美国核弹制造者1979年开始访问中国时，对中国人越来越直接的提问感到吃惊，他们的提问表明中国的同行急切希望得到制造现代核武器的秘密。那些秘密可以使氢弹变得很小，许多能装在一枚导弹上或者从卡车、潜艇等机动平台上发射。
中国在1992年9月25日取得成功，这个消息来自一名间谍，他对美国的接头人说，北京爆炸了一颗以小型化秘密技术为基础的炸弹。
新墨西哥州洛斯阿拉莫斯武器实验所一组科学家开始进行一项有重大影响的侦探调查：中国的进步是间谍活动的结果，还是自己艰苦劳动的结果，还是两者的结合?
这场辩论还在激烈地进行着。专家们一致认为出现了间谍活动，但是对多少秘密被窃和它对北京的进步有多大作用分歧很大。
洛斯阿拉莫斯的小组在1995年作出了结论，说中国的进步大概是以间谍活动为基础的。1999年国会一个委员会的报告走得更远，声称“没有从美国盗窃的核机密”，中国“几乎不可能”造出小型弹头。
国会的报告遭到政府内外科学家的批评，他们说报告夸大了间谍活动的作用，中国很可能靠自己的力量取得突破。
一项以几个月的采访以及透露的武器和情报秘密为基础的研究表明，国会的报告声称盗窃的机密是中国取得突破的主要原因没有证据。
这项研究支持争论不休的专家逐渐形成的一致看法：联邦调查立早地集中于洛斯阿拉莫斯国家实验所和那里的一名工作人员(李文和)。现在看来，失窃的机密全国与武器有关的单位数百人、甚至数干人都能接触到。
对洛斯阿拉莫斯侦探小组来说，整个间谍理论在1995年得到有力支持，当时中央情报局得到一份中国内部文件，其中有对美国最先进的小型弹头W-88的说明。联邦官员在首次透露他们在此案中最重要的证据即文件的秘密内容时说，中国的文件列出了这种弹头的五种关键的特点，包括精确到4/100英寸的两种测量法。
但是批评者坚持说，即使北京进行了间谍活动，小型化的突破也是靠它自己的力量，它进行这项工作至少有13年(从1979年到1992年)。
美国专家说，中国科学家的才能从他们制造的一种相机中可见一斑。这种相机拍摄的核爆炸的照片比美国类似的相机拍出的照片好得多。
前洛斯阿拉莫斯的负责人霍华德・阿格纽说：“他们不需要我们的任何帮助。他们只是好奇，就像我们对他们好奇一样。”阿格纽访问过中国，是联邦情报顾问。
解析间谍活动造成的损失是一门涉及到推理、求证和作出结论的不精确的艺术。在已知和怀疑之间的空白常常插进不少个人、党派或者机构的偏见。
同在多数间谍案中一样，证据可以作多种解释。一些了解中央情报局获得的中国文件的批评者说，文件对美国弹头的描述本身不足以制造小型化的弹头。
能源部主管洛斯阿拉莫斯调查的官员诺特拉・特鲁洛克同意这种估计，但是说这些情报是秘密的，从来没有在任何公开文件或者因特网上公布过。他和他手下的人分析，任何得到这些情报的人必定能够得到多得多的弹头设计秘密。
特鲁洛克在接受采访时说，知道各部件近似的体积和形状可以为中国核弹制造者提供一张线路图，也许可以使他们省去多年初步试验的时间。
然而特鲁洛克接着说，国会委员会的报告太绝对。报告部分是以他的证词为基础的。
他说：“我在作证时为防止误解而作了适当的解释。说明对我们的证据和结论是不肯定的。我们一般说：‘也许是这样，也许是那样。”’他说国会委员会对间谍活动在中国突破中的作用“作出判断”。
中国迟迟才加入核俱乐部，而在加人时却表现出相当高的技巧。
北京在1964年爆炸了第一颗原子弹。像美国投到日本广岛的原子弹一样，这种铀弹设计巧妙，不仅节约了价格不菲的燃料，而且使原子弹更轻巧，从而增加了它的军事价值。
北京仅在32个月后就制造出第一颗氢弹。
华盛顿负责监督核武器的私人机构自然资源保护委员会的罗伯特・诺里斯说，相比之下，从核武器发展到热核武器，伦敦花了历个月，莫斯科花了75个月，华盛顿花了解个月，巴黎花了103个月。
中国仅试爆6次就达到了热核武器的程度，而美国进行了31次。中国试爆的次数一向很少。尽管北京研制了至少6种类型的武器，但它几十年来进行的核试验相对而言较少，总共45次，美国则有1030次。
在理查德・尼克松总统1972年访华之前。美国情报机构对中国的核计划和现代化计划知之甚少。但在尼克松外交行动打开了大门之后，美中两国就建立了军事关系。
到了1979年，美国的核武器设计者和安全问题专家纷纷开始造访中国同行、武器实验所以及位于中国西部沙漠的罗布泊。中国的原型核武器就是在那里爆炸的。
据前中央情报局官员罗伯特・弗鲁曼说，从1979年至1990年，光是从格斯阿拉莫斯去中国访问的科学家和官员至少有85名。弗鲁曼当时负责洛斯阿拉莫斯的反情报工作。
高级访问者包括该武器实验所已故所长阿格纽；负责情报工作的丹尼・斯蒂尔曼；后来担任里根总统的科学顾问的物理学家乔治・基沃思。
根据判断，访问带来的好处远远超过武器专家在非正式环境或谈话中可能有意无意泄露机密带来的风险。其实，美国人了解到很多情况。
一个发现就是，中国计划的一部分是十分先进的，其中包括原子弹研制技术。
阿格纽说：“他们拥有精良的装置，其中一些强于我们。”阿格纽是在1979年和1982年首批访华者之一。
阿格纽说，例如，中国人能够利用一种先进的照相机观察烈焰滚滚的爆炸，从而把观察到的细节用于弹头研制工作。阿格纽说，美国的设备只有一个轴，而中国有两个，其用途自然也多出一倍。他说：“(中国的设备)要好得多。”
美国访问者还了解到中国的不足之处。根据多年来接二连三的调查结果，北京显然急于尽一切可能掌握有关如何缩小原子弹引爆装置的技术。美国官员说，这些是普遍问题。它们越来越突出，从未有过答案。他们坚持认为，北京在这个方面没有秘密。
美国官员们透露，中国最终在1992年9月25日成功地爆炸了一颗小型化的原子弹。情报机构分析家们花了两年多时间才完全弄明白中国取得了什么成就。在一位受雇为美国刺探情报的中国核专家向美国上司递交了一份报告之后，美国才了解到中国取得的这一成就。
这个间谍说，美国分析家最初认为中国在9月份的试爆是常规试爆，但实际情况决非如此。这位间谍说，当日试爆的这颗原子弹已被小型化，只有一个核心，形状为特殊的卵形，这表明中国已开始掌握现代化弹头的制造技术。
在90年代中期，追踪其它国家核计划底细的任务落到国家武器实验所头上。刺探者之一就是洛斯阿拉莫斯经验丰富的武器设计者罗伯特・亨森。自从1988年以来，亨森一直在分析有关外国计划的情报。
亨森在接受记者采访时说，1995年元月，他开始更加密切关注中国如何解决小型化的难题。为此，他求助于专门分析俄罗斯武器计划的劳伦斯・布思。
他们起草了一份分析报告，最后交给了特鲁洛克，此人一年前担任管理洛斯阿拉莫斯的能源部情报部主任。特鲁洛克获得过政治学学士学位，没有受过任何正规的技术训练。他说，他希望邀请其他一些核专家参加，尤其是那些长期从事氢弹的小型核引爆装置研究的专家。洛斯阿拉莫斯的约翰・L．里克特这位科学家正合需要，于是加入了这个小组。
官员们在一份刚刚公开的材料中说，这个小组仔仔细细地研究了那名中国间谍提供的线索，这名间谍说，这种炸弹的原子起动装置核心的大小类似一种寻常家用物品。从这个问题开始人手，科学家们对尺寸大小进行了更加精确的推算，亨森博士和里克特博士查遍了美国的核武器储备，看看有没有尺寸大小与此相符。
他们发现，W-88的原子弹起动装置在尺寸大小上非常接近，足以引起人们的怀疑。
能源部举行了多次会议，中央情报局和国防情报局的分析家们也同洛斯阿拉莫斯小组一道出席了会议。联邦官员们现在说，这些情报机构表示怀疑，它们认为一个外国人的粗略类比受到了过分的重视，但是，能源部和洛斯阿拉莫斯小组觉得这种证据令人振奋。
正如以前披露的那样，1995年出现了突破性的进展，当时，一名中国政府官员给美国官员寄来了一包秘密的中国文件。
特鲁洛克说，最能透露内情的一份文件注明日期是1988年，这份文件是为生产导弹和头锥的第一机械工业部制订的核现代化计划。这份文件不仅描述了中国的计划，还拿它们同美国武库中的核武器进行比较。
据联邦官员们透露，中国功这份文件翻译后大约长达20页，这份文件还提供了有关W-88的敏感资料。它精确地描述了这种初级装置的形状，说它不是圆形，并说它位于头锥狭窄的前部靠后的位置，美国的有些弹头也是这样安排，但不是全部。这份文件还正确地描述这种次级装置是一种圆形的形状。更让这个小组困惑的是，这份文件还描述了原子起动装置包装盒子的宽度，精确到了毫米的程度，即相当于一英寸的百分之四。特鲁洛克回忆说：“真是精确极了。”
中央情报局最终断定，寄来这些文件的那名特工一定受到了中国情报部门的指使。关于中国为何给美国间谍寄送这些文件，到现在为止，还没人提出一个更有说服力的解释。
能源部于1995年9月28日对可能存在的盗窃W-88机密的事件展开了调查，在过去的3年里，联邦官员们不动声色，试图查明他们中间是否有人就是中国的间谍。
特鲁洛克和他的小组分析道，如果存在间谍活动，那么，这种间谍活动肯定发生在1984年和1988年之间。1984年，这种弹头进入设计研制阶段；而中国这份材料的注明日期是1988年。
能源部的官员们把注意力集中到设计这种炸弹的洛斯阿拉莫斯的身上。他们尤其仔细地审查了这些年中每位去过中国或者接待过到访的中国科学家的人。
1996年5月，能源部向联邦调查局递交了一份12名受到怀疑的人的名单。然后联邦调查局进行了刑事调查，最终日标缩小到了李博士的身上。李是一名美国科学家，出生在台湾，在洛斯阿拉莫斯实验室工作。
弗鲁曼是当时洛斯阿拉莫斯实验室反情报部门的主任，他后来对这种调查提出了直言不讳的批评。他指责这种调查受到了种族偏见的站污，它的矛头是美籍华人。联邦官员们极力否认了这种指责。但是，参议院政府事务委员会共和党主席和民主党首席委员一致认为，联邦调查人员过早地把注意力集中到了李博士的身上。参议院政府事务委员会负责这一间谍案的调查，并且听取了弗鲁曼的证词。
如果不是一系列与此无关的有关中国问题的报道，这场调查很有可能不会引起公众的注意。1998年4月，《纽约时报》报道了两家美国航天工业公司因向中国科学家提供火箭资料而受到刑事调查的事情。
这个问题在国会引起了轰动。众议院成立了一个由考克斯牵头的特别委员会，负责调查政府在卫星出口方面日益开放的政策是否损害了国家安全。
当时，没有任何迹象表明这个委员会最终会调查到核炸弹问题上来。这个由5名共和党人和4名民主党人组成的委员会没有听说中国涉嫌从事核间谍活动这件事情，直到1998年10月，此时距该委员会使命结束仅仅还有几个月的时间。在11月12日和12月16日，该委员会举行了听证会，特鲁洛克作为主要证人被传到会作证。
1月，经过5个月的调查，该委员会完成了一份秘密报告。5月，在就什么东西可以公开的问题同白宫经过一番长时间的争论之后，该委员会公开了一份长达872页的报告。关于原子技术谍报活动的章节只有37页，但是各家媒体对此问题报道的最多。
该委员会的4名民主党人在这份报告上签了字。但是，就在这份报告公开之后不久，其中一名民主党人、来自南卡罗来纳州的约翰・斯普拉特众议员对这份报告提出了批评，说这是一份匆忙赶制的报告，草率肤浅，夸大事实。他又说，该委员会查访的一些证人“没有那种技术背景，可以充分评估丢失的那些情报的性质和价值”。
自那时以来，斯普拉特的批评意见引起了许多著名科学家和炸弹设计人员的共鸣，而且还有所增大。他们说不用进行间谍活动，北京可能已经自力更生实现了自己弹头的小型化。长期以来一直替华盛顿在核武器问题上出谋划策的物理学家理查德・加温最近参加了前国防部长唐纳德・拉姆斯菲尔德牵头的一个两党小组。他说： “没有理由相信中国不能依靠自己研制的核技术替一系列现代导弹生产完全合适的弹头。”
几名官员说，中国所走的道路与其他核大国完全一样，对美国所取得的成果的大致了解让它们从中获得了帮助，证据是：氢弹可能造得很小，但是仍然具有很大的威力。
洛斯阿拉莫斯实验室国际安全研究部门的负责人──因间谍案显然处于守势──霍金斯说，炸弹和导弹的基本物理特性促使武器设计人员朝着总体一致的方向前进。他说，为了取得最佳性能，工程技术人员不可避免地都会进行大约16度宽的狭窄头锥的研究──如果是一块蛋糕，这种也是非常小的一片。霍金斯说：“一旦认识到这个问题，它就会促使每个国家沿着相似的道路前进。最后，所有的武器系统看上去都差不多。这更多地关系到物理特性，而不是间谍活动。”
这种观点得到了普遍接受。
但首先在洛斯阿拉莫斯拉响警报的分析家亨森博士说，在导弹头锥设计方面，没有任何东西促使一名科学家把原子起动装置的核心制造成椭圆的形状。他说：“毫无疑问，严重的间谍活动肯定发生过。”
美国情报机构的态度则没有那么坚决。分析家们得出结论认为，间谍案在北京的改进方面确实起了作用，但是，他们却苦于找不到一个像在本世纪四十年代苏联盗取美国第一颗原子弹设计方案那样的硬性联系。
一位曾认真审查过这份秘密材料的政府高级官员说：“大家都得出了同样的结论。但我们没有确切的证据。”
美国情报机构曾对这起间谍案作过一份损失评估报告，并于1999年4月公布于世。这份评估报告说，间谍案、公开可以得到的资料和科学的敏锐等因素加起来使中国迈出的步伐更大了。报告说，这些“被盗的机密”“可能帮助”北京研制出一种移动式导弹并且“可能加速了研制未来的核武器的计划”。
6月份，也在独自进行调查的总统国外情报顾问委员会说，国会和政府领导人都参与了对这起间谍案的“简单化和夸大化”。该委员会说，无论是重大损失评估还是断然的否认，都没有充分的事实依据。
洛斯阿拉莫斯核试验和核技术实验室的官员霍金斯说，已确知被中国人盗走的具体机密，主要是在1995年文件中详细列举的那些，对于导弹制造者来说没有什么助益，而且距特鲁洛克的公路图相距甚远。至于被设计者称之为物理包的氢弹内部，霍金斯说，这些文件“没有介绍任何重要的内容”。
一位联邦官员用有关中国原子弹的情报数据来说明，这根本不是一个真正的翻版。
这位官员说：“事实证明，W-88要稍小些。”他认为，北京可能自行作了改进。
中国是如何从第一地点获得W-88的机密的，这个问题至今仍然是个谜，但是，人们已经普遍开始认为，把泄秘事件的范围一下子缩小到洛斯阿拉莫斯实验室实在是太快了。
参议院和总统的国外情报顾问委员会1999年进行的研究都提出了严肃的问题，即联邦调查局和能源部把目标对推这家武器实验室是否太快了。没有任何证据把该实验室作为泄秘事件的源头。
从1987年到1998年一直担任该实验室反间谍机构负责人的弗鲁曼注意到，描述W-88弹头设计的一份机密文件通过政府和军方发给了548个邮件地址。一些政府专家相信，中国人在1995年的文件中描述的资料取自于工程计划或军事基地的秘密手册。
曾参与总统顾问委员会的调查工作的斯坦福大学德雷尔博土说：“那种情报比比皆是，发行的手册既有照片也有数字。”
(美国《纽约时报》报道)(载于《参考消息》1995年9月15日15、16版)
=======
September 7, 1999
Spies vs. Sweat: The Debate Over China's Nuclear Advance
By WILLIAM J. BROAD
When American bomb makers began visiting China in 1979, they were startled by increasingly pointed questions that suggested their Chinese peers were hot on the trail of the secret to building a modern nuclear arsenal. It allows hydrogen bombs to be made so small that many can fit atop a single missile or be fired from trucks, submarines and other mobile platforms.
China succeeded on Sept. 25, 1992, the news coming from a spy who told his American handlers that Beijing had exploded a bomb based on the miniaturization secret.
A team of scientists at the Los Alamos weapons laboratory in New Mexico set to work on a whodunit with huge implications: Was China's advance the result of espionage, hard work or some mix of the two?
Today, the debate rages on. Experts agree that spying occurred, but clash violently on how much was stolen and what impact it had on Beijing's advance, if any.
The Los Alamos team concluded in 1995 that China's stride was probably based on espionage. A report this year by a Congressional committee that made the case public went further, claiming that it would have been ''virtually impossible'' for China to have made small warheads ''without the nuclear secrets stolen from the United States.''
The Congressional report unleashed criticism from scientists inside and outside the Government who said the importance of the espionage was overstated, and that China could well have achieved the breakthrough on its own, as it insists publicly.
A review of the dispute, based on months of interviews and disclosures of weapons and intelligence secrets, suggests that the Congressional report went beyond the evidence in asserting that stolen secrets were the main reason for China's breakthrough.
The review also bolsters a point of emerging agreement among feuding experts: that the Federal investigation focused too soon on the Los Alamos National Laboratory and one worker there, Wen Ho Lee, who was fired for security violations. The lost secrets, it now appears, were available to hundreds and perhaps thousands of individuals scattered throughout the nation's arms complex.
Federal officials asked in recent days that some details about weapons design and intelligence sources not be published, and The New York Times agreed to withhold them.
For the Los Alamos team of detectives, the overall spy theory was supported strongly in 1995 when the Central Intelligence Agency obtained an internal Chinese document that included a description of the United States' most advanced miniature warhead, the W-88. Revealing for the first time their top evidence in the case, the document's secret contents, Federal officials say the Chinese text cited five key attributes of the warhead, including two measurements accurate to within four-hundredths of an inch.
But the critics, who are also revealing new information, insist that Beijing, even if it spied, made the miniaturization breakthrough on its own, pursuing it for at least 13 years, from 1979 to 1992.
The prowess of Chinese scientists, American experts said, is suggested by a camera they built for photographing nuclear blasts, which was far better than a similar one made by the United States.
''They don't need any help from us,'' said Harold Agnew, a past Los Alamos director, visitor to China and Federal intelligence adviser. ''They're just curious, as we are curious about them.''
Deconstructing the damage wrought by espionage is an imprecise art that mixes inference, evidence and deduction. In the vacuum between what is known and what is suspected, personal, partisan or institutional bias often rushes in.
The debate over Chinese spying has been blurred by issues that include Republican distaste for President Clinton's China policy, accusations of racial bias in the investigation and fears among scientists that the uproar is prompting security measures so tight as to damage work, morale and recruitment.
As in most spy cases, the evidence is open to interpretation. Several critics familiar with the Chinese document obtained by the C.I.A. said that its description of the American warhead was not by itself sufficient to build a miniaturized warhead.
The Energy Department official who supervised the Los Alamos inquiry, Notra Trulock, agreed with this assessment but said the information was secret and had never been mentioned in any public document or Internet posting. Anyone who had it, he and his team reasoned, must have also obtained access to a much broader range of secrets about the warhead's design.
In addition, Mr. Trulock said in an interview, knowing the approximate size and shape of the components provided a road map to Chinese bomb makers, probably allowing them to skip years of preliminary testing.
Mr. Trulock added, however, that the Congressional committee was too categorical in its report, which was based in part on his testimony.
''When I testified, I used the appropriate caveats to express uncertainties in our evidence and our conclusions,'' said Mr. Trulock, formerly the Energy Department's intelligence chief. ''We typically said: 'Probably this. Probably that.' '' The committee, he said, ''made judgments'' about the centrality of spying in China's breakthrough.
Representative Christopher Cox, a California Republican who was chairman of the committee, defended the work of his staff of 47, which included no one with nuclear design experience. The panel, he said in a lengthy interview, drew largely on Clinton Administration witnesses for its expertise. The conclusion that espionage allowed Beijing to skip decades of research, he said, was an appropriate one, based on the Government's own evidence.
''Judgment matters,'' he said, responding to Mr. Trulock's criticism. ''We don't know everything to a certainty. The question is what is more likely than not.''
In the interview, Mr. Cox expressed surprise when told of the depth and breadth of China's interest in the miniaturization secret. He also played down the idea, cited by Federal skeptics of Chinese spying, that most of the world's nuclear powers have figured out the secret of miniaturization.
Can China, Mr. Cox asked, ''develop it indigenously because France did? That is a stretch. It's almost apples and oranges.''
The Secret
America Shrinks An Atomic Match
From the dawn of the nuclear age, miniaturization has been an obsession of weapons designers. The world's first atomic bomb, designed by the Los Alamos laboratory and detonated in the New Mexico desert in July 1945, was an awesome but cumbersome affair. A lump of plutonium the size of a softball was surrounded by a much larger ball of high explosives that was five feet wide and made up of 32 explosive charges and 64 detonators. Big as a car, it could not have fit into a small airplane, let alone a missile.
In 1952, American physicists made an important breakthrough: the H-bomb. Roughly a thousand times more powerful than the first atomic weapon, the hydrogen bomb was a two-stage device. Inside its dense casing, an atomic explosion -- called the primary -- worked as a match to kindle an even more powerful detonation by the bomb's hydrogen fuel, which was known as the secondary.
Size was an issue from the start. The first hydrogen bomb stood two stories high and weighed 82 tons. It would be militarily useful only if it could be shrunk, and over the next few years, the country's best physicists set out to do just that. After considerable trial and error, they figured out that they could obtain the same kind of explosive power from a smaller package. A main breakthrough centered on the large, heavy atomic match. By shaping its plutonium fuel into an ovoid, roughly like a watermelon, scientists were able to drastically shrink the size and number of the explosives that triggered the nuclear blast.
After at least one flop, the radical idea roared to life in July 1957 in a nuclear explosion in the Nevada desert, according to Chuck Hansen, author of a detailed history of America's early nuclear efforts. It had taken the United States a little more than five years to move from the first hydrogen bomb to its miniaturized cousin.
The development had profound implications for the cold war's nuclear competition.
Shrinking the atomic trigger from something roughly the size of a washing machine to something smaller than a football allowed weapons designers to put thermonuclear arms atop small missiles that could be launched from submarines or mobile platforms like trucks. Arms would no longer be confined to bombers or silos in the ground.
The advance meant weapons could now be carried, quite stealthily, closer to enemy shores and could be made safer from attack. It also meant warheads could fit into the cramped spaces of narrow nose cones, which streaked faster to Earth than blunter shapes and were less buffeted by winds during the fiery plunge, making them more accurate.
The first warhead in the new generation of weapons, the W-47, was less than half the size of the bomb that leveled Hiroshima but up to 80 times more powerful. In 1960, when the first Polaris submarine put to sea, each of its 16 missiles was armed with a W-47.
The weapons continued to evolve, and by all accounts, the apex was reached in the 1980's with the W-88, one of the most deadly weapons in the American arsenal.
The warhead, made for submarines, first went to sea a decade ago and is considered quite powerful for its small size. The precise size is secret. But at least eight W-88's can fit atop the Trident D-5 missile, which is less than seven feet wide. Since Trident subs have 24 missiles, a single submarine can carry up to 192 of the thermonuclear arms.
Today, American submarines on patrol in the Atlantic carry the small warheads. And the Navy is adding them to its Pacific fleet, so in the next few years the W-88 is likely to be aimed at China.
The Chinese
Late to Start, Quick to Excel
China was late in joining the nuclear club, but showed considerable skill when it did.
Beijing detonated its first bomb in 1964. The tricky design was based on uranium, like the Hiroshima bomb, but saved costly fuel and made the bomb lighter, increasing its military value.
Sidney D. Drell, a Stanford physicist and Clinton Administration adviser, writing in ''China Builds the Bomb'' (Stanford University Press, 1988), called the feat ''enormously impressive.'' Beijing's first hydrogen bomb came just 32 months later.
By comparison, the step from nuclear to thermonuclear took London 66 months, Moscow 75 months, Washington 87 months and Paris 103 months, said Robert S. Norris of the Natural Resources Defense Council, a private group in Washington that monitors nuclear arms.
China set off just 6 test blasts to get to the H-bomb stage, versus 31 for the United States. The low number was typical. While developing at least six types of weapons, Beijing over the decades conducted relatively few nuclear tests, 45 in all, versus 1,030 for the United States.
The evidence strongly suggests that China, in its first phases of missile building, had no idea how to shrink thermonuclear arms. According to ''China's Strategic Seapower'' (Stanford University Press, 1994), the warhead for the submarine missile deployed by Beijing in the 1970's weighed 1,300 pounds, more than twice the old American W-47, suggesting that the Chinese were still using a spherical atomic match to ignite hydrogen bombs.
China's land force was modest. Starting in the 1980's, it deployed about 20 missiles that can now reach anywhere in North America, each topped by a single warhead that can unleash a force equivalent to up to five million tons of high explosives. That is about 300 times stronger than the Hiroshima bomb.
The big warheads are not particularly accurate, but they fit China's professed war doctrine -- to fire nuclear arms only in retaliation. The big missiles can, if necessary, hit a city.
China's interest in building smaller weapons was spurred, in part, by the United States' development in the late 1970's of a high-accuracy design known as the Missile Experimental, or MX, that bristled with 10 warheads. Though meant primarily to unnerve Moscow, the weapon also worried Beijing, which quickly grasped that its handful of big land-based missiles looked like sitting ducks that could be destroyed in a first strike of precisely aimed hydrogen bombs.
Beijing's unease grew as the American Navy in the late 1970's unveiled plans for a new submarine-launched missile nearly as unerring as the MX and bearing an even more powerful warhead -- the W-88.
American intelligence agencies knew little about China's nuclear program and modernization plans, if any, before President Richard Nixon's visit to China in 1972. But the military ties that followed the Nixon diplomatic initiative opened the door.
By 1979, American nuclear arms designers and security experts were starting to visit their Chinese peers, weapons labs and Lop Nur, the sprawling site in China's western desert where prototype nuclear weapons were detonated.
From Los Alamos alone, at least 85 scientists and officials made trips from 1979 to 1990, according to Robert S. Vrooman, a former C.I.A. officer who at the time directed counterintelligence at Los Alamos.
Top visitors included Dr. Agnew, the past director of the weapons lab; Danny B. Stillman, its head of intelligence; and George A. Keyworth 2d, a physicist who later became President Reagan's science adviser.
The benefits were judged to far outweigh the risks that arms scientists in informal settings and conversations might, by accident or design, give away secrets. And indeed, the Americans learned much.
''This was a huge intelligence game for the United States,'' said a United States official who spoke on the condition of anonymity. ''At the beginning we knew zip about China.''
One discovery was that parts of the Chinese program were quite advanced, including technologies for bomb development.
''They have excellent facilities, some better than ours,'' said Dr. Agnew, who in 1979 and 1982 was among the first visitors.
For instance, he said, the Chinese were able to peer into fiery blasts with an advanced camera known as pinex, revealing details to aid warhead development.
The American version of the device had one axis, he said, the Chinese version two, doubling its usefulness. ''It's much better,'' Dr. Agnew said.
The American visitors also learned much about what China lacked. From a barrage of inquiries over the years, it became clear that Beijing was eager to learn everything it could about shrinking the atomic trigger. The questions were regular, increasingly pointed and never answered, American officials said, insisting that Beijing got no secrets that way.
But in one case, investigators became suspicious about an American scientist at the Livermore weapons lab in California who in 1979 had talked with Chinese scientists.
The suspect, born in Taiwan, never confessed. But some Federal investigators, in an investigation code-named Tiger Trap, feared the scientist had compromised not only the design of the W-70, a neutron bomb, but the secret to making small atomic triggers.
Weapons experts say that the crucial insight of the watermelon shape can be communicated with a few comments, a hand motion or a simple drawing on the back of an envelope, although years of computing, calculation, experiment and factory labor are then needed to turn the idea into nuclear blasts.
''The real challenge is not in the design, it's in the manufacturing,'' said Houston T. Hawkins, head of international security studies at Los Alamos. For example, he said, plutonium, one of the most complex metals known to science, is difficult to cast because of its odd ways of reacting with other metals and materials. ''It's a strange beast,'' he said of the dense metal that fuels most atom bombs.
The Breakthrough
China Takes Giant Nuclear Step
China finally succeeded in exploding a miniaturized bomb on Sept. 25, 1992, American officials revealed. It took intelligence analysts more than two years to fully understand what China had accomplished, its feat becoming clear only after a Chinese nuclear expert who had been recruited to spy for the United States delivered an intriguing report to his American handlers.
The spy said that China's September test blast, initially viewed by American analysts as routine, was anything but. The bomb detonated that day was miniaturized with a core, the spy said, in the distinctive shape of an ovoid, indicating China had begun to master the art of making modern warheads.
In the mid-1990's, the task of tracking the technical ins and outs of other nations' nuclear programs fell to the national weapons labs. Among the sleuths was Dr. Robert M. Henson, an experienced weapons designer at Los Alamos who had been analyzing intelligence on foreign programs since 1988.
In January 1995, Dr. Henson said in an interview, he began looking more closely at how China had solved the miniaturization puzzle. For help he turned to Lawrence A. Booth, a friend who specialized in Russian analyses.
''We kept looking into it for two weeks,'' Dr. Henson recalled. ''Then, we decided to do something.''
They drew up their analysis and eventually took it to Mr. Trulock, who the previous year had become director of intelligence at the Energy Department, which oversees Los Alamos. Mr. Trulock, who has a bachelor's degree in political science and no formal technical training, said he wanted to bring in other nuclear experts, particularly ones who had long experience in developing the miniaturized nuclear triggers for hydrogen bombs. John L. Richter of Los Alamos, a scientist who filled that void, joined the team.
The group looked more closely at a clue provided by the Chinese spy, who described the size of the bomb's atomic core with an analogy to a common household object, officials said in a new disclosure. Working from that, the scientists calculated a more precise size and Dr. Henson and Dr. Richter went through the American stockpile of nuclear arms, looking up measurements to see if any matched.
The atomic trigger of the W-88, they discovered, was close enough in size to raise suspicions.
The Energy Department held meetings in which the Los Alamos team was joined by analysts from the C.I.A. and the Defense Intelligence Agency. Federal officials now say the intelligence agencies were skeptical, reasoning that too much was being made of a foreigner's rough analogy. But the Energy Department and the Los Alamos team felt the evidence was provocative.
The breakthrough came in 1995, as has been previously disclosed, when a Chinese Government official sent a package of secret Chinese documents to American officials.
Mr. Trulock said the most revealing document, dated 1988, laid out China's nuclear modernization plans for Beijing's First Ministry of Machine Building, which, among other things, made missiles and nose cones. It not only described China's plans but compared them to the nuclear arms of the American arsenal.
Relatively crude hand drawings sketched out the nose cones enveloping the W-88, the W-87, the W-78, the W-76, the W-62 and the W-56 -- warheads of the Trident, MX and Minuteman missiles -- and also gave their overall weights and dimensions.
In itself, these were not damning. Though still officially classified secret in some cases, such information by then was widely available in many unclassified American papers and articles.
But the Chinese document, some 20 pages in translation, went on to give sensitive data about the W-88, Federal officials revealed. It accurately described the shape of the atomic trigger as not spherical and said it was situated in the nose cone's narrow forward end -- an arrangement used in some but not all American warheads. And it correctly described the hydrogen fuel, or secondary, as having a spherical shape.
More unsettling to the team, it described the width of the casing that surrounds the atomic trigger to within a millimeter, or four-hundredths of an inch. ''That's pretty damn accurate,'' Mr. Trulock recalled.
A senior Federal official agreed. ''That opened eyes,'' he said. ''It seemed to confirm earlier assessments that had seemed insubstantial.''
Mr. Trulock said his team later found that the Chinese document gave a similarly exact measure for the width of the W-88's secondary, or hydrogen stage. ''Primaries are the long pole in the tent,'' he said, referring to the importance of the atomic trigger. ''But that measurement was as good as the one for the primary.''
The C.I.A. eventually concluded that the agent who sent the documents was acting under the instruction of Chinese intelligence. No one has ever come up with a persuasive explanation of why China sent the documents to American spies.
From 1992 to 1996, American officials revealed, China used its new atomic match to ignite a variety of hydrogen bombs, including one similar in some respects to the W-88. After this series of blasts shook the ground at the Lop Nur test site, China signed the Comprehensive Test Ban Treaty, signaling an end to its nuclear experimentation.
The Investigation
Federal Sleuths Hunt for a Spy
The Energy Department opened an investigation into the possible theft of W-88 secrets on Sept. 28, 1995, and over the next three years, Federal officials quietly tried to find out whether there was a Chinese spy in their midst.
If espionage occurred, Mr. Trulock and his team reasoned, it must have happened between 1984, when the warhead entered engineering development, and 1988, the date of the Chinese document.
Energy Department officials focused on Los Alamos, which had designed the bomb. They looked particularly closely at anyone who had traveled to China in those years or met visiting Chinese scientists.
Mr. Vrooman, then head of counterintelligence at Los Alamos and later a vocal critic of the inquiry, said investigators scrutinized only those people whose trips to China were paid for by the Energy Department.
Left unexamined, he said, were at least 15 additional people whose trips were paid for by the Chinese, the C.I.A., the Air Force or privately. These travelers tended to be top weapons designers and high officials -- the people who knew the most American arms secrets and had the most intimate contact with Chinese peers, Mr. Vrooman said.
In May 1996, the Energy Department turned over a list of a dozen suspects to the Federal Bureau of Investigation, which began a criminal case that eventually narrowed to Dr. Lee, an American scientist of Taiwanese birth working at Los Alamos.
Dr. Lee and his wife, Sylvia, had traveled to China in 1986 and 1988. Mrs. Lee was a secretary at Los Alamos who often met visiting Chinese delegations. And Dr. Lee, though a mechanical engineer by training and never a weapons designer, was familiar with the W-88 and many other nuclear arms and secrets (including the atomic trigger advance) because of his work on secret computer codes.
The F.B.I. believed it had enough evidence to seek a secret wiretap on Dr. Lee's phone calls, citing 20 reasons he was a prime suspect. But the Justice Department found the evidence unpersuasive and refused to seek a court order for the eavesdropping, a routine step in most spy cases.
Mr. Vrooman has charged that the inquiry was marred by a racist bias to target Chinese-Americans, an assertion Federal officials have vehemently denied. But the Republican chairman and the ranking Democrat of the Senate Governmental Affairs Committee, which investigated the spy case and heard testimony from Mr. Vrooman, concluded that Federal investigators had focused prematurely on Dr. Lee.
After the spy case became in public, Bill Richardson, the Secretary of Energy, recommended that Mr. Vrooman be disciplined for letting Dr. Lee have continuing access to secrets even after doubts about him had been raised.
Dr. Lee, fired this year from Los Alamos for security violations, including failing to report foreign contacts, has been charged with no crime and has denied any spying. After his ouster, investigators found that he had loaded many secret files onto an unsecured computer, raising the risk that they could have fallen into the wrong hands.
The inquiry most likely would not have come into public view had it not been for a series of unrelated disclosures about China.
In April 1998, The Times reported that two United States aerospace companies were under criminal investigation for providing rocket data to Chinese scientists.
A furor erupted in Congress. The House created a select committee, led by Mr. Cox, who had recently vied unsuccessfully for the House speakership, to look into whether the Administration's increasingly open policies on satellite exports had compromised national security.
There was no hint the committee would end up studying nuclear bombs. Composed of five Republicans and four Democrats, the committee did not learn of the suspected Chinese nuclear espionage until October 1998, just a few months before its mandate expired. On Nov. 12 and Dec. 16 it held secret hearings in which Mr. Trulock was called as the star witness.
In January, after three months of investigation, the committee completed a secret manuscript. In May, after a long argument with the White House over what could be made public, it released an 872-page report. The chapter on atomic espionage, just 37 pages, garnered most of the headlines.
In fiery prose accompanied by vivid color pictures and charts, the committee charged that Chinese spies had carried off vital secrets about seven of America's most advanced arms.
The People's Republic of China, it alleged, ''has stolen classified information on all of the United States' most advanced thermonuclear warheads,'' leaping from the clumsy designs of the 1950's to those that are far more modern and deadly.
The main evidence cited was the Chinese document obtained by the C.I.A. in 1995 and an inquiry in the 1980's into spying at the Livermore lab that concluded China had most likely obtained design secrets of the neutron bomb. The unclassified version of the committee's report gave no details of the 1995 document's secret details about the W-88.
The report was signed by the committee's four Democrats. But immediately after its release, Representative John M. Spratt Jr. of South Carolina, one of the Democrats, criticized it as rushed, superficial and exaggerated. The witnesses heard by the committee, he added, ''did not have the technical background to fully assess the nature or value of the information lost.''
The Debate
Analysts Sift For the Truth
Since then, Mr. Spratt's critique has been echoed and amplified by a range of top scientists and bomb designers who say Beijing could have miniaturized its warheads on its own without spying.
Richard L. Garwin, a physicist who has long advised Washington on nuclear arms, recently on a bipartisan team led by former Defense Secretary Donald H. Rumsfeld, said, ''there is no reason to believe that China could not have built perfectly adequate warheads'' for a range of modern missiles ''from nuclear technology that it developed itself.''
China, several officials said, simply went down the same path as other nuclear powers, helped along by the general knowledge of what the United States had achieved: proof that hydrogen bombs can be made very small but nonetheless very powerful.
''Every state has come to it,'' said one Federal official, referring to breakthroughs in atomic triggers by the Soviet Union, Britain and France. ''Now they've got it too.''
Mr. Hawkins, the head of international security studies at Los Alamos, which is clearly on the defensive because of the spy scandal, said the basic physics of bombs and missiles push weapons designers in roughly the same direction. To obtain the best performance, he said, engineers are invariably led toward narrow nose cones about 16 degrees wide -- if cut from a pie, a very modest slice.
''Once you realize that,'' Mr. Hawkins said, ''it drives every nation down similar paths. Eventually, all weapons systems will look alike. It has to do more with physics than espionage.''
That view is not universally accepted.
Dr. Henson, the analyst who first sounded the alarm at Los Alamos, said there was nothing in the design of missile nose cones that propelled a scientist to shape the core of an atomic trigger into an oval.
Do scientific and technical analyses automatically ''draw you to a watermelon?'' he asked, alluding to the shape of the top-secret design. ''That's not true.''
''It's beyond a shadow of a doubt,'' Dr. Henson added. ''Major espionage took place.''
American intelligence agencies are less categorical. Analysts have concluded that espionage played a role in Beijing's advance, but cannot identify a hard link comparable to the Soviet Union's theft in the 1940's of the American design for the first atom bomb.
''Everybody has come to the same conclusion,'' said a top Administration official who has closely scrutinized the secret data. ''We don't have a smoking gun.''
A Federal intelligence study done last year, which the Cox committee drew on, said American secrets lost between 1984 and 1988 let the Chinese ''accelerate their nuclear weapons program well beyond indigenous capabilities,'' a view that echoed the original Los Alamos finding.
A damage assessment by the American intelligence community, made public in April, said a mix of espionage, openly available data and scientific acumen had greatly lengthened Chinese strides. Stolen secrets, it said, ''could help'' Beijing develop a mobile missile and ''probably accelerated its program to develop future nuclear weapons.''
In June, the President's Foreign Intelligence Advisory Board, which did its own investigation, said both Congressional and Administration leaders had engaged in ''simplification and hyperbole'' in the spy case. Neither dramatic damage assessments nor categorical reassurances, it said, were wholly substantiated.
And Mr. Hawkins, the Los Alamos official, said the specific secrets known to have been seized by the Chinese, principally those detailed in the 1995 document, would have been little help to a bomb maker, and far from Mr. Trulock's road map. As for an H-bomb's innards, what designers call the physics package, Mr. Hawkins said the documents ''describe nothing significant.''
Mr. Cox insisted that highly classified intelligence data available to his committee showed a more persuasive case than has emerged publicly. ''There are more interpolating facts'' that closely tie lost W-88 secrets to Beijing's advance, he said.
But a Federal official cited intelligence data about China's atomic trigger showing it to be anything but an exact copy.
''It turns out the W-88's is slightly smaller,'' said the official, who believes Beijing may have made the advance on its own.
It remains unresolved how China got the W-88 secrets in the first place, but a consensus is emerging that the search for the leak narrowed too quickly to Los Alamos.
Studies by the Senate as well as the President's foreign intelligence board this year raised serious questions about whether the F.B.I. and Energy Department had too quickly focused on the weapons lab. No evidence has pinpointed it as the leak's source.
Mr. Vrooman, the head of counterintelligence at the laboratory from 1987 until 1998, noted that one secret document describing the design of the W-88 warhead went to 548 mailing addresses throughout the Government and military. Some Administration experts believe the data described by the Chinese in the 1995 document came from engineering plans or from secret manuals on military bases.
''That kind of information was widely available,'' said Dr. Drell of Stanford, who served on the President's advisory board investigation. ''The manuals that went out had pictures and numbers. If a submarine came in, and there was a problem, they had to know what they were dealing with.''
However Beijing made its miniaturization advance -- on its own, by theft or a combination of the two -- it is apparently proud enough to boast about it publicly, at least among its friends in the mountains of New Mexico. Dr. Henson said a Chinese arms scientist, Sun Cheng Wei, bragged of the breakthrough at Los Alamos a few years ago, telling an open seminar that China had forged significantly ahead in nuclear arms.
''What he said,'' recalled Dr. Henson, who attended the talk, ''was that for a long time they were dealing only with round designs, and then only watermelons.''
Photos: The world's first nuclear explosive device, above, was code-named the Gadget. Dr. Norris E. Bradbury, director of Los Alamos National Laboratory from 1945 to 1970, posed with it in 1945. Below, the first hydrogen bomb, called Mike, in 1952. (Jornada Del Muerto Valley, New Mexico, July 1945); (Los Alamos National Laboratory)(pg. A14); The United States' most advanced nuclear warhead is the W-88, at least eight of which can fit on the Trident D-5 missile. (Courtesy of Federation of American Scientists)(pg. A15) Diagrams: Some secret specifications of the W-88, an American miniature hydrogen bomb, that were found in a Chinese document. SHAPE OF ATOMIC BOMB TRIGGER Described as not spherical SHAPE OF HYDROGEN BOMB FUEL -- Described as spherical SIZES OF CASINGS BOMB PLACEMENT -- Atomic bomb trigger is placed above the hydrogen bomb fuel. (Illustration by Mika Grondahl)(pg. A1) ''Less but More'' Early bombs were behemoths compared with modern miniature devices. The W-88, the U.S. arsenal's most modern warhead, though tiny, is dozens of times more powerful than the bomb that leveled Hiroshima. A-BOMBS Little Boy -- 1940's Dropped on Hiroshima Fat Man -- 1940's Dropped on Nagasaki H-BOMBS B-17 -- 1950's Largest bomb deployed by U.S. W-88 -- 1980's As many as eight fit atop Trident II (D-5) missiles (Sources: Federation of American Scientists; The Cox Report)(pg. A14) ''A Missile That Carries Miniature Bombs'' The miniaturization of the W-88 hydrogen warhead, considered the most advanced in the American arsenal, made the Trident submarine missile one of the most deadly weapons of all time. Trident II (D-5) missile W-88 WARHEADS The Trident can carry eight of these, each of which can seek a different target. THIRD STAGE MOTOR Accelerates the warhead to approximately 15,000 m.p.h. SECOND STAGE MOTOR FIRST STAGE MOTOR Ignites after missile clears the water's surface. Missile is fired from a submarine. (Source: Federation of American Scientists)(pg. A15)
中国在1992年9月25日取得成功，这个消息来自一名间谍，他对美国的接头人说，北京爆炸了一颗以小型化秘密技术为基础的炸弹。
新墨西哥州洛斯阿拉莫斯武器实验所一组科学家开始进行一项有重大影响的侦探调查：中国的进步是间谍活动的结果，还是自己艰苦劳动的结果，还是两者的结合?
这场辩论还在激烈地进行着。专家们一致认为出现了间谍活动，但是对多少秘密被窃和它对北京的进步有多大作用分歧很大。
洛斯阿拉莫斯的小组在1995年作出了结论，说中国的进步大概是以间谍活动为基础的。1999年国会一个委员会的报告走得更远，声称“没有从美国盗窃的核机密”，中国“几乎不可能”造出小型弹头。
国会的报告遭到政府内外科学家的批评，他们说报告夸大了间谍活动的作用，中国很可能靠自己的力量取得突破。
一项以几个月的采访以及透露的武器和情报秘密为基础的研究表明，国会的报告声称盗窃的机密是中国取得突破的主要原因没有证据。
这项研究支持争论不休的专家逐渐形成的一致看法：联邦调查立早地集中于洛斯阿拉莫斯国家实验所和那里的一名工作人员(李文和)。现在看来，失窃的机密全国与武器有关的单位数百人、甚至数干人都能接触到。
对洛斯阿拉莫斯侦探小组来说，整个间谍理论在1995年得到有力支持，当时中央情报局得到一份中国内部文件，其中有对美国最先进的小型弹头W-88的说明。联邦官员在首次透露他们在此案中最重要的证据即文件的秘密内容时说，中国的文件列出了这种弹头的五种关键的特点，包括精确到4/100英寸的两种测量法。
但是批评者坚持说，即使北京进行了间谍活动，小型化的突破也是靠它自己的力量，它进行这项工作至少有13年(从1979年到1992年)。
美国专家说，中国科学家的才能从他们制造的一种相机中可见一斑。这种相机拍摄的核爆炸的照片比美国类似的相机拍出的照片好得多。
前洛斯阿拉莫斯的负责人霍华德・阿格纽说：“他们不需要我们的任何帮助。他们只是好奇，就像我们对他们好奇一样。”阿格纽访问过中国，是联邦情报顾问。
解析间谍活动造成的损失是一门涉及到推理、求证和作出结论的不精确的艺术。在已知和怀疑之间的空白常常插进不少个人、党派或者机构的偏见。
同在多数间谍案中一样，证据可以作多种解释。一些了解中央情报局获得的中国文件的批评者说，文件对美国弹头的描述本身不足以制造小型化的弹头。
能源部主管洛斯阿拉莫斯调查的官员诺特拉・特鲁洛克同意这种估计，但是说这些情报是秘密的，从来没有在任何公开文件或者因特网上公布过。他和他手下的人分析，任何得到这些情报的人必定能够得到多得多的弹头设计秘密。
特鲁洛克在接受采访时说，知道各部件近似的体积和形状可以为中国核弹制造者提供一张线路图，也许可以使他们省去多年初步试验的时间。
然而特鲁洛克接着说，国会委员会的报告太绝对。报告部分是以他的证词为基础的。
他说：“我在作证时为防止误解而作了适当的解释。说明对我们的证据和结论是不肯定的。我们一般说：‘也许是这样，也许是那样。”’他说国会委员会对间谍活动在中国突破中的作用“作出判断”。
中国迟迟才加入核俱乐部，而在加人时却表现出相当高的技巧。
北京在1964年爆炸了第一颗原子弹。像美国投到日本广岛的原子弹一样，这种铀弹设计巧妙，不仅节约了价格不菲的燃料，而且使原子弹更轻巧，从而增加了它的军事价值。
北京仅在32个月后就制造出第一颗氢弹。
华盛顿负责监督核武器的私人机构自然资源保护委员会的罗伯特・诺里斯说，相比之下，从核武器发展到热核武器，伦敦花了历个月，莫斯科花了75个月，华盛顿花了解个月，巴黎花了103个月。
中国仅试爆6次就达到了热核武器的程度，而美国进行了31次。中国试爆的次数一向很少。尽管北京研制了至少6种类型的武器，但它几十年来进行的核试验相对而言较少，总共45次，美国则有1030次。
在理查德・尼克松总统1972年访华之前。美国情报机构对中国的核计划和现代化计划知之甚少。但在尼克松外交行动打开了大门之后，美中两国就建立了军事关系。
到了1979年，美国的核武器设计者和安全问题专家纷纷开始造访中国同行、武器实验所以及位于中国西部沙漠的罗布泊。中国的原型核武器就是在那里爆炸的。
据前中央情报局官员罗伯特・弗鲁曼说，从1979年至1990年，光是从格斯阿拉莫斯去中国访问的科学家和官员至少有85名。弗鲁曼当时负责洛斯阿拉莫斯的反情报工作。
高级访问者包括该武器实验所已故所长阿格纽；负责情报工作的丹尼・斯蒂尔曼；后来担任里根总统的科学顾问的物理学家乔治・基沃思。
根据判断，访问带来的好处远远超过武器专家在非正式环境或谈话中可能有意无意泄露机密带来的风险。其实，美国人了解到很多情况。
一个发现就是，中国计划的一部分是十分先进的，其中包括原子弹研制技术。
阿格纽说：“他们拥有精良的装置，其中一些强于我们。”阿格纽是在1979年和1982年首批访华者之一。
阿格纽说，例如，中国人能够利用一种先进的照相机观察烈焰滚滚的爆炸，从而把观察到的细节用于弹头研制工作。阿格纽说，美国的设备只有一个轴，而中国有两个，其用途自然也多出一倍。他说：“(中国的设备)要好得多。”
美国访问者还了解到中国的不足之处。根据多年来接二连三的调查结果，北京显然急于尽一切可能掌握有关如何缩小原子弹引爆装置的技术。美国官员说，这些是普遍问题。它们越来越突出，从未有过答案。他们坚持认为，北京在这个方面没有秘密。
美国官员们透露，中国最终在1992年9月25日成功地爆炸了一颗小型化的原子弹。情报机构分析家们花了两年多时间才完全弄明白中国取得了什么成就。在一位受雇为美国刺探情报的中国核专家向美国上司递交了一份报告之后，美国才了解到中国取得的这一成就。
这个间谍说，美国分析家最初认为中国在9月份的试爆是常规试爆，但实际情况决非如此。这位间谍说，当日试爆的这颗原子弹已被小型化，只有一个核心，形状为特殊的卵形，这表明中国已开始掌握现代化弹头的制造技术。
在90年代中期，追踪其它国家核计划底细的任务落到国家武器实验所头上。刺探者之一就是洛斯阿拉莫斯经验丰富的武器设计者罗伯特・亨森。自从1988年以来，亨森一直在分析有关外国计划的情报。
亨森在接受记者采访时说，1995年元月，他开始更加密切关注中国如何解决小型化的难题。为此，他求助于专门分析俄罗斯武器计划的劳伦斯・布思。
他们起草了一份分析报告，最后交给了特鲁洛克，此人一年前担任管理洛斯阿拉莫斯的能源部情报部主任。特鲁洛克获得过政治学学士学位，没有受过任何正规的技术训练。他说，他希望邀请其他一些核专家参加，尤其是那些长期从事氢弹的小型核引爆装置研究的专家。洛斯阿拉莫斯的约翰・L．里克特这位科学家正合需要，于是加入了这个小组。
官员们在一份刚刚公开的材料中说，这个小组仔仔细细地研究了那名中国间谍提供的线索，这名间谍说，这种炸弹的原子起动装置核心的大小类似一种寻常家用物品。从这个问题开始人手，科学家们对尺寸大小进行了更加精确的推算，亨森博士和里克特博士查遍了美国的核武器储备，看看有没有尺寸大小与此相符。
他们发现，W-88的原子弹起动装置在尺寸大小上非常接近，足以引起人们的怀疑。
能源部举行了多次会议，中央情报局和国防情报局的分析家们也同洛斯阿拉莫斯小组一道出席了会议。联邦官员们现在说，这些情报机构表示怀疑，它们认为一个外国人的粗略类比受到了过分的重视，但是，能源部和洛斯阿拉莫斯小组觉得这种证据令人振奋。
正如以前披露的那样，1995年出现了突破性的进展，当时，一名中国政府官员给美国官员寄来了一包秘密的中国文件。
特鲁洛克说，最能透露内情的一份文件注明日期是1988年，这份文件是为生产导弹和头锥的第一机械工业部制订的核现代化计划。这份文件不仅描述了中国的计划，还拿它们同美国武库中的核武器进行比较。
据联邦官员们透露，中国功这份文件翻译后大约长达20页，这份文件还提供了有关W-88的敏感资料。它精确地描述了这种初级装置的形状，说它不是圆形，并说它位于头锥狭窄的前部靠后的位置，美国的有些弹头也是这样安排，但不是全部。这份文件还正确地描述这种次级装置是一种圆形的形状。更让这个小组困惑的是，这份文件还描述了原子起动装置包装盒子的宽度，精确到了毫米的程度，即相当于一英寸的百分之四。特鲁洛克回忆说：“真是精确极了。”
中央情报局最终断定，寄来这些文件的那名特工一定受到了中国情报部门的指使。关于中国为何给美国间谍寄送这些文件，到现在为止，还没人提出一个更有说服力的解释。
能源部于1995年9月28日对可能存在的盗窃W-88机密的事件展开了调查，在过去的3年里，联邦官员们不动声色，试图查明他们中间是否有人就是中国的间谍。
特鲁洛克和他的小组分析道，如果存在间谍活动，那么，这种间谍活动肯定发生在1984年和1988年之间。1984年，这种弹头进入设计研制阶段；而中国这份材料的注明日期是1988年。
能源部的官员们把注意力集中到设计这种炸弹的洛斯阿拉莫斯的身上。他们尤其仔细地审查了这些年中每位去过中国或者接待过到访的中国科学家的人。
1996年5月，能源部向联邦调查局递交了一份12名受到怀疑的人的名单。然后联邦调查局进行了刑事调查，最终日标缩小到了李博士的身上。李是一名美国科学家，出生在台湾，在洛斯阿拉莫斯实验室工作。
弗鲁曼是当时洛斯阿拉莫斯实验室反情报部门的主任，他后来对这种调查提出了直言不讳的批评。他指责这种调查受到了种族偏见的站污，它的矛头是美籍华人。联邦官员们极力否认了这种指责。但是，参议院政府事务委员会共和党主席和民主党首席委员一致认为，联邦调查人员过早地把注意力集中到了李博士的身上。参议院政府事务委员会负责这一间谍案的调查，并且听取了弗鲁曼的证词。
如果不是一系列与此无关的有关中国问题的报道，这场调查很有可能不会引起公众的注意。1998年4月，《纽约时报》报道了两家美国航天工业公司因向中国科学家提供火箭资料而受到刑事调查的事情。
这个问题在国会引起了轰动。众议院成立了一个由考克斯牵头的特别委员会，负责调查政府在卫星出口方面日益开放的政策是否损害了国家安全。
当时，没有任何迹象表明这个委员会最终会调查到核炸弹问题上来。这个由5名共和党人和4名民主党人组成的委员会没有听说中国涉嫌从事核间谍活动这件事情，直到1998年10月，此时距该委员会使命结束仅仅还有几个月的时间。在11月12日和12月16日，该委员会举行了听证会，特鲁洛克作为主要证人被传到会作证。
1月，经过5个月的调查，该委员会完成了一份秘密报告。5月，在就什么东西可以公开的问题同白宫经过一番长时间的争论之后，该委员会公开了一份长达872页的报告。关于原子技术谍报活动的章节只有37页，但是各家媒体对此问题报道的最多。
该委员会的4名民主党人在这份报告上签了字。但是，就在这份报告公开之后不久，其中一名民主党人、来自南卡罗来纳州的约翰・斯普拉特众议员对这份报告提出了批评，说这是一份匆忙赶制的报告，草率肤浅，夸大事实。他又说，该委员会查访的一些证人“没有那种技术背景，可以充分评估丢失的那些情报的性质和价值”。
自那时以来，斯普拉特的批评意见引起了许多著名科学家和炸弹设计人员的共鸣，而且还有所增大。他们说不用进行间谍活动，北京可能已经自力更生实现了自己弹头的小型化。长期以来一直替华盛顿在核武器问题上出谋划策的物理学家理查德・加温最近参加了前国防部长唐纳德・拉姆斯菲尔德牵头的一个两党小组。他说： “没有理由相信中国不能依靠自己研制的核技术替一系列现代导弹生产完全合适的弹头。”
几名官员说，中国所走的道路与其他核大国完全一样，对美国所取得的成果的大致了解让它们从中获得了帮助，证据是：氢弹可能造得很小，但是仍然具有很大的威力。
洛斯阿拉莫斯实验室国际安全研究部门的负责人──因间谍案显然处于守势──霍金斯说，炸弹和导弹的基本物理特性促使武器设计人员朝着总体一致的方向前进。他说，为了取得最佳性能，工程技术人员不可避免地都会进行大约16度宽的狭窄头锥的研究──如果是一块蛋糕，这种也是非常小的一片。霍金斯说：“一旦认识到这个问题，它就会促使每个国家沿着相似的道路前进。最后，所有的武器系统看上去都差不多。这更多地关系到物理特性，而不是间谍活动。”
这种观点得到了普遍接受。
但首先在洛斯阿拉莫斯拉响警报的分析家亨森博士说，在导弹头锥设计方面，没有任何东西促使一名科学家把原子起动装置的核心制造成椭圆的形状。他说：“毫无疑问，严重的间谍活动肯定发生过。”
美国情报机构的态度则没有那么坚决。分析家们得出结论认为，间谍案在北京的改进方面确实起了作用，但是，他们却苦于找不到一个像在本世纪四十年代苏联盗取美国第一颗原子弹设计方案那样的硬性联系。
一位曾认真审查过这份秘密材料的政府高级官员说：“大家都得出了同样的结论。但我们没有确切的证据。”
美国情报机构曾对这起间谍案作过一份损失评估报告，并于1999年4月公布于世。这份评估报告说，间谍案、公开可以得到的资料和科学的敏锐等因素加起来使中国迈出的步伐更大了。报告说，这些“被盗的机密”“可能帮助”北京研制出一种移动式导弹并且“可能加速了研制未来的核武器的计划”。
6月份，也在独自进行调查的总统国外情报顾问委员会说，国会和政府领导人都参与了对这起间谍案的“简单化和夸大化”。该委员会说，无论是重大损失评估还是断然的否认，都没有充分的事实依据。
洛斯阿拉莫斯核试验和核技术实验室的官员霍金斯说，已确知被中国人盗走的具体机密，主要是在1995年文件中详细列举的那些，对于导弹制造者来说没有什么助益，而且距特鲁洛克的公路图相距甚远。至于被设计者称之为物理包的氢弹内部，霍金斯说，这些文件“没有介绍任何重要的内容”。
一位联邦官员用有关中国原子弹的情报数据来说明，这根本不是一个真正的翻版。
这位官员说：“事实证明，W-88要稍小些。”他认为，北京可能自行作了改进。
中国是如何从第一地点获得W-88的机密的，这个问题至今仍然是个谜，但是，人们已经普遍开始认为，把泄秘事件的范围一下子缩小到洛斯阿拉莫斯实验室实在是太快了。
参议院和总统的国外情报顾问委员会1999年进行的研究都提出了严肃的问题，即联邦调查局和能源部把目标对推这家武器实验室是否太快了。没有任何证据把该实验室作为泄秘事件的源头。
从1987年到1998年一直担任该实验室反间谍机构负责人的弗鲁曼注意到，描述W-88弹头设计的一份机密文件通过政府和军方发给了548个邮件地址。一些政府专家相信，中国人在1995年的文件中描述的资料取自于工程计划或军事基地的秘密手册。
曾参与总统顾问委员会的调查工作的斯坦福大学德雷尔博土说：“那种情报比比皆是，发行的手册既有照片也有数字。”
(美国《纽约时报》报道)(载于《参考消息》1995年9月15日15、16版)
=======
September 7, 1999
Spies vs. Sweat: The Debate Over China's Nuclear Advance
By WILLIAM J. BROAD
When American bomb makers began visiting China in 1979, they were startled by increasingly pointed questions that suggested their Chinese peers were hot on the trail of the secret to building a modern nuclear arsenal. It allows hydrogen bombs to be made so small that many can fit atop a single missile or be fired from trucks, submarines and other mobile platforms.
China succeeded on Sept. 25, 1992, the news coming from a spy who told his American handlers that Beijing had exploded a bomb based on the miniaturization secret.
A team of scientists at the Los Alamos weapons laboratory in New Mexico set to work on a whodunit with huge implications: Was China's advance the result of espionage, hard work or some mix of the two?
Today, the debate rages on. Experts agree that spying occurred, but clash violently on how much was stolen and what impact it had on Beijing's advance, if any.
The Los Alamos team concluded in 1995 that China's stride was probably based on espionage. A report this year by a Congressional committee that made the case public went further, claiming that it would have been ''virtually impossible'' for China to have made small warheads ''without the nuclear secrets stolen from the United States.''
The Congressional report unleashed criticism from scientists inside and outside the Government who said the importance of the espionage was overstated, and that China could well have achieved the breakthrough on its own, as it insists publicly.
A review of the dispute, based on months of interviews and disclosures of weapons and intelligence secrets, suggests that the Congressional report went beyond the evidence in asserting that stolen secrets were the main reason for China's breakthrough.
The review also bolsters a point of emerging agreement among feuding experts: that the Federal investigation focused too soon on the Los Alamos National Laboratory and one worker there, Wen Ho Lee, who was fired for security violations. The lost secrets, it now appears, were available to hundreds and perhaps thousands of individuals scattered throughout the nation's arms complex.
Federal officials asked in recent days that some details about weapons design and intelligence sources not be published, and The New York Times agreed to withhold them.
For the Los Alamos team of detectives, the overall spy theory was supported strongly in 1995 when the Central Intelligence Agency obtained an internal Chinese document that included a description of the United States' most advanced miniature warhead, the W-88. Revealing for the first time their top evidence in the case, the document's secret contents, Federal officials say the Chinese text cited five key attributes of the warhead, including two measurements accurate to within four-hundredths of an inch.
But the critics, who are also revealing new information, insist that Beijing, even if it spied, made the miniaturization breakthrough on its own, pursuing it for at least 13 years, from 1979 to 1992.
The prowess of Chinese scientists, American experts said, is suggested by a camera they built for photographing nuclear blasts, which was far better than a similar one made by the United States.
''They don't need any help from us,'' said Harold Agnew, a past Los Alamos director, visitor to China and Federal intelligence adviser. ''They're just curious, as we are curious about them.''
Deconstructing the damage wrought by espionage is an imprecise art that mixes inference, evidence and deduction. In the vacuum between what is known and what is suspected, personal, partisan or institutional bias often rushes in.
The debate over Chinese spying has been blurred by issues that include Republican distaste for President Clinton's China policy, accusations of racial bias in the investigation and fears among scientists that the uproar is prompting security measures so tight as to damage work, morale and recruitment.
As in most spy cases, the evidence is open to interpretation. Several critics familiar with the Chinese document obtained by the C.I.A. said that its description of the American warhead was not by itself sufficient to build a miniaturized warhead.
The Energy Department official who supervised the Los Alamos inquiry, Notra Trulock, agreed with this assessment but said the information was secret and had never been mentioned in any public document or Internet posting. Anyone who had it, he and his team reasoned, must have also obtained access to a much broader range of secrets about the warhead's design.
In addition, Mr. Trulock said in an interview, knowing the approximate size and shape of the components provided a road map to Chinese bomb makers, probably allowing them to skip years of preliminary testing.
Mr. Trulock added, however, that the Congressional committee was too categorical in its report, which was based in part on his testimony.
''When I testified, I used the appropriate caveats to express uncertainties in our evidence and our conclusions,'' said Mr. Trulock, formerly the Energy Department's intelligence chief. ''We typically said: 'Probably this. Probably that.' '' The committee, he said, ''made judgments'' about the centrality of spying in China's breakthrough.
Representative Christopher Cox, a California Republican who was chairman of the committee, defended the work of his staff of 47, which included no one with nuclear design experience. The panel, he said in a lengthy interview, drew largely on Clinton Administration witnesses for its expertise. The conclusion that espionage allowed Beijing to skip decades of research, he said, was an appropriate one, based on the Government's own evidence.
''Judgment matters,'' he said, responding to Mr. Trulock's criticism. ''We don't know everything to a certainty. The question is what is more likely than not.''
In the interview, Mr. Cox expressed surprise when told of the depth and breadth of China's interest in the miniaturization secret. He also played down the idea, cited by Federal skeptics of Chinese spying, that most of the world's nuclear powers have figured out the secret of miniaturization.
Can China, Mr. Cox asked, ''develop it indigenously because France did? That is a stretch. It's almost apples and oranges.''
The Secret
America Shrinks An Atomic Match
From the dawn of the nuclear age, miniaturization has been an obsession of weapons designers. The world's first atomic bomb, designed by the Los Alamos laboratory and detonated in the New Mexico desert in July 1945, was an awesome but cumbersome affair. A lump of plutonium the size of a softball was surrounded by a much larger ball of high explosives that was five feet wide and made up of 32 explosive charges and 64 detonators. Big as a car, it could not have fit into a small airplane, let alone a missile.
In 1952, American physicists made an important breakthrough: the H-bomb. Roughly a thousand times more powerful than the first atomic weapon, the hydrogen bomb was a two-stage device. Inside its dense casing, an atomic explosion -- called the primary -- worked as a match to kindle an even more powerful detonation by the bomb's hydrogen fuel, which was known as the secondary.
Size was an issue from the start. The first hydrogen bomb stood two stories high and weighed 82 tons. It would be militarily useful only if it could be shrunk, and over the next few years, the country's best physicists set out to do just that. After considerable trial and error, they figured out that they could obtain the same kind of explosive power from a smaller package. A main breakthrough centered on the large, heavy atomic match. By shaping its plutonium fuel into an ovoid, roughly like a watermelon, scientists were able to drastically shrink the size and number of the explosives that triggered the nuclear blast.
After at least one flop, the radical idea roared to life in July 1957 in a nuclear explosion in the Nevada desert, according to Chuck Hansen, author of a detailed history of America's early nuclear efforts. It had taken the United States a little more than five years to move from the first hydrogen bomb to its miniaturized cousin.
The development had profound implications for the cold war's nuclear competition.
Shrinking the atomic trigger from something roughly the size of a washing machine to something smaller than a football allowed weapons designers to put thermonuclear arms atop small missiles that could be launched from submarines or mobile platforms like trucks. Arms would no longer be confined to bombers or silos in the ground.
The advance meant weapons could now be carried, quite stealthily, closer to enemy shores and could be made safer from attack. It also meant warheads could fit into the cramped spaces of narrow nose cones, which streaked faster to Earth than blunter shapes and were less buffeted by winds during the fiery plunge, making them more accurate.
The first warhead in the new generation of weapons, the W-47, was less than half the size of the bomb that leveled Hiroshima but up to 80 times more powerful. In 1960, when the first Polaris submarine put to sea, each of its 16 missiles was armed with a W-47.
The weapons continued to evolve, and by all accounts, the apex was reached in the 1980's with the W-88, one of the most deadly weapons in the American arsenal.
The warhead, made for submarines, first went to sea a decade ago and is considered quite powerful for its small size. The precise size is secret. But at least eight W-88's can fit atop the Trident D-5 missile, which is less than seven feet wide. Since Trident subs have 24 missiles, a single submarine can carry up to 192 of the thermonuclear arms.
Today, American submarines on patrol in the Atlantic carry the small warheads. And the Navy is adding them to its Pacific fleet, so in the next few years the W-88 is likely to be aimed at China.
The Chinese
Late to Start, Quick to Excel
China was late in joining the nuclear club, but showed considerable skill when it did.
Beijing detonated its first bomb in 1964. The tricky design was based on uranium, like the Hiroshima bomb, but saved costly fuel and made the bomb lighter, increasing its military value.
Sidney D. Drell, a Stanford physicist and Clinton Administration adviser, writing in ''China Builds the Bomb'' (Stanford University Press, 1988), called the feat ''enormously impressive.'' Beijing's first hydrogen bomb came just 32 months later.
By comparison, the step from nuclear to thermonuclear took London 66 months, Moscow 75 months, Washington 87 months and Paris 103 months, said Robert S. Norris of the Natural Resources Defense Council, a private group in Washington that monitors nuclear arms.
China set off just 6 test blasts to get to the H-bomb stage, versus 31 for the United States. The low number was typical. While developing at least six types of weapons, Beijing over the decades conducted relatively few nuclear tests, 45 in all, versus 1,030 for the United States.
The evidence strongly suggests that China, in its first phases of missile building, had no idea how to shrink thermonuclear arms. According to ''China's Strategic Seapower'' (Stanford University Press, 1994), the warhead for the submarine missile deployed by Beijing in the 1970's weighed 1,300 pounds, more than twice the old American W-47, suggesting that the Chinese were still using a spherical atomic match to ignite hydrogen bombs.
China's land force was modest. Starting in the 1980's, it deployed about 20 missiles that can now reach anywhere in North America, each topped by a single warhead that can unleash a force equivalent to up to five million tons of high explosives. That is about 300 times stronger than the Hiroshima bomb.
The big warheads are not particularly accurate, but they fit China's professed war doctrine -- to fire nuclear arms only in retaliation. The big missiles can, if necessary, hit a city.
China's interest in building smaller weapons was spurred, in part, by the United States' development in the late 1970's of a high-accuracy design known as the Missile Experimental, or MX, that bristled with 10 warheads. Though meant primarily to unnerve Moscow, the weapon also worried Beijing, which quickly grasped that its handful of big land-based missiles looked like sitting ducks that could be destroyed in a first strike of precisely aimed hydrogen bombs.
Beijing's unease grew as the American Navy in the late 1970's unveiled plans for a new submarine-launched missile nearly as unerring as the MX and bearing an even more powerful warhead -- the W-88.
American intelligence agencies knew little about China's nuclear program and modernization plans, if any, before President Richard Nixon's visit to China in 1972. But the military ties that followed the Nixon diplomatic initiative opened the door.
By 1979, American nuclear arms designers and security experts were starting to visit their Chinese peers, weapons labs and Lop Nur, the sprawling site in China's western desert where prototype nuclear weapons were detonated.
From Los Alamos alone, at least 85 scientists and officials made trips from 1979 to 1990, according to Robert S. Vrooman, a former C.I.A. officer who at the time directed counterintelligence at Los Alamos.
Top visitors included Dr. Agnew, the past director of the weapons lab; Danny B. Stillman, its head of intelligence; and George A. Keyworth 2d, a physicist who later became President Reagan's science adviser.
The benefits were judged to far outweigh the risks that arms scientists in informal settings and conversations might, by accident or design, give away secrets. And indeed, the Americans learned much.
''This was a huge intelligence game for the United States,'' said a United States official who spoke on the condition of anonymity. ''At the beginning we knew zip about China.''
One discovery was that parts of the Chinese program were quite advanced, including technologies for bomb development.
''They have excellent facilities, some better than ours,'' said Dr. Agnew, who in 1979 and 1982 was among the first visitors.
For instance, he said, the Chinese were able to peer into fiery blasts with an advanced camera known as pinex, revealing details to aid warhead development.
The American version of the device had one axis, he said, the Chinese version two, doubling its usefulness. ''It's much better,'' Dr. Agnew said.
The American visitors also learned much about what China lacked. From a barrage of inquiries over the years, it became clear that Beijing was eager to learn everything it could about shrinking the atomic trigger. The questions were regular, increasingly pointed and never answered, American officials said, insisting that Beijing got no secrets that way.
But in one case, investigators became suspicious about an American scientist at the Livermore weapons lab in California who in 1979 had talked with Chinese scientists.
The suspect, born in Taiwan, never confessed. But some Federal investigators, in an investigation code-named Tiger Trap, feared the scientist had compromised not only the design of the W-70, a neutron bomb, but the secret to making small atomic triggers.
Weapons experts say that the crucial insight of the watermelon shape can be communicated with a few comments, a hand motion or a simple drawing on the back of an envelope, although years of computing, calculation, experiment and factory labor are then needed to turn the idea into nuclear blasts.
''The real challenge is not in the design, it's in the manufacturing,'' said Houston T. Hawkins, head of international security studies at Los Alamos. For example, he said, plutonium, one of the most complex metals known to science, is difficult to cast because of its odd ways of reacting with other metals and materials. ''It's a strange beast,'' he said of the dense metal that fuels most atom bombs.
The Breakthrough
China Takes Giant Nuclear Step
China finally succeeded in exploding a miniaturized bomb on Sept. 25, 1992, American officials revealed. It took intelligence analysts more than two years to fully understand what China had accomplished, its feat becoming clear only after a Chinese nuclear expert who had been recruited to spy for the United States delivered an intriguing report to his American handlers.
The spy said that China's September test blast, initially viewed by American analysts as routine, was anything but. The bomb detonated that day was miniaturized with a core, the spy said, in the distinctive shape of an ovoid, indicating China had begun to master the art of making modern warheads.
In the mid-1990's, the task of tracking the technical ins and outs of other nations' nuclear programs fell to the national weapons labs. Among the sleuths was Dr. Robert M. Henson, an experienced weapons designer at Los Alamos who had been analyzing intelligence on foreign programs since 1988.
In January 1995, Dr. Henson said in an interview, he began looking more closely at how China had solved the miniaturization puzzle. For help he turned to Lawrence A. Booth, a friend who specialized in Russian analyses.
''We kept looking into it for two weeks,'' Dr. Henson recalled. ''Then, we decided to do something.''
They drew up their analysis and eventually took it to Mr. Trulock, who the previous year had become director of intelligence at the Energy Department, which oversees Los Alamos. Mr. Trulock, who has a bachelor's degree in political science and no formal technical training, said he wanted to bring in other nuclear experts, particularly ones who had long experience in developing the miniaturized nuclear triggers for hydrogen bombs. John L. Richter of Los Alamos, a scientist who filled that void, joined the team.
The group looked more closely at a clue provided by the Chinese spy, who described the size of the bomb's atomic core with an analogy to a common household object, officials said in a new disclosure. Working from that, the scientists calculated a more precise size and Dr. Henson and Dr. Richter went through the American stockpile of nuclear arms, looking up measurements to see if any matched.
The atomic trigger of the W-88, they discovered, was close enough in size to raise suspicions.
The Energy Department held meetings in which the Los Alamos team was joined by analysts from the C.I.A. and the Defense Intelligence Agency. Federal officials now say the intelligence agencies were skeptical, reasoning that too much was being made of a foreigner's rough analogy. But the Energy Department and the Los Alamos team felt the evidence was provocative.
The breakthrough came in 1995, as has been previously disclosed, when a Chinese Government official sent a package of secret Chinese documents to American officials.
Mr. Trulock said the most revealing document, dated 1988, laid out China's nuclear modernization plans for Beijing's First Ministry of Machine Building, which, among other things, made missiles and nose cones. It not only described China's plans but compared them to the nuclear arms of the American arsenal.
Relatively crude hand drawings sketched out the nose cones enveloping the W-88, the W-87, the W-78, the W-76, the W-62 and the W-56 -- warheads of the Trident, MX and Minuteman missiles -- and also gave their overall weights and dimensions.
In itself, these were not damning. Though still officially classified secret in some cases, such information by then was widely available in many unclassified American papers and articles.
But the Chinese document, some 20 pages in translation, went on to give sensitive data about the W-88, Federal officials revealed. It accurately described the shape of the atomic trigger as not spherical and said it was situated in the nose cone's narrow forward end -- an arrangement used in some but not all American warheads. And it correctly described the hydrogen fuel, or secondary, as having a spherical shape.
More unsettling to the team, it described the width of the casing that surrounds the atomic trigger to within a millimeter, or four-hundredths of an inch. ''That's pretty damn accurate,'' Mr. Trulock recalled.
A senior Federal official agreed. ''That opened eyes,'' he said. ''It seemed to confirm earlier assessments that had seemed insubstantial.''
Mr. Trulock said his team later found that the Chinese document gave a similarly exact measure for the width of the W-88's secondary, or hydrogen stage. ''Primaries are the long pole in the tent,'' he said, referring to the importance of the atomic trigger. ''But that measurement was as good as the one for the primary.''
The C.I.A. eventually concluded that the agent who sent the documents was acting under the instruction of Chinese intelligence. No one has ever come up with a persuasive explanation of why China sent the documents to American spies.
From 1992 to 1996, American officials revealed, China used its new atomic match to ignite a variety of hydrogen bombs, including one similar in some respects to the W-88. After this series of blasts shook the ground at the Lop Nur test site, China signed the Comprehensive Test Ban Treaty, signaling an end to its nuclear experimentation.
The Investigation
Federal Sleuths Hunt for a Spy
The Energy Department opened an investigation into the possible theft of W-88 secrets on Sept. 28, 1995, and over the next three years, Federal officials quietly tried to find out whether there was a Chinese spy in their midst.
If espionage occurred, Mr. Trulock and his team reasoned, it must have happened between 1984, when the warhead entered engineering development, and 1988, the date of the Chinese document.
Energy Department officials focused on Los Alamos, which had designed the bomb. They looked particularly closely at anyone who had traveled to China in those years or met visiting Chinese scientists.
Mr. Vrooman, then head of counterintelligence at Los Alamos and later a vocal critic of the inquiry, said investigators scrutinized only those people whose trips to China were paid for by the Energy Department.
Left unexamined, he said, were at least 15 additional people whose trips were paid for by the Chinese, the C.I.A., the Air Force or privately. These travelers tended to be top weapons designers and high officials -- the people who knew the most American arms secrets and had the most intimate contact with Chinese peers, Mr. Vrooman said.
In May 1996, the Energy Department turned over a list of a dozen suspects to the Federal Bureau of Investigation, which began a criminal case that eventually narrowed to Dr. Lee, an American scientist of Taiwanese birth working at Los Alamos.
Dr. Lee and his wife, Sylvia, had traveled to China in 1986 and 1988. Mrs. Lee was a secretary at Los Alamos who often met visiting Chinese delegations. And Dr. Lee, though a mechanical engineer by training and never a weapons designer, was familiar with the W-88 and many other nuclear arms and secrets (including the atomic trigger advance) because of his work on secret computer codes.
The F.B.I. believed it had enough evidence to seek a secret wiretap on Dr. Lee's phone calls, citing 20 reasons he was a prime suspect. But the Justice Department found the evidence unpersuasive and refused to seek a court order for the eavesdropping, a routine step in most spy cases.
Mr. Vrooman has charged that the inquiry was marred by a racist bias to target Chinese-Americans, an assertion Federal officials have vehemently denied. But the Republican chairman and the ranking Democrat of the Senate Governmental Affairs Committee, which investigated the spy case and heard testimony from Mr. Vrooman, concluded that Federal investigators had focused prematurely on Dr. Lee.
After the spy case became in public, Bill Richardson, the Secretary of Energy, recommended that Mr. Vrooman be disciplined for letting Dr. Lee have continuing access to secrets even after doubts about him had been raised.
Dr. Lee, fired this year from Los Alamos for security violations, including failing to report foreign contacts, has been charged with no crime and has denied any spying. After his ouster, investigators found that he had loaded many secret files onto an unsecured computer, raising the risk that they could have fallen into the wrong hands.
The inquiry most likely would not have come into public view had it not been for a series of unrelated disclosures about China.
In April 1998, The Times reported that two United States aerospace companies were under criminal investigation for providing rocket data to Chinese scientists.
A furor erupted in Congress. The House created a select committee, led by Mr. Cox, who had recently vied unsuccessfully for the House speakership, to look into whether the Administration's increasingly open policies on satellite exports had compromised national security.
There was no hint the committee would end up studying nuclear bombs. Composed of five Republicans and four Democrats, the committee did not learn of the suspected Chinese nuclear espionage until October 1998, just a few months before its mandate expired. On Nov. 12 and Dec. 16 it held secret hearings in which Mr. Trulock was called as the star witness.
In January, after three months of investigation, the committee completed a secret manuscript. In May, after a long argument with the White House over what could be made public, it released an 872-page report. The chapter on atomic espionage, just 37 pages, garnered most of the headlines.
In fiery prose accompanied by vivid color pictures and charts, the committee charged that Chinese spies had carried off vital secrets about seven of America's most advanced arms.
The People's Republic of China, it alleged, ''has stolen classified information on all of the United States' most advanced thermonuclear warheads,'' leaping from the clumsy designs of the 1950's to those that are far more modern and deadly.
The main evidence cited was the Chinese document obtained by the C.I.A. in 1995 and an inquiry in the 1980's into spying at the Livermore lab that concluded China had most likely obtained design secrets of the neutron bomb. The unclassified version of the committee's report gave no details of the 1995 document's secret details about the W-88.
The report was signed by the committee's four Democrats. But immediately after its release, Representative John M. Spratt Jr. of South Carolina, one of the Democrats, criticized it as rushed, superficial and exaggerated. The witnesses heard by the committee, he added, ''did not have the technical background to fully assess the nature or value of the information lost.''
The Debate
Analysts Sift For the Truth
Since then, Mr. Spratt's critique has been echoed and amplified by a range of top scientists and bomb designers who say Beijing could have miniaturized its warheads on its own without spying.
Richard L. Garwin, a physicist who has long advised Washington on nuclear arms, recently on a bipartisan team led by former Defense Secretary Donald H. Rumsfeld, said, ''there is no reason to believe that China could not have built perfectly adequate warheads'' for a range of modern missiles ''from nuclear technology that it developed itself.''
China, several officials said, simply went down the same path as other nuclear powers, helped along by the general knowledge of what the United States had achieved: proof that hydrogen bombs can be made very small but nonetheless very powerful.
''Every state has come to it,'' said one Federal official, referring to breakthroughs in atomic triggers by the Soviet Union, Britain and France. ''Now they've got it too.''
Mr. Hawkins, the head of international security studies at Los Alamos, which is clearly on the defensive because of the spy scandal, said the basic physics of bombs and missiles push weapons designers in roughly the same direction. To obtain the best performance, he said, engineers are invariably led toward narrow nose cones about 16 degrees wide -- if cut from a pie, a very modest slice.
''Once you realize that,'' Mr. Hawkins said, ''it drives every nation down similar paths. Eventually, all weapons systems will look alike. It has to do more with physics than espionage.''
That view is not universally accepted.
Dr. Henson, the analyst who first sounded the alarm at Los Alamos, said there was nothing in the design of missile nose cones that propelled a scientist to shape the core of an atomic trigger into an oval.
Do scientific and technical analyses automatically ''draw you to a watermelon?'' he asked, alluding to the shape of the top-secret design. ''That's not true.''
''It's beyond a shadow of a doubt,'' Dr. Henson added. ''Major espionage took place.''
American intelligence agencies are less categorical. Analysts have concluded that espionage played a role in Beijing's advance, but cannot identify a hard link comparable to the Soviet Union's theft in the 1940's of the American design for the first atom bomb.
''Everybody has come to the same conclusion,'' said a top Administration official who has closely scrutinized the secret data. ''We don't have a smoking gun.''
A Federal intelligence study done last year, which the Cox committee drew on, said American secrets lost between 1984 and 1988 let the Chinese ''accelerate their nuclear weapons program well beyond indigenous capabilities,'' a view that echoed the original Los Alamos finding.
A damage assessment by the American intelligence community, made public in April, said a mix of espionage, openly available data and scientific acumen had greatly lengthened Chinese strides. Stolen secrets, it said, ''could help'' Beijing develop a mobile missile and ''probably accelerated its program to develop future nuclear weapons.''
In June, the President's Foreign Intelligence Advisory Board, which did its own investigation, said both Congressional and Administration leaders had engaged in ''simplification and hyperbole'' in the spy case. Neither dramatic damage assessments nor categorical reassurances, it said, were wholly substantiated.
And Mr. Hawkins, the Los Alamos official, said the specific secrets known to have been seized by the Chinese, principally those detailed in the 1995 document, would have been little help to a bomb maker, and far from Mr. Trulock's road map. As for an H-bomb's innards, what designers call the physics package, Mr. Hawkins said the documents ''describe nothing significant.''
Mr. Cox insisted that highly classified intelligence data available to his committee showed a more persuasive case than has emerged publicly. ''There are more interpolating facts'' that closely tie lost W-88 secrets to Beijing's advance, he said.
But a Federal official cited intelligence data about China's atomic trigger showing it to be anything but an exact copy.
''It turns out the W-88's is slightly smaller,'' said the official, who believes Beijing may have made the advance on its own.
It remains unresolved how China got the W-88 secrets in the first place, but a consensus is emerging that the search for the leak narrowed too quickly to Los Alamos.
Studies by the Senate as well as the President's foreign intelligence board this year raised serious questions about whether the F.B.I. and Energy Department had too quickly focused on the weapons lab. No evidence has pinpointed it as the leak's source.
Mr. Vrooman, the head of counterintelligence at the laboratory from 1987 until 1998, noted that one secret document describing the design of the W-88 warhead went to 548 mailing addresses throughout the Government and military. Some Administration experts believe the data described by the Chinese in the 1995 document came from engineering plans or from secret manuals on military bases.
''That kind of information was widely available,'' said Dr. Drell of Stanford, who served on the President's advisory board investigation. ''The manuals that went out had pictures and numbers. If a submarine came in, and there was a problem, they had to know what they were dealing with.''
However Beijing made its miniaturization advance -- on its own, by theft or a combination of the two -- it is apparently proud enough to boast about it publicly, at least among its friends in the mountains of New Mexico. Dr. Henson said a Chinese arms scientist, Sun Cheng Wei, bragged of the breakthrough at Los Alamos a few years ago, telling an open seminar that China had forged significantly ahead in nuclear arms.
''What he said,'' recalled Dr. Henson, who attended the talk, ''was that for a long time they were dealing only with round designs, and then only watermelons.''
Photos: The world's first nuclear explosive device, above, was code-named the Gadget. Dr. Norris E. Bradbury, director of Los Alamos National Laboratory from 1945 to 1970, posed with it in 1945. Below, the first hydrogen bomb, called Mike, in 1952. (Jornada Del Muerto Valley, New Mexico, July 1945); (Los Alamos National Laboratory)(pg. A14); The United States' most advanced nuclear warhead is the W-88, at least eight of which can fit on the Trident D-5 missile. (Courtesy of Federation of American Scientists)(pg. A15) Diagrams: Some secret specifications of the W-88, an American miniature hydrogen bomb, that were found in a Chinese document. SHAPE OF ATOMIC BOMB TRIGGER Described as not spherical SHAPE OF HYDROGEN BOMB FUEL -- Described as spherical SIZES OF CASINGS BOMB PLACEMENT -- Atomic bomb trigger is placed above the hydrogen bomb fuel. (Illustration by Mika Grondahl)(pg. A1) ''Less but More'' Early bombs were behemoths compared with modern miniature devices. The W-88, the U.S. arsenal's most modern warhead, though tiny, is dozens of times more powerful than the bomb that leveled Hiroshima. A-BOMBS Little Boy -- 1940's Dropped on Hiroshima Fat Man -- 1940's Dropped on Nagasaki H-BOMBS B-17 -- 1950's Largest bomb deployed by U.S. W-88 -- 1980's As many as eight fit atop Trident II (D-5) missiles (Sources: Federation of American Scientists; The Cox Report)(pg. A14) ''A Missile That Carries Miniature Bombs'' The miniaturization of the W-88 hydrogen warhead, considered the most advanced in the American arsenal, made the Trident submarine missile one of the most deadly weapons of all time. Trident II (D-5) missile W-88 WARHEADS The Trident can carry eight of these, each of which can seek a different target. THIRD STAGE MOTOR Accelerates the warhead to approximately 15,000 m.p.h. SECOND STAGE MOTOR FIRST STAGE MOTOR Ignites after missile clears the water's surface. Missile is fired from a submarine. (Source: Federation of American Scientists)(pg. A15)