovery order. See Kronisch v. United States, No. 83 Civ. 2458, 1995 WL 303625 (S.D.N.Y. May 18, 1995) ("Kronisch II"). Substantially for the reasons stated by the district court, we believe that the court did not abuse its "wide discretion in managing pre-trial discovery,” B.F. Goodrich v. Betkoski, 99 F.3d 505, 523 (2d Cir.1996), and, having viewed the unredacted versions of the documents ourselves, we affirm the district court’s discovery order. We have considered plaintiff's challenges to the district court’s order, and find them to be without merit.
. Defendants do not contest that the diligence-discovery rule of accrual applies in this case. The only dispute is when plaintiff's claims accrued under this rule.
. The record is not clear as to exactly when Glickman first sought legal advice, although Glickman’s deposition testimony suggests that it was as early as 1977 or 1978. In any event, there is no suggestion in the record that Glick-man's decision to seek legal advice — even assuming arguendo that he first did so after December 22, 1979 — was based on any new information beyond that which he possessed as early as 1977. The relevant point is that before December 22, 1979, Glickman's belief that he had been drugged by the CIA was sufficient for him to have been aware of the need to protect himself by seeking legal counsel. Although Glickman was not successful at the time in obtaining counsel, this fact alone is not dispositive. Inasmuch as Glickman, at the time he sought counsel, knew enough to assert the basic outlines of a cause of action capable of surviving a motion to dismiss for failure to state a claim, the fact that counsel was unwilling to take the case, and to develop further facts through court-ordered discovery, does not prevent the statute of limitations from running. See Kubrick, 444 U.S. at 124, 100 S.Ct. 352 ("If [plaintiff] fails to bring suit because he is ... mistakenly told that he does not have a case, we discern no sound reason for visiting the consequences of such error on the defendant by delaying the accrual of the claim_").
. The original concealment that occurred by virtue of the destruction of MKULTRA documents in 1973 did not continue to toll the statute of limitations once Glickman became aware of the basic facts of his claim in 1977-1978. See, e.g., Pinaud v. County of Suffolk, 52 F.3d 1139, 1157 (2d Cir.1995) ("[W]hen a defendant fraudulently conceals the wrong, the time limit of the statute of limitations does not begin running until the plaintiff discovers, or by the exercise of reasonable diligence should have discovered, the cause of action.") (emphasis added; internal punctuation and citation omitted). To the extent that plaintiff argues otherwise, he seems to be suggesting that without the destroyed documents, he could not have been reasonably certain that he was drugged by the CIA. However, if plaintiff needed to be reasonably certain that the government caused his injury in order for his FTCA claim to accrue, Glickman's claim would not have accrued to this day, and might never accrue. A plaintiff need not have compelling proof of the *123validity of his claim in order for his claim to accrue — indeed, in cases where plaintiffs claim is ultimately misconceived, such proof will not exist — but must simply have formed a firm belief in his claim based on his awareness of the basic facts of injury and causation.
. The district court observed that “[c]onceivably, plaintiff could argue that the limitations period for his Bivens claims should be tolled during the pendency of his administrative tort claim,” Kron-isch III, 1997 WL 907994, at *15, although the court did not cite any authority in support of this proposition. The court went on to find, in any event, that plaintiff's Bivens claims accrued more than three years before the filing of his administrative claim in December 1981, and therefore *124the court had no occasion to decide whether the limitations period should be tolled during the pendency of the administrative claim. Plaintiff does not argue on appeal that the filing of an administrative claim tolls the statute of limitations for a Bivens suit, and we therefore have no occasion to consider the question.
. John Marks was a former Senate aide and State Department official whose book, The Search for the “Manchurian Candidate," based on thousands of pages of CIA documents obtained through the FOIA, gave a detailed description of the CIA’s drug-testing programs. The documents reviewed by Corren in 1981 were apparently the ones originally made public pursuant to Marks’s FOIA request. See Affidavit of Dean R. Corren V 2 (August 19, 1983).
. Apart from Glickman's primary claim that he was drugged at the Café Select by an agent of the United States government, who may have been Gottlieb himself, he also claims that he was administered electroshock treatment and additional hallucinatory drugs at the American Hospital of Paris. The suggestion is that this treatment was also part of the CIA's drug-testing program. In support of this argument, plaintiff introduced evidence that: (1) he had been treated at the hospital for hepatitis earlier in 1952, and a CIA Information Report summarizing information acquired between November 1952 and September 1953 (which spans the time of Glickman's admission to the hospital) states that “[s]ubjects in whom even only a slight modification of hepatic function is present ... make a very marked response to LSD”; and (2) the psychiatrist who treated Glickman after the incident at the Café Select had published an article regarding his experiments with LSD on rabbits that was cited in CIA materials. Whatever else might be said about the merits of plaintiff’s claim regarding his treatment at the American Hospital of Paris, this claim is part and parcel of plaintiff's more general (untimely) claim that he was the subject of a CIA experiment, and not his more specific (timely) claim that he was drugged by Gottlieb himself. In other words, this claim would have accrued when plaintiff first came to believe that he was a victim of CIA drug tests, and is therefore untimely.
. It would not avail the plaintiff to argue that his putative inability to make this factual connection prior to April 1981 also tolls the accrual of his FTCA claim and his Bivens claims against Helms and Gottlieb arising from their roles in the CIA’s LSD testing program. While Glick-man's ability, as of April 1981, to potentially link Gottlieb to the man in the café strengthened all of Glickman's claims, his awareness of the existence of the FTCA claim and the broader Bivens claim — and hence the commencement of the limitations period for those claims — did not depend on his ability to identify Gottlieb as the possible direct perpetrator of the drugging. So long as a plaintiff has enough information to identify the basic facts giving rise to his claim, his ignorance of additional supportive facts does not postpone the vety accrual of his claim.
. The principle that an adverse inference may be drawn against a party responsible for the loss or destruction of evidence is often associated with the famous common-law case of Armory v. Delamirie, 1 Strange 505, 93 Eng. Rep. 664 (K.B. 1722), in which a chimney sweep who found a jewel sued a jeweler for the loss of the jewel, and was entitled, based on the jeweler’s return of the ring without the stone, to an inference that the stone was "of the finest water.” See Welsh v. United States, 844 F.2d 1239, 1246 (6th Cir.1988); Nation-Wide Check, 692 F.2d at 218.
. Plaintiff has also offered evidence that Gott-lieb met with George White — the Bureau of Narcotics Agent who administered LSD to unwitting subjects in New York City — on October 20, 1952 (in New York), and then again in Washington, D.C. on October 30, 1952. Plaintiff argues that there is therefore a window within the relevant time period (Glickman's admission to the American Hospital of Paris on November 11, 1952 occurred approximately two weeks or so after the alleged incident at the Café Select, suggesting, that the incident would have occurred in the vicinity of the last week of October) during which Gottlieb could have been in Paris. See Kronisch *130III, 1997 WL 907994, at *20 n. 19. Although, of course, it is not defendants’ burden to prove a negative, we note that CIA Agent Frank H. Lau-binger testified in his deposition that he recalled either preparing or helping to prepare a memorandum that would have reflected all international travel by Gottlieb in 1952 — and no such document has been forthcoming from defendants in response to plaintiff’s specific document request.
. As described supra, note 12, the alleged incident at the Café Select occurred somewhere in the vicinity of the last week of October 1952. Assuming that a jury were to credit Glickman’s account, the fact that Gottlieb met with White in New York on October 20, 1952, and in Washington on October 30, 1952, would only strengthen the nexus between plaintiff's claim and Gottlieb’s work with White in New York.