The Wizards of Langley

by Jeffrey T Richelson

  The concept behind the Acoustic Kitty project was that unlike a mechanical bugging device, a cat’s ear had a cochlea, as did a human ear, that could filter out irrelevant noise. Project staffers attempted to train a cat to listen to conversations and not to the background noise—as an interim step in designing a microphone that could filter out extraneous noise.72 Then, according to Marchetti,

  they slit the cat open, put batteries in him, wired him up. The tail was used as an antenna. They made a monstrosity. They tested him and tested him. They found he would walk off the job when he got hungry, so they put another wire in to override that. Finally, they’re ready. They took it out to a park bench and said “Listen to those two guys. Don’t listen to anything else—not the birds, no cat or dog—just those two guys!”. . . They put him out of the van, and a taxi comes and runs him over. There they were, sitting in the van with all those dials, and the cat was dead!73

  An undated agency memo that was heavily redacted, possibly largely out of embarrassment, on “[Deleted] Views on Trained Cats [Deleted] for [Deleted] Use,” apparently quashed the project. The memo stated that “the program would not lend itself in a practical sense to our highly specialized needs.” It also concluded that “the environmental and security factors in using this technique in a real foreign situation force us to conclude that . . . it would not be practical.”74 That would seem to have been an obvious conclusion that staffers could have reached before the first cat was cut open.

  In the late 1960s, ORD was also interested in attempting to turn birds—both real and mechanical ones—into spies. One project followed from the refusal, first by the 303 Committee in 1967 and then by President Johnson, to approve an A-12 mission to get better imagery on the Flat Twin engagement radar associated with the Tallinn system.75

  Over a period of five to six months in California, ORD trained a red-tailed hawk to carry a camera over the site and return with its photographs. The process involved construction of a wooden model of a Flat Twin so that the red-tailed spy would know when to activate the camera—which could take only a single photograph—that would be hung around its neck.* But after training was completed, the CIA discovered that legislation restricted the transport of the bird.76 Not surprisingly, no one had thought to write a national security exemption into the law.

  ORD turned to an unprotected crow, which completed its training in half the time. But the crow would have to be escorted much of the way, carried on a boat up the Baltic and through the canals in Denmark and Sweden. Knowing that the Soviets paid close attention to boat movements through the canals, the CIA decided to cancel the project.77

  Sometimes the spies were just supposed to look like birds, in one way or another. The objective of another late 1960s ORD project, appropriately code-named ORNITHOPTER (an airplane that flaps its wings), was to produce small mechanical birds, which could perch on windowsills while their little microphone-equipped bodies eavesdropped on conversations within the room.78

  Project AQUILINE was intended to produce a Remotely Piloted Vehicle (RPV) that would have a radar profile similar to a condor and carry photographic, ELINT, and air-sampling sensors. Its targets were to include the Chinese missile test center and the nuclear test site at Lop Nur and possibly Sary Shagan and the Flat Twin radar that had been the target of Project UPWIND. The vehicle could not be seen or heard at 1,000 feet.But a key problem emerged during development—directing AQUILINE’s movements after it passed over the horizon. Initial experiments employed C-130s, but such aircraft could not communicate with the AQUILINE vehicles if they were more than 250 miles away. U-2s, because of their high altitude, would theoretically permit communication and guidance at far greater distances—except that when flying at 70,000 feet, U-2 pilots would have too many other tasks to permit them to direct the remote spy’s movements. Eventually, the program was terminated—although much of the development work was turned over to Israel, which used it in developing its RPVs.79

  Shortly after the end of AQUILINE, an ORD employee, initially using his own funds, bought a twin-engine model plane set, put a television camera in its belly, and convinced the agency to continue work on the project. Once again a C-130, flying in the same direction, was used to control the aerial spy, which proved capable of producing photographs with resolution of three feet. The CIA turned the technology over to the Army, which used it in its Aquila RPV.80

  DIFFERENCES OF OPINION

  On January 20, 1969, Richard Nixon, having eked out a close victory over Hubert Humphrey in the November elections, became the thirty-seventh president of the United States. On March 21, his Secretary of Defense, Melvin Laird, during prolonged testimony before a skeptical Senate Foreign Relations Committee’s subcommittee on disarmament, announced that the Soviet Union had embarked on a buildup of its strategic nuclear forces aimed at wiping out America’s ICBM force in a single blow. “There is no question about that,” he added.81

  A week earlier, in accord with a recommendation by the Joint Chiefs of Staff, Nixon had announced that he was scrapping the Johnson administration’s plans for the Sentinel ABM defense system. Sentinel’s objective had been to protect urban areas from a Chinese attack or a limited Soviet attack, but the proposal had been the focus of protests—particularly from the fifteen metropolitan areas around which the nuclear-armed ABMs were to be installed. The new administration explained that it was changing the system’s name and objective. The goal of Safeguard, as it was now called, would be to protect not the U.S. population but the country’s land-based missile force—the ICBMs stored in silos in Nebraska, Wyoming, North and South Dakota, Missouri, Kansas, and Arkansas—from a Soviet first strike.82 The strategic theory was simple. By ensuring that U.S. missiles would survive any Soviet attempt to destroy them in their silos, the United States would preserve its retaliatory capability, and any Soviet incentive to engage in a first strike would be removed.

  The feasibility of Safeguard was challenged by both politicians and scientists, including Herbert Scoville, who had left the Arms Control and Disarmament Agency with the advent of the new administration. Opponents suggested that the exposed radars and command and control centers required to operate the system could easily be neutralized, in contrast to ICBMs in hardened silos. The radars could be blinded by the first nuclear detonations. The capability of radars to sort out missiles from debris, decoys, and chaff was questioned, as was the ability of the system’s computers to manage a massive missile-against-missile engagement. In addition, it was argued that the Soviets could build additional ICBMs more cheaply than the United States could build the antimissiles to knock them down.83

  An even more fundamental question revolved around a Soviet missile, designated the SS-9 or Scarp by the U.S. intelligence community. First deployed in 1966, the missile was estimated to weigh 450,000 pounds, stand ten stories high, and be capable of carrying a twenty-five-megaton warhead (weighing between 10,000 and 15,000 pounds) on a 7,000-mile journey. In August and September 1968, the United States monitored an SS-9 test in which three separate warheads, each judged to be capable of carrying enough nuclear material for a five-megaton blast, were dispensed by the missile.84

  The question facing the intelligence community was whether the new Mod-4 version of the SS-9 was the Soviet Union’s first missile with Multiple Independent Reentry Vehicles—MIRVs. A missile with three independently targetable warheads could drop each one on a different target. If they were sufficiently accurate, the warheads would have a high probability of destroying even hardened missile silos—and a Soviet force of 500 SS-9s would have a good chance of destroying a substantial portion of the Strategic Air Command’s 1,000 ICBMs. A conclusion that the new version of the SS-9 carried three accurate MIRVed warheads, and would be deployed in sufficient numbers, would provide support to the administration’s argument that the United States needed to deploy Safeguard. In contrast, a missile with multiple reentry but not independently targetable warheads—with MRVs—could only dump its warheads within a very
narrow area.

  The capabilities of the SS-9 Mod-4 had been addressed five months earlier in the annual national intelligence estimate on Soviet strategic forces (NIE 11-8-68). The estimate observed that the three-warhead tests were “not incompatible with tests leading to a multiple independently-targeted reentry vehicle (MIRV) capability.” The study observed that “we believe the Soviets could achieve an operational MRV employing three RVs in a modified SS-9 payload by late 1969.” However, the reentry vehicle system, as observed, would “degrade the overall accuracy and reliability of the SS-9 system.” Thus, an SS-9 with three warheads could still be effective against a single target but not multiple targets. MIRVs that could be employed against Minuteman silos probably could not be deployed until 1972.85

  The intelligence community’s collective judgment was accepted by outgoing Defense Secretary Clark Clifford, whose January 1969 posture statement declared the modified SS-9 triplet to be a MRV missile but not a MIRVed missile.86 Although the issue did not become public until March, the debate within the administration was already under way within weeks after Nixon assumed the presidency.

  Providing some of the intellectual ammunition for the new administration’s view was the Air Force’s Foreign Technology Division (FTD). FTD argued that the Mod-4 might possess some primitive characteristics of a MIRV. If the Soviets could delay the separation of each reentry vehicle from the launch platform for seconds or even a fraction of a second, they could send the missiles off on different ballistic paths. If the targets were close enough together, the SS-9 would be the functional equivalent of a MIRVed missile.87

  Among those left unconvinced by such arguments was David Brandwein, Carl Duckett’s successor as the director of the Foreign Missile and Space Analysis Center. On February 5, he wrote in his personal notebook that Lloyd Lauderdale had passed him a draft of a memo on the SS-9/triple RV prepared for John Foster, the Defense Department’s research and engineering chief. He observed that “it was pretty much the . . . party line, which says it could be made into a MIRV, they might do it without our knowing it, and it looks like this is what the Sovs have in mind. I disagree! My gut feeling is that it is MRV, not MIRV, and that is all it will ever be.”88

  For the next six weeks, Brandwein would attend a variety of meetings on the SS-9 triplet as the CIA and Defense Department battled over the MIRV issue—the outcome of which could support or undermine the new administration’s ABM program. Attendance at some meetings was part of a process of debate between rival viewpoints; at others, the objective was to ensure that the views of FMSAC and the CIA were represented along with those of the Defense Department.

  On February 7, Brandwein attended a DOD presentation to Secretary of the Air Force Harold Brown, along with Wayne Boring, a CIA analyst. Brandwein recorded in his diary that “as soon as we got into Brown’s office he [Boring] proceeded to tell Lloyd [Wilson of DDR&E] why the DDR&E . . . diagnosis of the SS-9/Triple RV = MIRV was much too positive. He had . . . found all the weaknesses which I had found independently! Wayne said I was grinning like a cat & that he felt like standing up and cheering! Afterwards Wilson was thoroughly deflated.”89

  But such efforts had no impact on Laird or the administration’s commitment to Safeguard. During his Senate testimony, Laird predicted that unless the United States deployed such an ABM system, the Soviets would have a first-strike capability in five years. At the time, there were only 228 SS-9s deployed, but Laird pointed to intelligence indicating that the Soviets had prepared six new SS-9 silos in December in an entirely new missile field. He suggested that might indicate a new wave of deployment—contradicting the view expressed in NIE 11-8-68 that deployment would soon come to a halt.90

  Research and engineering director Foster, convinced of the SS-9 MIRV potential by FTD’s analysis, provided some specific numbers during congressional testimony that spring. He argued that by the mid-1970s, the Soviet strategic rocket forces could have 420 MIRVed SS-9s with sufficient accuracy and reliability to wipe out 95 percent of the Minuteman force in a single strike. In such briefings, Foster ignored the cost to the Soviets for each missile ($25–30 million) as well as the need to reprogram missiles during flight.91

  Although Foster was unable to convince Brandwein, Duckett, or others from the CIA with such arguments, he did, at least for as long as it would matter, convince national security adviser Henry Kissinger. The consequence was a meeting in the White House Situation Room on Memorial Day during which Kissinger “beat up on” DCI Richard Helms, Deputy Director for Intelligence R. Jack Smith, Abbot Smith, chairman of the Board of National Estimates, and presumably Duckett, who was also in attendance, for their heretical views on the SS-9—views that were preventing Kissinger from obtaining congressional funding for Safeguard.92

  Between Laird’s March testimony and Kissinger’s tantrum, further SS-9 testing seemed to strengthen the Pentagon’s position. From April 20 to mid-May, three SS-9 triplets were tested. After analysis of the telemetry, TRW concluded that the triplets were landing in a triangular pattern that resembled the Minuteman deployment pattern. Analysts used such findings to claim that even if the SS-9 triplet was not a true MIRV, it was the “functional equivalent” of one—able to attack three distinct, geographically separated targets.93

  For a brief period, Laird appeared to believe that new evidence showed the SS-9 Mod-4 to be more than functionally equivalent. On June 6, Duckett asked Brandwein to investigate Laird’s claim to Helms that the Defense Intelligence Agency had demonstrated that the last SS-9 MRV test showed a three-in-a-line pattern, and that such evidence proved that a MIRVed missile was involved. Brandwein checked into the story and determined that Laird had been unintentionally misinformed by his research and engineering chief, John Foster, who at the time believed the evidence showed three warheads in a line because the refined, and contrary, data had not been passed on to him by an aide.94

  By mid-June, the intelligence community was working on a “memorandum to holders” of NIE 11-8-68. Such memos were sometimes prepared to update portions of NIEs concerning issues of immediate importance. On June 12, the United States Intelligence Board had agreed to a memorandum suggesting that the Soviet Union’s actions to improve its nuclear forces fell short of an effort to achieve a first-strike capability.95

  On Saturday June 13, the day after the USIB meeting, Helms, Abbot Smith, Jack Smith, Duckett, and Brandwein were summoned to a meeting with Kissinger and officials from Defense, State, and the NSC Staff. According to Brandwein, “Kissinger made it pretty plain he was unhappy with it! He kept saying he didn’t want to influence our judgements—but! It was a difficult 3 hours. Kissinger kept saying ‘his most important client’ wanted the facts separated from the judgements and identified as such. He also kept claiming he wasn’t a technical man, neither was his client—but the ‘facts’ are the technical details.”96

  Kissinger asked that the memorandum be rewritten to provide more clarification of some points and additional discussion pro and con about the MRV-MIRV issue, including more evidence of the judgment that the SS-9 was not MIRVed. Smith rewrote the draft with assistance from several FMSAC staffers as well as Brandwein, who rewrote the sections on “Is it a MIRV?” and the SS-9 range, but did not change the conclusion.97

  Then, on June 19, Nixon told a press conference that in proposing Safeguard, he did so “on intelligence information at that time,” and since that time, “new intelligence . . . with regard to Soviet success in testing multiple reentry vehicles . . . has convinced me that Safeguard is even more important.”98

  Meanwhile, Duckett had been assigned to help Kissinger understand the technical issues involved. Kissinger referred to Duckett as “Professor” because of his ability to explain technical matters in understandable terms, and Duckett later recalled the relationship:

  What was difficult was, when you have the national security adviser saying, “Look, the president of the United States and secretary of defense have said the following. Now, are you telling me
that you are going to argue with them. And the answer was, ‘No sir, I’m not going to argue with them, other than to tell you that they’re wrong.’ But that’s when a senior intelligence officer . . . has to be prepared to say, the fact is, we do have enough information . . . And in this case, we felt the technical argument was overwhelming.”99

  On June 23, at a hearing before the Senate Foreign Relations Committee, Richard Helms and Melvin Laird sat side by side. Helms noted that since he had last briefed the committee, the Soviets had conducted two additional long-range tests of the SS-9, and that “the intelligence community is in agreement that vehicles in these tests were not independently guided.”100 Laird, however, pointed to the recent SS-9 tests and argued that although the data were not conclusive, they were consistent with a MIRV capability. In addition, he noted that the level of deployment was not tapering off. He also claimed that there were no differences with regard to the meaning of the intelligence data, only an institutional difference, with the Defense Department planning for the worst case.101

  But the battle continued behind the scenes. Brandwein addressed the President’s Science Advisory Committee on June 25 and gave a two-hour briefing on the SS-9, covering accuracy, range, and whether it was MIRVed. After the FMSAC director presented both sides of the argument, PSAC member Richard Latter, Brandwein wrote, “went on to a blackboard lecture to tell them how his calculations showed SS-9 accuracy had improved. I didn’t think he convinced the others.”102