The Sagamihara glassworks is the repository of the highly secret, world beating techniques that have propelled Nikon to a dominant position in the manufacture of key production equipment.

This plant leads the world in making glass with uniquely powerful optical qualities that are vital in the production of semiconductors. Thanks to this glass, Nikon has come from nowhere in less then two decades to become the world's dominant manufacturer in so-called steppers.

Steppers play an important role in world manufacturing in the late twentieth century as blast furnaces did in the nineteenth. They are lithographic machines that perform the crucial function of printing circuit patterns on silicon wafers.

Lithographic-a printing process in which the image is rendered on a flat surface and treated to retain designs while the nonimage areas are treated to repel designs.

The printing is done by focusing an image of the circuit pattern onto a photosensitive material coated on each wafer. The image is formed by shining a light through a "mask"- a stencil- like metal replica of the circuit pattern- and is progressively reduced through a series of huge lenses. It is then reproduced on the silicon wafers.

Because the resolution power of each new generation of steppers is greater then the last, the semiconductor industry can print even finer lines on each new generation of chips. And of course, the finer a chip's lines, the greater its information- processing capacity.

Much of the stepper industry's growth in sales revenues comes from price hikes as each succeeding generation of machines offers greater performance.

As of 1998, a state-of -the-art stepper cost more then five million. This represented a five-fold increase over Nikon's first steppers in the early 1980's. Moreover, a next generation stepper, due to go to volume production in 2000, is expected to cost more then $8 million.

For the Japanese economy, the most striking benefit has been soaring exports. Nikon exports more then fifty percent of its steppers. Moreover, even those Nikon steppers that are installed in Japanese semiconductor plants contribute strongly, if indirectly, to Japans exports because they are engaged mainly in making semiconductors for export.

Nikon pay levels are well among the best in Japan. As of 1997 its workers earned on average Y- 404, 900 a month. That was about $3, 300 at the 1997 exchange rate. Workers received large biannual bonuses that brought their total income to around $55, 000 a year. Nikon's pay levels were comparable to those in Japan's world beating automobile industry.

The stepper industry's need for highly precise mechanics stems in part from the ultra demanding level of accuracy required in positioning silicon wafers for the lithography process. If a wafer's position is off by much more then .1 microns- little more then one -thousandth of a hair's breadth- the resulting chips may prove fatally flawed.

Microchips are three-dimensional devices. Like the floors in a tall building, layers of circuits are imposed one on top of another. If each succeeding layer is not precisely aligned over the one below, the circuit connections between the layers will not line up. The mechanical challenge is greatly compounded by the fact that steppers print the same tiny pattern repeatedly on each wafer.

A final benefit of Nikon's base in Japan was a plentiful supply of long- term capital. Nikon could count on its Mitsubishi connections to ensure that it had privileged access to the torrent of patient capital available from the Mitsubishi groups various banks and insurance companies.

Nikon's main rival, Canon, had already established an important beach-head in copiers and used this advantage to achieve leadership in the emerging laser printing business (Canon supplies the key enabling components in laser printers to companies like Hewlett- Packard).

When American semiconductor makers were under strong competitive pressure from fast expanding Japanese rivals in the mid- 1980s, a furious policy debate raged in Washington. For many in the Reagan administration, the appropriate response seemed clear: do nothing. Their position was famously encapsulated.

In a remark attributed to the top Reagan Administration Economic Advisor Michael Boskin: " Computer chips, wood chips, potato chips. What the difference? They're all chips."

Reaganites believed that the choice of which goods the United States should or should not produce was best left to the unfettered free market. If American companies found quicker, easier profits in making potato chips or wood chips then computer chips, then this is where the nation's best prospects for prosperity lay.

There were several flaws in this argument, of which the most obvious was that if the American economy was to remain one of the world's strongest, the United States needed not only profitable corporations but good jobs.

The fact that the profit downturn the American semiconductor makers had suffered stemmed from no fundamental slowdown in the industry's phenomenal growth rate, but rather from short- term oversupply problems created by the aggressive expansion of notably far sighted Japanese semiconductor companies.

In just ten years to 1996, worldwide sales of microchips multiplied five fold to total $132 billion.

Higashi Hiroshima- a medium sized Japanese town- has one of the world's largest and most advanced semiconductor fabrication plants. Measuring 400 metres long and about thirty metres high, this plant is located in a semirural setting on the outskirts of town.

The plant is operated by NEC, which ranks as the world's second largest semiconductor maker after Santa Clara-based Intel Corporation. For our purposes, NEC is a more interesting company to look at then Intel because it has based its success almost entirely on world leading manufacturing skills, Whereas Intel's growth has come from a so called market lock- in - its virtual monopoly in making chips for Wintel-standard companies.

In the summer of 1997, NEC had just finished building a new extension full of state-of-the-art manufacturing technology. The extension, code named A2, was expected to lead the world in mass producing chips with lines as thin as .25 microns- one quarter of one millionth of a metre.

This was an important advance on the semiconductor industry's previous mass production best .28 microns, and it paved the way for NEC to pioneer the production of massively powerful 256 megabyt memories.

Perhaps the most important thing about A2 is its price tag- a cool $630 million. It is a good example of the pattern for Japanese companies to make their largest investments at home- a pattern reflected in the fact that though NEC operates more semiconductor factories abroad then in Japan, fully two- thirds of all the investment capital it pours into its worldwide semiconductor business each year goes to domestic operations.

Clearly the Higashi Hiroshima plant is a classic example of manufacturing's ability to create superb First World jobs. Moreover, in common with other manufacturing operations, the plant has created an excellent range of jobs: nearly 80% of NEC's workers are mere high school graduates, a considerably higher proportion then in the Japanese workforce as a whole.

For a fifty-year-old worker, wages plus bonuses came to well over $60,000 a year.

The evidence of the 1990's is that the job prospects are better in semiconductors than in mushrooms - not to mention wood chips or potato chips.