Two photographs decorating the walls of Phytotech Inc.'s new offices show employees in white suits and surgical masks working in fields of green plants. One picture is set in a vacant lot in New Jersey; the other shows the Chernobyl atomic power plant in the background.

The photos suggest a day when a simple and cheap method will be used to clean up soil contaminated by the residues of old smelters, metals finishers, nuclear weapons plants and scores of other industrial operations.

Phytotech, based in Monmouth Junction, N.J., is testing the use of plants to leach toxic heavy metals from soil. In the New Jersey photograph, mustard plants are pulling chromium from dirt that was carted in years ago from the site of an old chromate smelter. In the Ukraine, the mustard plants are leaching radioactive strontium and cesium from soil that was so heavily contaminated by the nuclear-reactor disaster there that crops grown in it are poisonous.

The little two-year-old company is attempting to exploit a technology called phytoremediation, ``phyto'' being from the Greek word for plant. The technology and its name were conceived by a Russian-born, U.S.-educated plant biologist, Ilya Raskin, at Rutgers University in New Brunswick, N.J.

Dr. Raskin, who came to the U.S. in 1976, recalls that in 1989, when he moved to Rutgers from DuPont Co., he encountered scientists with a Lawrenceville, N.J., start-up biotech company, Envirogen Inc. They were using bacteria and other microorganisms to clean up polluted soil. While bacteria are good at degrading oils and chemicals, ``they can't degrade heavy metals,'' Dr. Raskin says. But ``a lot of contaminated sites contain mixtures'' of hydrocarbons and heavy metals.

``I remembered reading Russian papers from the 1930s and 1940s about geobotany, in which they prospected for minerals by looking at plants,'' Dr. Raskin says. ``Some plants have a high capability of accumulating metals from the soil,'' thus providing clues to what lies just below the surface.

The idea that such specialized plants might be used to deliberately remove metals from soil wasn't new. But most research had involved wild plants that had managed to thrive on metal-soaked soils. For cleaning up polluted land, Dr. Raskin needed the kind of plant that farmers knew how to grow on a large scale. Moreover, the plant had to be able to take the metals in through its roots, and move them above ground, where the leaves and stalks could be harvested and destroyed.

``We screened hundreds of plants,'' Dr. Raskin recalls. The Brassicaceae family of herbs that includes mustard, broccoli, cabbage and rape showed some promise, however. Finally, the search narrowed down to the Indian mustard plant, Brassica juncea, grown throughout the world for its oilseed (and only to a small extent for making the condiment).

Greenhouse experiments showed that after two to three weeks of growth, the Indian mustard soaked up twice as much lead in its shoots as the tobacco plant and 50 times as much as the corn plant. But not all strains were equally adept at grabbing metals from the soil. Screening more than 100 Indian-mustard strains finally led to one that is particularly good at leaching out not only lead but cadmium, chromium, nickel, zinc and copper.

The experiments indicated that one planting of mustard could soak up as much as two metric tons of lead per hectare (2.5 acres), or six tons a year if three plantings could be squeezed in.

The economics of phytoremediation are still being worked out. Dr. Raskin and his colleagues estimate that using plants would cost $60,000 to $100,000 to clean an acre of soil to a depth of about 20 inches. In contrast it would cost $400,000 to dig up the same amount of soil and haul it to a hazardous-waste site, they estimate.

With New Jersey's blessing, Rutgers spun off Phytotech in 1993 to exploit phytoremediation. A former Envirogen executive, Burt D. Ensley, was brought in as chief executive officer. Earlier this year Phytotech raised more than $3 million in a private placement. The company, in turn, made a $1 million grant to Dr. Raskin's laboratory for basic research.

Until Phytotech completes a rooftop greenhouse of its own, its scientists are working side by side with Rutgers researchers in the sprawling set of green houses at the Rutgers agricultural college. The plants range from six-inch high sproutings of Indian mustard to 12-foot tall sunflowers growing on perforated plates suspended over tanks of running water.

The sunflowers' roots, Drs. Ensley and Raskin explain, are extremely efficient at soaking up metals but the plant isn't very good at moving the metals up to the stalk and leaves for harvest. So, while it may not be useful in polluted soils its mass of roots could be used to filter metals from running water, or rhizofiltration (from the Greek word for root). One of the few paying projects Phytotech has landed is a Department of Energy experiment using rhizofiltration to clean up water contaminated with uranium from a government atomic bomb facility in Ashtabula, Ohio.

Besides screening new plants for their metal-grabbing talents, the scientists are testing various chemical soil additives to boost the amount leached out by the plants.

Phytotech, meanwhile, is leaping at every chance it can get to test phytoremediation in the field, usually free of charge or at nominal cost. That is why it is growing a plot of Indian mustard within half a mile of the Chernobyl atomic reactor disaster. In England, a test plot of mustard is extracting cadmium from tailings of a lead mine dating back to Roman times.

The biggest challenge is a long-since-dismantled plant that for many years treated chromium ore for smelting. That plant left soil contaminated with cancer-causing chromium that has been distributed as landfill to 180 sites in New Jersey.