The Twilight of the Bombs

by Richard Rhodes

  The solution the administration found to its problem was denial. When the story of the test broke, on 25 October, Cyrus Vance informed a press conference that it was “not clear that there34 has been a nuclear detonation” and that the U.S. didn’t know “that anything has happened in South Africa.” Curiously, the South Africans took a similar line, a CIA document noted: “On 26 October, immediately following the announcement in Washington of the Vela indication, Jacobus de Villiers, President of South Africa’s Atomic Energy Board, told the press, ‘If there was anything of the sort, my first reaction would be that some other power might have undertaken a test, but it was definitely not South Africa.’”35 The CIA was suspicious, however, of a comment Botha had made three days after the Vela flashes. In a speech to a provincial congress, the South African prime minister said his country’s enemies ought to watch out, or they “might find out36 we have military weapons they do not know about.”

  To obfuscate the evidence of a nuclear test, the Carter administration appointed an outside ad hoc panel of scientists to consider if the evidence could have derived from natural phenomena. Led by Jack Ruina, an engineering professor at MIT and the former head of the Defense Advanced Research Projects Agency, the panel of nine included the physicist Luis W. Alvarez of the University of California at Berkeley, a Manhattan Project veteran, Nobel laureate, and skilled scientific detective. The panel met three times between 1 November 1979 and early April 1980. It concluded that the Vela signals came not from a nuclear explosion but from a micrometeorite dislodging material from the satellite body that reflected sunlight into the Vela’s sensors as it spun past—a sequence of unlikely coincidences to which a January 1980 Stanford Research Institute study gave odds37 of between one billion and one hundred billion to one.

  As it happened, I worked with Alvarez a decade later helping him edit several thousand pages of dictation into his popular memoir Alvarez: The Adventures of a Physicist.38 A tall, ruddy, California-born ice blond despite his Spanish name, the Berkeley physicist was stubbornly independent, politically conservative, and not likely to have participated in a cover-up for a Democratic administration (or a Republican one, for that matter). The challenge of finding an explanation for what he called a “zoo-on,” meaning an event so rare that it belonged in a zoo, evidently overcame his usual good sense. When I knew him, he and his geologist son Walter had recently parlayed the curious absence of fossils in a layer of otherwise fossiliferous sedimentary rock that Walter found in Gubbio, Italy, into a theory, controversial then but now strongly supported, that an asteroid impact sixty-five million years ago induced an “asteroid winter” across the world that led to the extinction of the dinosaurs. If Alvarez had personally investigated the Vela event, my guess is he would have concluded that it registered a nuclear test. The ad hoc panel, in fact, began its deliberations assuming that the satellite had seen the double flash of an atomic bomb. It then listened to various expert presentations, in the course of which the panel members changed their minds, dismissing a great deal of corroborative evidence in order to do so. Ruina understood that his assignment was deliberately narrowly framed. “My mandate was39 to only look at technical data,” he told the journalist Seymour Hersh in 1991. “We didn’t include any political data—like are the Israelis interested in nuclear weapons? That was not in our charter.”

  An important clue to the origin of the 22 September 1979 test was the modest scale of the explosion—unlikely for a gun bomb overloaded with HEU as South Africa’s bombs were. A U.S. interagency intelligence memo prepared for the NSC in December 1979, after reviewing evidence of South African naval activity on 22 September and systematically eliminating other candidates, proposed that the Vela event might have been “a secret test by Israel”:

  The Israelis might have conceivably40 foreseen needs for more advanced weapons, such as low-yield nuclear weapons that could be used on the battlefield. Or they might have considered desirable a small tactical nuclear warhead for Israel’s short-range Lance surface-to-surface missiles. Israeli strategists might even have been interested in developing the fission trigger for a thermonuclear weapon. If they were to have developed reliable nuclear devices for any of these weapons without access to tested designs, moreover, Israeli nuclear weapons designers would probably have wanted to test prototypes. A low-yield nuclear test conducted clandestinely at sea could have enabled them to make basic measurements of the device’s performance.… Indeed, of all the countries which might have been responsible for the 22 September event, Israel would probably have been the only one for which a clandestine approach would have been virtually its only option.

  Conducting a clandestine nuclear test off South Africa’s shore and allowing its ally to take the blame, however, was “not consistent with Israel’s policy or attitude toward Pretoria,” the interagency memo argued. That conclusion raised the possibility that the test had been conducted jointly by South Africa and Israel:

  If the South Africans had considered41 testing Israeli designs in exchange for Israeli technical assistance, the benefits of cooperation would have been carefully weighed by both parties against the security risks inherent in such joint operations. On the one hand, the Israelis would have calculated that South Africa, as a pariah state in need of reliable friends, would have had every reason to preserve security and to remain silent in the face of inevitable speculation about its complicity with Tel Aviv. The Israelis also could have counted as a high probability that responsibility for any nuclear test in the area under investigation would be attributed to South Africa.… Israelis have not only participated in certain South African nuclear research activities over the last few years, but they have also offered and transferred various sorts of advanced nonnuclear weapons technology to South Africa. So clandestine arrangements between South Africa and Israel for joint testing operations might have been negotiable.

  A joint test of an Israeli nuclear-artillery shell or the primary stage of a thermonuclear device would explain the discrepancy between the brute-force South African gun bombs and the low yield of the 22 September event. A joint test of an Israeli weapon would also account for Jacobus de Villiers’ comment that “some other power might have undertaken a test, but it was definitely not South Africa” and Botha’s warning to his country’s enemies to watch out. In these and other comments, the South Africans appear to have been toying contemptuously with the announced nuclear powers that maintained their own large stockpiles of nuclear weapons but punished South Africa’s efforts to acquire such weapons when it believed itself to be surrounded and at mortal risk.

  And a joint South African–Israeli test, as it turned out, was indeed what the U.S. Vela satellite had recorded on 22 September 1979. Seymour Hersh broke the story in his 1991 book The Samson Option, quoting “former Israeli government officials”42 who told him “that the warhead tested that Saturday morning was a low-yield nuclear artillery shell.” The Vela satellite event, his sources also told him, “was not the first but the third test of a nuclear device over the Indian Ocean. At least two Israeli Navy ships had sailed to the site in advance, and a contingent of Israeli military men and nuclear experts—along with the South African Navy—was observing the tests. ‘We wouldn’t send ships down there for one test,’ one Israeli said. ‘It was a fuck-up,’ he added, referring to the capture of a test by the Vela satellite. ‘There was a storm and we figured it would block Vela, but there was a gap in the weather—a window—and Vela got blinded by the flash.’”

  A well-informed but anonymous source told me that only one weapon was tested, however, despite Israeli claims to Hersh, and that South Africa’s participation was limited to supplying the island test site. The weapon was, indeed, an Israeli artillery shell, the source said.

  BY THE MID-1980S, Israel’s military and security association with South Africa had become an acute embarrassment in a world increasingly active in opposing apartheid. Besides the 1979 nuclear test, the two countries had collaborated on developing a variety of convent
ional arms and other military materiel. “We created the South43 African arms industry,” Alon Liel, a former Israeli ambassador to South Africa, told the British journalist Chris McGreal in 2006. “They assisted us to develop all kinds of technology because they had a lot of money. When we were developing things together we usually gave the know-how and they gave the money. After 1976, there was a love affair between the security establishments of the two countries and their armies. We were involved in Angola as consultants to [the South African] army. You had Israeli officers there cooperating with the army. The link was very intimate.” The Israeli security establishment resisted Israel’s mid-1980s withdrawal. “When we came to the crossroads44 in ’86-’87,” Liel told McGreal, “in which the foreign ministry said we have to switch from white to black, the security establishment said, ‘You’re crazy, it’s suicidal.’ They were saying we wouldn’t have military and aviation industries unless we had had South Africa as our main client from the mid-1970s; they saved Israel. By the way, it’s probably true.”

  Israel’s withdrawal from cooperation was less painful for South Africa than it might have been. Since the late 1970s, even as the number of South Africans drafted into military service doubled45—a measure of the country’s perception of internal and external threats—the apartheid state had been restaging its nuclear ambitions. While it researched implosion and boosted-fission weapons, which would require higher-enriched uranium as well as tritium, the country had given immediate priority to developing a minimal deterrent that would work, uniquely, not only by threatening South Africa’s enemies but also by challenging her nominal allies, the United States in particular, to come to the pariah country’s aid.

  In July 1979, when South Africa decided to build seven gun bombs, it had turned the gun-bomb program over to Armscor for development. The nuclear analyst David Albright has speculated that South Africa limited its nuclear arsenal because it realized that using nuclear weapons “would have been akin46 to committing suicide.” But the fact that it was actively researching more advanced implosion systems and had purchased tritium from Israel at great expense to investigate boosting yields* argues for another, hedged scenario: build a few gun bombs first as a deterrent or as weapons of last resort (prodigal as they were of HEU) while developing smaller, more efficient tactical nuclear weapons for war-fighting in the event deterrence failed.

  This scenario would explain South Africa’s willingness to risk censure or worse to host the September 1979 Israeli test of a nuclear artillery shell—exactly the sort of weapon that South Africa itself might want to add to its arsenal eventually but that was currently beyond its technical means. It would explain why South Africa in 1977 and 1978 had exchanged fifty tons of yellowcake uranium for thirty grams of Israeli tritium,47 supplied in 2.5-gram capsules over a period of months. It would explain why Israel and South Africa collaborated extensively in the 1980s on missile development. It is supported in part by something Armscor officials told Albright in 1993: that Armscor had planned to “replace the seven48 cannon-type devices with seven upgraded devices when [the gun bombs] reached the end of their estimated life by the year 2000.” The upgrade, Albright understood, would involve switching from gun assembly to implosion.

  South Africa’s deterrent strategy, AEC chairman de Villiers said in 1993, had involved three phases:

  Phase 1 called for49 strategic ambiguity: internationally, the government would neither confirm nor deny whether it possessed a nuclear weapons capability. If the country were threatened militarily, Phase 2 required South Africa covertly to reveal its nuclear capability to leading Western governments, principally the United States. Should Phase 2 fail to persuade the international community to intervene to alleviate an armed attack from outside South Africa’s borders, Phase 3 required Pretoria publicly to disclose its nuclear arsenal, either by official acknowledgement or an underground nuclear test.

  There was an unspecified Phase 4 to the South African strategy: actual use of the arsenal against an enemy. Armscor designed its gun bombs to be airdropped by bomber if no other country came to South Africa’s aid. But beyond even that phase of nuclear strategy, wrote Los Alamos’s Frank Pabian, “the questions that must be answered50 are: Why did the South Africans go to the added expense of constructing a second generation nuclear weapons manufacturing complex [at Pelindaba] when they already had a ‘credible’ deterrent, and why did South Africa complete the infrastructure for building 3,000-kilometer-range ballistic missiles designed for delivering nuclear weapons?”

  South Africa never answered these questions. Its claim of having built only seven gun bombs, like its claim not to have exploded a nuclear weapon on 22 September 1979, is evidently a half-truth. The whole truth is that its small cache of gun bombs was a picket guard in advance of a full-scale nuclear arsenal, with means of delivery that would have included everything from atomic artillery to ICBMs—the ICBMs to hold Soviet cities at risk, a “Samson option,” as Hersh calls it, that Israel had already implemented. Deliverable nuclear weapons equalize the capacity of smaller powers to do unacceptable harm even to superpowers, however invulnerable those superpowers may imagine themselves to be—which is why superpowers have, or ought to have, a fundamental security interest in eliminating nuclear weapons from the arsenals of the world.

  Given South Africa’s belligerence and dogged determination, why did it turn around at the end of the 1980s and dismantle its arsenal under IAEA supervision? It did so primarily because its security needs had been met. The decision, Pabian points out, “was due in large part51 to the removal of the external military security threat brought about by the Angolan/Namibian peace settlement and the collapse of the Soviet Union. As President F. W. de Klerk stated [later], ‘In these circumstances, a nuclear deterrent had become, not only superfluous, but in fact an obstacle to the development of South Africa’s international relations.’”

  An uglier reason was Afrikaner determination not to allow nuclear weapons to pass into the hands of the inevitably approaching South African black-majority government. Waldo Stumpf admitted as much in an interview with a Washington Post correspondent in 1991. When “questioned about the reasons52 behind the change of government thinking about the objectives of its nuclear program,” the Post’s David B. Ottaway wrote, “Stumpf said that the prospect of black-majority rule in the ‘New South Africa’ within a few years was a major consideration.” This consideration explains, at least in part, the Armscor/AEC joint effort, beginning in 1990, to eliminate not only the seven gun bombs in storage at Pelindaba but also to destroy all the program’s documentation.

  De Villiers describes the stark dénouement:

  The Y Plant was closed53 on February 1, 1990. The working group reported that it would need about 18 months to fully dismantle the country’s nuclear deterrent capability, and Armscor and the AEC were jointly entrusted with this task. They studied the problem for five months before beginning in July 1990. An independent auditor, directly responsible to de Klerk, was charged with overseeing the dismantlement of the six assembled nuclear devices and ensuring that the HEU from each, as well as from the incomplete seventh device, was removed from Armscor’s custody and returned to the AEC. He was also to confirm that all technology and hardware were destroyed. The entire process was completed by early July 1991.

  On 10 July 1991 South Africa signed the NPT. De Klerk, to avoid the kind of intrusive inspections ongoing at the time in Iraq, waited until 1993 to announce his country’s nuclear disarming to the world. However racist South Africa’s domestic purpose was in disposing of its nuclear arsenal, its fundamental reason for doing so was to increase—not decrease—its security. It not only came in from the cold of international pariahdom; it also set a unique and instructive example for the world.

  * Thorium, element 90, atomic weight 232, is fertile but not fissile; it can be irradiated in a nuclear reactor to produce uranium-233, which has fissile properties similar to plutonium (its bare critical mass equals 16.13 kilograms compared to
10 kilograms for plutonium and 47.53 kilograms for U235), although its intense gamma radiation (from an impurity, U232) makes it more difficult to fabricate into bombs. The United States tested a composite U233/plutonium-core implosion device during the Teapot series at Yucca Flats, Nevada, in April 1955; the 800-pound device yielded 22 kilotons.

  * A laager is an encampment made by circling the wagons.

  * Most of the energy of a nuclear explosion is released in the first one-millionth of a second, before the fireball has had time to expand beyond the volume of the physical bomb itself. This energy-dense early fireball radiates ten-million-degree X-rays, which cook up so much ozone and nitrous oxide in the surrounding air that the fireball radiation is veiled; this initial opacity marks the low starting point on the left side of the graph. As the fireball continues to expand, a shock front of heated air forms around it that engulfs cold air and cools to visibility. At the graph’s first peak, the temperature of the shock front is about 10,000° C, the temperature of intense blue-white light. The shock front is still opaque to the fireball inside, but it becomes increasingly transparent as it continues to expand and cool. Within a few ten-millionths of a second after detonation the shock front has cooled to about 3000° and the interior fireball—still radiating beyond the range of visible light—shines through. This is the first minimum on the graph. Now the fireball itself cools to visibility—the second peak of light, with a temperature of about 7000°. From that point the decline in brightness simply registers the progressive loss of energy as the fireball continues to cool with expansion.