Stanford Ovshinsky may not be a household name, but his inventions have the power to change the world

Nov 30th 2006 | from the print edition

“THE ages of mankind have been classified by the materials they use—the Bronze Age, the Iron Age, the Age of Silicon. We are at the dawn of the Hydrogen Age.” So proclaims Stanford Ovshinsky, co-founder of Energy Conversion Devices (ECD), a company based near Detroit, Michigan. “What is more,” he says, “the hydrogen economy is happening already.”

There have been plenty of grandiose but unsubstantiated claims made over the past five years about the potential for hydrogen to replace fossil fuels as an energy carrier, so some scepticism is certainly in order. In particular, President George Bush and the big carmakers have been trumpeting hydrogen fuel cells—electrochemical devices that turn hydrogen into electricity and water vapour—as the replacement for the internal-combustion engine. But the date of commercialisation seems forever slipping just beyond the horizon.

That has prompted a backlash from advocates of rival technologies (such as ethanol-based engines and novel batteries) and from greens, who argue that hydrogen is just a cynical long-term diversion used by Mr Bush and Detroit to avoid short-term action on fuel-economy standards, plug-in hybrids and other here-and-now options. And yet here is Mr Ovshinsky, still trumpeting hydrogen’s virtues despite bitter opposition.

Three things set Mr Ovshinsky apart from the hydrogen hypesters. First of all, he is no newcomer. He first outlined his vision for what he calls a “hydrogen loop” some five decades ago as an alternative to fossil fuels. (The loop goes from water to stored hydrogen via solar-powered electrolysis, and from hydrogen back to water, generating electricity in the process, via a fuel cell.) Unlike others, he can hardly be accused of opportunistically seizing upon this obscure techno-fix for political reasons.

The second difference is that Mr Ovshinsky’s green credentials are impeccable. He and his wife Iris, who died recently, founded ECD in 1960 with the explicitly stated goal of “using creative science to solve societal problems”. Astonishingly, they had the foresight to predict—long before the oil shocks of the 1970s—that the world’s addiction to oil would have unacceptable side effects, from resource wars to climate change. Spend time with Mr Ovshinsky and his employees, and it becomes plain that his social values permeate his organisation.

But what lifts Mr Ovshinsky into the league of genius inventors is something rather less common: success. He is the inventor of the nickel-metal hydride (NiMH) battery, which is used to power everything from portable electronics to hybrid cars; around 1 billion such batteries are sold every year. He has also made advances in information technology (he calls information “encoded energy”) and holds critical patents relating to thin-film solar cells, rewriteable optical discs, a new form of non-volatile memory and flat-panel displays. These technologies are being commercialised through deals with Intel, Samsung, STMicroelectronics, General Electric, Chevron, United Solar Ovonic, and others.


Innovation from disorder


What all these apparently disparate inventions have in common is that they rely on Mr Ovshinsky’s path-breaking discoveries in the field of disordered or “amorphous” materials, since named “ovonics” in his honour. Such materials can be used for energy generation (in fuel cells and solar cells), for energy storage (in batteries), for computing (to store data on discs or in chips) and to create custom materials with novel properties.

Mr Ovshinsky has spent the past five decades devising actual working products, based on amorphous materials, that fill every niche in his hydrogen loop, from thin-film solar panels to solid-hydrogen storage tanks to “regenerative” fuel cells that can store energy captured while a car is braking. ECD has even “hacked” a Toyota Prius hybrid car so that it runs on pure hydrogen rather than petrol, which he says proves that “we don’t have to wait for fuel cells to move into the hydrogen economy.”

All this makes it tempting to compare ECD’s co-founder with Thomas Edison, the great inventor from another age who founded General Electric. Both established themselves early on not only as brilliant innovators, but inventors with their feet firmly planted on the ground. Both arose from humble roots: Edison was not born to privilege, while Mr Ovshinsky’s father collected scrap by buggy. Mr Ovshinsky did not even go to college, and credits his vast knowledge of science to the public libraries of his native Ohio. He likes to say, “invention comes to the prepared mind.” And Edison, like Mr Ovshinsky, straddled the fields of energy and information technology: he originally made his name with the invention of the quadruplex, a device that increased the capacity of telegraph lines, before moving on to electrification.

Another similarity between the two inventors is that both thought of their inventions as entire systems. They had the verve to envisage a radically different world, but were good at inventing the practical things needed to get there. In Edison’s case, his vision was that of mass electrification. He was not the first to make a light bulb, but he vastly improved it and, more importantly, created the generation and distribution technologies needed to make it work, from power stations to electricity meters. His company, now called GE, helped to light up America and then the world.

Despite his lack of formal training, the charming, soft-spoken Mr Ovshinsky is not at all threatened by scientists with fancy degrees: he hires many of them, and has hosted lively debates around a round table at ECD with such prominent scientists as Hellmut Fritzsche and Morrel Cohen of the University of Chicago, David Adler of MIT and Sir Neville Mott of Cambridge University (who went on to win a Nobel prize for work on amorphous materials). Ask him whether he expects his own Nobel, and he responds matter of factly: “Oh, never. I’ve been nominated before, and Mott gave me credit when he won his, but I’ll never get one.” Without a hint of bitterness he adds softly, “I’m not a part of their world.”

Mr Ovshinsky’s vision for a hydrogen loop was just a blackboard exercise five decades ago. But since then he has produced the inventions needed to make it work. “Stan starts with a vision, and then goes out to invent what we need to get from here to there,” says Joachim Doehler, a senior scientist at ECD. Doing this requires more than scientific theory: it requires a practical engineer’s mind too. “Stan is a very good toolmaker,” says Robert Stempel, ECD’s chairman (and a former boss of General Motors, a big carmaker). Mr Ovshinsky’s collaborators say that he has an astonishing ability to juggle the permutations of eight or ten novel materials in his head, which gives him an intuitive grasp of which scientific leads to follow. That said, his colleagues joke, he still sometimes cannot remember names correctly.

The best evidence of Mr Ovshinsky’s systems approach at work is his shiny new solar factory in Michigan. Several decades ago, he argued that solar panels ought to be made not as brittle crystalline panels in costly batch processes—how everyone else does it today—but in a continuous process, “by the mile”. He was ridiculed. But he refused to yield, and asked his team to devise processes for producing miles of thin-film solar material. Dr Doehler, a veteran of AT&T’s legendary Bell Labs research centre, recalls telling his boss it was impossible. The boss proved him wrong, personally designing much of the solar factory from scratch. Crucially, his approach does not require the expensive silicon used in conventional solar panels.


A sunny future


Mr Ovshinsky points to the happy result on the shop floor: a flexible, self-adhesive strip of solar material that makes power even on cloudy days and is virtually indestructible. The factory, which Mr Bush visited in February, has an order backlog of six months and profit margins approaching 30%, he says. He has another factory in the works nearby, and plans for more: “I see ECD’s future as a factory for factories. That’s how you build entirely new industries for the future.” So does he see ECD as the GE of the 21st century? “Oh, ECD will be much more than that,” says Mr Ovshinsky merrily. “Energy and information are the twin pillars of the global economy, after all.”

How justified is this boast? Few question his intellect, but some do challenge his record as a corporate boss. An article in Forbes magazine asked in 2003 why investors “keep giving money to Stan Ovshinsky, the inventor who can create anything but profits.” ECDhas lost money for most of the 40-plus years that it has been a public company. As even one of Mr Ovshinsky’s loyal lieutenants confesses, “This company would have gone bust six times already if it were not for the personal loyalty people felt for Stan and Iris; we went the extra mile for them because this place is unique.”

Inspired by the family’s links to the peace and civil-rights movements, the Ovshinsky motto is “with the oppressed, against the oppressor”, and ECD retains the feel of a family firm with those values. What is more, ECD is visibly committed to clean energy—and Mr Ovshinsky is clearly not motivated by money. The New York Times recently analysed executive pay in America and found that bosses typically get 500 times the salary of the average worker at their firms; the ratio at ECD is five to one. He even points out that he is “probably the only chief executive that is a union member”.

The loss of his wife, collaborator and co-founder has clearly devastated Mr Ovshinsky, but do not expect to see him retire anytime soon. He may be 84, but he evidently has plenty of unfinished business to attend to. He still rises early, dresses in natty suits, and moves with the agility and energy of a young man. His intellectual curiosity appears entirely undiminished by a life of learning: his desk at ECD is buried under neat stacks of annotated scientific papers, business plans and other reading material. And he remains as audaciously inventive as ever.

He has worked out how his next generation of solar films will be produced not at 2.5 feet per minute, he says, but 100 times faster. He is convinced he can radically improve the efficiency of fuel-cell electrodes. He thinks he will be able to scale up his firm’s hydrogen-storage system to megawatt scale, thus enabling grid storage of renewable power. And so on. As your correspondent departed at the end of a day-long visit, Mr Ovshinsky still had a dinner interview with a television crew, and then planned to work on a cosmology paper at home. As I.I. Rabi, a Nobel prize-winning physicist, is reported to have said when asked if his friend was another Edison: “He’s an Ovshinsky, and he’s brilliant.”

from the print edition | Technology Quarterly


兹韦尔(Ray Kurzweil)

本文作者库兹韦尔(Ray Kurzweil)曾发明第一种供盲人使用的阅读机、第一种词汇量庞大的商用语音识别软件及许多种其他技术。他指出,对发明家来说,今天的一切事物都在迅速变化。他在本文中提出了一些建议,指点同行如何赶上时势。

常常有人问我,怎样才可以成为一位成功的发明家。我从事发明30 多年,已明白了一些道理:从事发明颇象玩冲浪游戏,你必须在适当的时间冲上浪头。这就是我热中于研究科技趋势的理由。现在我有一组研究人员负责收集各种不同科技的数据,而我则利用数据制造数学模型,用来显示科技在不同的领域中如何发展。从这些模型可以看出,创新本身的速度每10 年增快一倍。

当一种技术的飞速发展期渐近尾声时,我们能在适当的时候发明出另一种新技术至关重要。你必须让你的发明着眼于未来的世界,而不是你的研究项目开展时的世界。将来你的发明问世时,这个世界肯定不再是今天的模样了。到时候一切都已经改变,市场需求、竞争、分销渠道、研发工具和辅助技术等都会不同了。为了能在适当的时间发明,首先必须了解一种发明的整个生命周期。一种技术的发展过程可以分为7 个阶段:设想、发明、发展、成熟、出现虚假的威胁、过时、成为古董。必须在几个关键阶段(即设想、发明、发展和成熟阶段)做得较好,一项发明才可能广受欢迎,最终演变为成功的产品。


在设想阶段,有助于发明出新技术的因素已基本具备,甚至有些设想还可能包括了运作情况和目标,但新技术还有待发明。例如,达芬奇(Leonardo da Vinci)曾描述过飞行机器,但我们不会认为他是飞机的发明者。我们的社会特别称颂发明,然而发明阶段仅仅具有承前启后的作用。发明家必须兼具科学知识和解决实际问题的技巧。显然,他们还必须坚强不屈。例如,爱迪生(Thomas Edison)在找到令人满意的灯泡灯丝之前,曾试验过几千种材料。此外,就象我说过的那样,发明家必须善于审时度势,并具备推销技巧来吸引必需的资源,其中包括投资、合作者、客户等。

第三个阶段是发展。一项新发明刚问世时,通常只是笨拙且缺乏实用价值的装置。如果要以莱特兄弟(Wright brothers)所发明的飞机为中心发展出一套有效的商业模式将会十分困难。飞机经过许多改进之后,我们才真正进入航空时代。


但随着时间的流逝,新一代发明家找到了方法来弥补新技术的缺点,从而将旧技术推入了过时阶段。在整个生命周期中,过时阶段占了5%~10%。一种技术的最终结局就是成为古董。今天的马和马车、手动打字机等都是例子,而音乐CD 很快就会步其后尘。

我曾经参与发明一种新技术来取代另一种已成熟了很久的技术:钢琴。钢琴的前身是羽管键琴(Harpsichord)。但音乐家对羽管键琴并不满意,因为它发出的声音在强度上没有变化。于是,意大利人克利斯托弗利(Bartolomeo Cristofori)发明了一种能做到这一点的琴。他将这种琴叫做“Gravicembalo col piano e forte”(能发出弱音和强音的大键琴),简称“钢琴”(Piano)。起先钢琴并不普及,但经过改进之后,它终于在整个19 和20 世纪成为了西方人最喜爱的键盘乐器。

在20 世纪80 年代初期,电子钢琴成为虚假的威胁。电子钢琴有很多优点,例如不需要调音、能发出全套声音、自动伴奏等。但电子钢琴缺少了一个最重要的特点:真正的钢琴音。后来,通过使用先进的信号处理技术和模式识别技术弥补了这种缺陷。今天,电子钢琴的音质已经超越立式钢琴(这种钢琴曾在传统钢琴市场中居于主导地位)。如今电子钢琴销售量与日俱增,正渐渐主宰钢琴市场,传统钢琴的销售量则持续下降。

发明的3 个步骤



对于跨学科技术的发展来说,一个主要的障碍就是不同的学科往往使用不同的术语来代表同一个概念。美国数学家维纳(Norbert Wiener)在他写于1948 年的名著《控制论》(Cybernetics)中曾针对这个问题作出过一番评论:“科学工作有很多领域⋯⋯分别从纯粹数学、统计学、电机工程学和神经生理学等不同的方面进行探究⋯⋯每一个概念都在不同的小组中有不同的名字,有一项重要的工作重复做了三四次,另一项重要的工作则因为在某个领域中尚未得出结果而被耽误,其实,这些结果可能在另一领域中早已存在。”



更重要的是,我设法招揽一些将受惠于我准备发明的那种技术的专家加入团队并成为主要成员。例如,我在20 世纪70 年代研发供盲人使用的阅读机时,聘请了美国全国盲人联盟(The National Federation of the Blind)的失明科学家和工程师。又如,我在20 世纪80 年代研发音乐合成技术时,要求团队里的每一位工程师都精通音乐。对于一种技术中的微妙问题,该技术的使用者会很敏感,而非使用者则无法察觉。





库兹韦尔以他自己的发明为基础创建了9 家企业。美国国家杰出发明家纪念馆(The National Inventors Hall of Fame)已将他列入名人录


nucifera: At last, 大家可以把目光回到佛陀教育理所说的“戒、定、慧”。无戒不生定,无定慧不存。




By Clare Baldwin
AUGUST 24, 2009

SAN FRANCISCO (Reuters) – Multitaskers of media activities like watching YouTube, writing e-mail and talking on the phone are not very good at any of their tasks, according to a Stanford University report on Monday.

Researchers who published the report in the Proceedings of the National Academy of Sciences said the results had surprised them. They were looking for the secret to good media multitaskers but instead found broad-based incompetence.

“Heavy multitaskers are lousy at multitasking… The more you do it, the worse you get,” said Stanford communications professor Clifford Nass.

Compulsive media multitaskers are worse at focusing their attention, worse at organizing information, and worse at quickly switching between tasks, the Stanford scientists wrote.

After testing about 100 Stanford students, the scientists concluded that chronic media multitaskers have difficulty focusing and are not able to ignore irrelevant information.

Nass said that multitasking is becoming more widespread — some jobs require workers to keep an instant message window open — and the scientists were surprised at the results.

“We knew that multitasking was difficult from a cognitive perspective. We thought, ‘What’s this special ability that people have that allows them to multitask?’ … Rather than finding things that they were doing better, we found things they were doing worse,” Stanford symbolic systems professor Eyal Ophir said.

A bright side to such distraction may mean that the media multitaskers will be first to notice anything new, Ophir said.

(Reporting by Clare Baldwin; Editing by Tim Dobbyn)

nucifera: 原来公主等待白马王子的降临,还真是场白日梦——科研如是说。


July 7, 2009

Scientists have long observed that women tend to be pickier than men when choosing a mate. The usual explanation is evolutionary: because women have a bigger investment in reproduction — they are the ones who have to endure pregnancy, childbirth and breast-feeding — they need to hedge their bets against selecting a dud to be the father.

In recent years, the emergence of speed dating has given psychologists, economists and political scientists new ways to test this and other hypotheses about mating. Because participants can be randomly assigned to groups and have no prior information about other participants, three-minute speed-dating sessions are about as close to a controlled experiment as researchers are likely to get.

Now, two scientists at Northwestern University have published an experiment that challenges the evolutionary hypothesis. The study by Eli J. Finkel and Paul W. Eastwick was published last month in the journal Psychological Science.The experiment looked at speed-dating sessions to determine whether men or women were choosier. The answer, it turned out, was neither. Regardless of gender, people who were instructed to approach other daters were less selective — that is, they were more likely to ask to meet later for a date.

Dr. Finkel and Mr. Eastwick write that this does not mean men were just as selective as women. But the scientists suggest that the explanation for the gap lies in social conditioning rather than evolution.

By making the first move, a person gains confidence and then finds more people attractive, the theory goes. Culturally, men are expected to approach women more often, which may boost their confidence and make them less selective. Citing what social psychologists call the scarcity principle, Mr. Eastwick and Dr. Finkel write that “individuals tend to place less value on objects or opportunities that are plentiful than those that are rare.” By contrast, they say, women are accustomed to being approached, which may make them feel more desirable and thus more selective.

Scientists have also used speed-dating experiments to examine the tendency for people to mate with people like themselves. A 2006 paper by economists at the University of Essex in England analyzed data from 3,600 male and female speed daters to see if people selected mates with similar traits, like height and education, because that is what they prefer or because they are most likely to encounter them in the dating market.

The economists, Michèle Belot and Marco Francesconi, found that men’s preferences for occupation, height and smoking had little effect on whom they asked out. Those factors also did not matter to women, but age did.

In homogeneous environments, Dr. Belot and Dr. Francesconi wrote, people are more likely to marry others like themselves, while more diverse communities are likely to produce more varied pairings.

“Mating requires meeting,” they wrote. “The pool of potential partners shapes the type of people to whom subjects propose and ultimately with whom they form long-term relationships.”

People narrow their market opportunities, the economists suggested, by selecting for height, weight and age, which tend to be proxies for socioeconomic status.

So how does a person increase the odds of crossing paths with someone who matches his or her preferences? Maybe by tapping into social networks. In “Connected: The Surprising Power of Our Social Networks and How They Shape Our Lives,” a book to be published in September, Dr. Nicholas A. Christakis of Harvard Medical School and James H. Fowler, a political scientist at the University of California, San Diego, argue that dating is not a random process.

They cite a landmark 1992 Chicago sex survey of 3,432 adults ages 18 to 59, which found that 68 percent of married people in the survey reported meeting their spouse through a friend, family member or other mutual acquaintance.

“If you are single and you know 20 people reasonably well, and if each of them knows 20 other people, and each of them knows 20 other people,” Dr. Christakis and Dr. Fowler write, “then you are connected to 8,000 people who are three degrees away. And one of them is likely to be your future spouse.”

nucifera: 有人对我说,他们注定学不好数理,无法与有天分的朋友相较。或是有人对我说,他们注定没法掌握文学和艺术,这与天分有关系。这篇文章还有好些研究告诉我们,这些看法都没法立足。




May 1, 2009
From The New York Times

Some people live in romantic ages. They tend to believe that genius is the product of a divine spark. They believe that there have been, throughout the ages, certain paragons of greatness — Dante, Mozart, Einstein — whose talents far exceeded normal comprehension, who had an other-worldly access to transcendent truth, and who are best approached with reverential awe.

We, of course, live in a scientific age, and modern research pierces hocus-pocus. In the view that is now dominant, even Mozart’s early abilities were not the product of some innate spiritual gift. His early compositions were nothing special. They were pastiches of other people’s work. Mozart was a good musician at an early age, but he would not stand out among today’s top child-performers.

What Mozart had, we now believe, was the same thing Tiger Woods had — the ability to focus for long periods of time and a father intent on improving his skills. Mozart played a lot of piano at a very young age, so he got his 10,000 hours of practice in early and then he built from there.

The latest research suggests a more prosaic, democratic, even puritanical view of the world. The key factor separating geniuses from the merely accomplished is not a divine spark. It’s not I.Q., a generally bad predictor of success, even in realms like chess. Instead, it’s deliberate practice. Top performers spend more hours (many more hours) rigorously practicing their craft.

The recent research has been conducted by people like K. Anders Ericsson, the late Benjamin Bloom and others. It’s been summarized in two enjoyable new books: “The Talent Code” by Daniel Coyle; and “Talent Is Overrated” by Geoff Colvin.

If you wanted to picture how a typical genius might develop, you’d take a girl who possessed a slightly above average verbal ability. It wouldn’t have to be a big talent, just enough so that she might gain some sense of distinction. Then you would want her to meet, say, a novelist, who coincidentally shared some similar biographical traits. Maybe the writer was from the same town, had the same ethnic background, or, shared the same birthday — anything to create a sense of affinity.

This contact would give the girl a vision of her future self. It would, Coyle emphasizes, give her a glimpse of an enchanted circle she might someday join. It would also help if one of her parents died when she was 12, infusing her with a profound sense of insecurity and fueling a desperate need for success.

Armed with this ambition, she would read novels and literary biographies without end. This would give her a core knowledge of her field. She’d be able to chunk Victorian novelists into one group, Magical Realists in another group and Renaissance poets into another. This ability to place information into patterns, or chunks, vastly improves memory skills. She’d be able to see new writing in deeper ways and quickly perceive its inner workings.

Then she would practice writing. Her practice would be slow, painstaking and error-focused. According to Colvin, Ben Franklin would take essays from The Spectator magazine and translate them into verse. Then he’d translate his verse back into prose and examine, sentence by sentence, where his essay was inferior to The Spectator’s original.

Coyle describes a tennis academy in Russia where they enact rallies without a ball. The aim is to focus meticulously on technique. (Try to slow down your golf swing so it takes 90 seconds to finish. See how many errors you detect.)

By practicing in this way, performers delay the automatizing process. The mind wants to turn deliberate, newly learned skills into unconscious, automatically performed skills. But the mind is sloppy and will settle for good enough. By practicing slowly, by breaking skills down into tiny parts and repeating, the strenuous student forces the brain to internalize a better pattern of performance.

Then our young writer would find a mentor who would provide a constant stream of feedback, viewing her performance from the outside, correcting the smallest errors, pushing her to take on tougher challenges. By now she is redoing problems — how do I get characters into a room — dozens and dozens of times. She is ingraining habits of thought she can call upon in order to understand or solve future problems.

The primary trait she possesses is not some mysterious genius. It’s the ability to develop a deliberate, strenuous and boring practice routine.

Coyle and Colvin describe dozens of experiments fleshing out this process. This research takes some of the magic out of great achievement. But it underlines a fact that is often neglected. Public discussion is smitten by genetics and what we’re “hard-wired” to do. And it’s true that genes place a leash on our capacities. But the brain is also phenomenally plastic. We construct ourselves through behavior. As Coyle observes, it’s not who you are, it’s what you do.

nucifera: 这无疑是很重要的报导,但毕竟这份报导挺久的了。去年的时候The economist也有做过类似的报导,那时这个概念比起先前获得了更多的证实和肯定。


Published: December 21, 2007

Within the next few months, researchers at three medical centers expect to start the first test in patients of one of the most promising — and contentious — ideas about the cause and treatment of cancer.

The idea is to take aim at what some scientists say are cancerous stem cells — aberrant cells that maintain and propagate malignant tumors.

Although many scientists have assumed that cancer cells are immortal — that they divide and grow indefinitely — most can only divide a certain number of times before dying. The stem-cell hypothesis says that cancers themselves may not die because they are fed by cancerous stem cells, a small and particularly dangerous kind of cell that can renew by dividing even as it spews out more cells that form the bulk of a tumor. Worse, stem cells may be impervious to most standard cancer therapies.

Not everyone accepts the hypothesis of cancerous stem cells. Skeptics say proponents are so in love with the idea that they dismiss or ignore evidence against it. Dr. Scott E. Kern, for instance, a leading pancreatic cancer researcher at Johns Hopkins University, said the hypothesis was more akin to religion than to science.

At stake in the debate is the direction of cancer research. If proponents of the stem-cell hypothesis are correct, it will usher in an era of hope for curing once-incurable cancers.

If the critics are right, the stem-cell enthusiasts are heading down a blind alley that will serve as just another cautionary tale in the history of medical research.

In the meantime, though, proponents are looking for ways to kill the stem cells, and say that certain new drugs may be the solution.

“Within the next year, we will see medical centers targeting stem cells in almost every cancer,” said Dr. Max S. Wicha, director of the University of Michigan Comprehensive Cancer Center, one of the sites for the preliminary study that begins in the next few months (the other participating institutions are Baylor College of Medicine in Houston and the Dana-Farber Cancer Institute in Boston).

“We are so excited about this,” Dr. Wicha said. “It has become a major thrust of our cancer center.”

At the National Cancer Institute, administrators seem excited, too.

“If this is real, it could have almost immediate impact,” said Dr. R. Allan Mufson, chief of the institute’s Cancer Immunology and Hematology Branch.

The cancer institute is financing the research, he said, and has authorized Dr. Mufson to put out a request for proposals, soliciting investigators to apply for cancer institute money to study cancer stem cells and ways to bring the research to cancer patients. The institute has agreed to contribute $5.4 million.

“Given the current fiscal situation, which is terrible, it’s a surprising amount,” Dr. Mufson said. “We actually asked for less,” he added, but the cancer institute’s executive committee asked that the amount be increased.

Proponents of the hypothesis like to use the analogy of a lawn dotted with dandelions: Mowing the lawn makes it look like the weeds are gone, but the roots are intact and the dandelions come back.

So it is with cancer, they say. Chemotherapy and radiation often destroy most of a tumor, but if they do not kill the stem cells, which are the cancer’s roots, it can grow back.

Cancerous stem cells are not the same as embryonic stem cells, the cells present early in development that can turn into any cell of the body. Cancerous stem cells are different. They can turn into tumor cells, and they are characterized by distinctive molecular markers.

The stem-cell hypothesis answered a longstanding question: does each cell in a tumor have the same ability to keep a cancer going? By one test the answer was no. When researchers transplanted tumor cells into a mouse that had no immune system, they found that not all of the cells could form tumors.

To take the work to the next step, researchers needed a good way to isolate the cancer-forming cells. Until recently, “the whole thing languished,” said Dr. John E. Dick, director of the stem cell biology program at the University of Toronto, because scientists did not have the molecular tools to investigate.

But when those tools emerged in the early 1990s, Dr. Dick found stem cells in acute myelogenous leukemia, a blood cancer. He reported that such cells made up just 1 percent of the leukemia cells and that those were the only ones that could form tumors in mice.

Yet Dr. Dick’s research, Dr. Wicha said, “was pretty much ignored.” Cancer researchers, he said, were not persuaded — and even if they had accepted the research — doubted that the results would hold for solid tumors, like those of the breast, colon, prostate or brain.

That changed in 1994, when Dr. Wicha and a colleague, Dr. Michael Clarke, who is now at Stanford, reported finding cancerous stem cells in breast cancer patients.

“The paper hit me like a bombshell,” said Robert Weinberg, a professor of biology at M.I.T. and a leader in cancer research. “To my mind, that is conceptually the most important paper in cancer over the past decade.”

Dr. Weinberg and others began pursuing the stem-cell hypothesis, and researchers now say they have found cancerous stem cells in cancers of the colon, head and neck, lung, prostate, brain, and pancreas.

Symposiums were held. Leading journals published paper after paper.

But difficult questions persisted. One problem, critics say, is that the math does not add up. The hypothesis only makes sense if a tiny fraction of cells in a tumor are stem cells, said Dr. Bert Vogelstein, a colon cancer researcher at Johns Hopkins who said he had not made up his mind on the validity of the hypothesis.

But some studies suggest that stem cells make up 10 percent or even 40 percent or 50 percent of tumor cells, at least by the molecular-marker criterion. If a treatment shrinks a tumor by 99 percent, as is often the case, and 10 percent of the tumor was stem cells, then the stem cells too must have been susceptible, Dr. Vogelstein says.

Critics also question the research on mice. The same cells that can give rise to a tumor if transplanted into one part of a mouse may not form a tumor elsewhere.

“A lot of things affect transplants,” Dr. Kern, the Johns Hopkins researcher, said, explaining that transplanting tumors into mice did not necessarily reveal whether there were stem cells.

Other doubts have been raised by Dr. Kornelia Polyak, a researcher at the Dana-Farber Cancer Institute. Dr. Polyak asked whether breast cancer cells remain true to type, that is, whether stem cells remain stem cells and whether others remain non-stem cells? The answer, she has found, is “not necessarily.”

Cancer cells instead appear to be moving targets, changing from stem cells to non-stem cells and back again. The discovery was unexpected because it had been thought that cell development went one way — from stem cell to tumor cell — and there was no going back.

“You want to kill all the cells in a tumor,” Dr. Polyak said. “Everyone assumes that currently-used drugs are not targeting stem cell populations, but that has not been proven.”

“To say you just have to kill the cancer stem cell is oversimplified,” she added. “It’s giving false hope.”

The criticisms make sense, Dr. Weinberg said. But he said he remained swayed by the stem cell hypothesis.

“There are a lot of unanswered questions, mind you,” he said. “Most believe cancer stem cells exist, but that doesn’t mean they exist. We believe it on the basis of rather fragmentary evidence, which I happen to believe in the aggregate is rather convincing.”

Dr. Wicha said he was convinced that the hypothesis was correct, and said it explained better than any other hypothesis what doctors and patients already know.

“Not only are some of the approaches we are using not getting us anywhere, but even the way we approve drugs is a bad model,” he said. Anti-cancer drugs, he noted, are approved if they shrink tumors even if they do not prolong life. It is the medical equivalent, he said, of mowing a dandelion field.

He said the moment of truth would come soon, with studies like the one planned for women with breast cancer.

The drug to be tested was developed by Merck to treat Alzheimer’s disease. It did not work on Alzheimer’s but it kills breast cancer stem cells in laboratory studies, Dr. Wicha says.

The study will start with a safety test on 30 women who have advanced breast cancer. Hopes are that it will be expanded to find out if the drug can prolong lives.

“Patient survival,” Dr. Wicha said, “is the ultimate endpoint.”