FREDERICK, Md. (AP) _ Researcher Marcia Gray-Goodrich spends her working hours testing a simple theory: that somewhere in the world's forests, jungles or seas lies the cure to cancer.

At a National Cancer Institute laboratory housed in a converted bacteriological warfare center, Ms. Gray-Goodrich and scientists like her expose extracts of plants and marine life to the deadliest forms of cancer.

Of tens of thousands of specimens, only a few will show enough cancer- killing potential to be tested in rats and mice. An even smaller percentage will be promising enough to be tested in humans.

''It is frustrating at times but the long-term rewards are what we're looking for,'' Ms. Gray-Goodrich said.

Rewards like taxol, a drug derived from the bark of the yew tree that has been effective in treating advanced ovarian cancer.

Or CPT11, a drug developed in Japan from an Indian plant called camptothescin. The drug, now undergoing clinical trials in the United States, represents a new class of anti-cancer compounds that prevents malignant cells from dividing.

But those discoveries are rare. Most of the exotic extracts tested at the NCI lab show little or no effect in halting cancer's growth.

''Some of these drugs will actually cause the cells to grow better,'' said Ms. Gray-Goodrich.

Ms. Gray-Goodrich has no idea what substance she's testing. That's to prevent a lab worker's bias from creeping into the process. And she often doesn't hear right away if a drug she's identified continues to show promise in subsequent testing.

''We sometimes don't get instant gratification,'' she said. ''It could be months down the line.''

Still, she's cheerful and energetic, even when she has a migraine, and remembers to tell someone to ''have a nice day'' as she slides a bottle of an amber liquid into a special cabinet. The liquid is cancer in its purest form.

''You have to maintain an optimism in this business,'' says Dr. Michael Boyd, who oversees the NCI lab.

In the two decades that followed the signing of the National Cancer Act in 1971, the institute has tested some 150,000 new compounds, both natural and synthetic. Of that only 150 showed enough promise to be tested in humans. Thirty new drugs were the result. That means just 0.02 percent of the compounds tested wind up as drugs.

Similar work is going on in labs around the world.

The NCI lab at Fort Detrick, Md., is focusing its research on nature.

''Nature produces chemicals of a complexity which no chemist is going to dream of,'' says Dr. Gordon Cragg, chief of NCI's natural products branch.

So divers explore the ocean, and botanists comb the forests.

''This is still very much a shot in the dark,'' said botanist Jim Miller of Missouri Botanical Garden in St. Louis, one of three organizations under contract to pick up plants and send them to NCI.

Miller and his colleagues engage in what he calls ''chemical prospecting'' in the jungles of Africa and Madagascar. The pay is poor, the conditions are lousy, and the work has its hazards.

Miller said he often comes back from his trips sick, once discovering that what he thought was a bout of mononucleosis was really a case of five intestinal parasites. Others have come home with typhoid, one fellow almost died of malaria.

When asked whether he thinks there will be a cure found, Miller replied:

''I don't know. It's pretty easy to be convinced that it's out there somewhere. Obviously I hope I am the one to pick it up.''

So far, the most productive method of finding new cancer drugs is one the scientists euphemistically call ''serendipity.'' The drugs are found by accident.

''I don't think we'll find a cure,'' Cragg said. ''But I think we're going to find new drugs which are effective in causing some partial remissions of cancers.''

''We hate to talk about cures,'' he said. ''A cure implies that you literally wiped that disease out totally.''

Still, the prospect of finding a cure is what keeps Ms. Gray-Goodrich going.

''Somebody's going to do it,'' she said.