WASHINGTON -- Philip Forsline stalks wild apple trees for a living. Just last month he returned from a mountain wilderness in Kazakhstan -- the probable evolutionary home of the fruit -- having bagged cuttings from trees that he hopes will solve some of the problems with apples.


In a forest far from civilization and prowled by apple-eating bears, for example, Forsline found "just an extraordinary apple, one we call a super elite." It was a wild tree indigenous to the area, not a cultivated variety, but the apples were big and "tasted great," and, in contrast to others in the area, the tree was unblemished by apple scab.


Apple scab, a fungal disease, can kill trees and makes the fruit's skin look so unappealing that most commercial growers spray many tons of pesticides to prevent it.


"If we could develop these into a commercial apple with genetic resistance to scab," Forsline said, "we could cut down on the use of pesticides."


Forsline is curator of the apple collection of the U.S. Department of Agriculture's Agricultural Research Service. Based at the agency's Plant Genetic Resources Unit in Geneva, New York, the collection includes 2,210 domestic varieties of apples, including many that were popular centuries ago and are kept today not only as living antiques but as repositories of potentially useful genes. Also in the collection are seeds from about 1,000 wild trees, undomesticated forms of the species collected in wilderness regions.


One of the best ways to find new genes, plant breeders know, is to go to the place where the species is thought to have evolved. This "center of origin," if not obliterated by agricultural or industrial development, often harbors relatives of the wild ancestors of domesticated specimens. They are close enough genetically to interbreed but often possess genes that were lost in domesticated varieties.


The Tien Shan mountains of eastern Kazakhstan may not be the Garden of Eden, but they are thought to be the first place on Earth where apples grew. That is because the greatest diversity in wild apple species exists in this area.


In 1989, after the advent of glasnost in the Soviet Union opened the region to outsiders, Forsline first went to Kazakhstan to search for lost apple genes. Some 400 apple trees are now growing at the Geneva collection from seeds collected on that trip. Only a few are old enough yet to bear fruit. Forsline and his group returned in 1993 and again this fall, timing the trips to the fruiting season.


After brief visits to domesticated orchards, he and his group headed off the roads and into the wilderness, armed with rifles because the local bears also are interested in apples.


"We wanted to go to remote places were nobody had collected before," Forsline said. "We found this one area that was fantastic. We found apples that tasted great and really almost had what we call commercial character," meaning they looked and tasted almost good enough to sell.


Forsline speculates the apples evolved to taste good because that draws bears, which help the tree species by dispersing its seeds.


Moreover, the area was at the same latitude as northern Minnesota, a region where winter lows can drop to minus 40 degrees. Descendants of these Kazakhstan trees might someday thrive in northerly latitudes of this country too cold for today's commercial varieties.


Of key importance, the apples on one tree were largely free of apple scab, even though similar trees in the area were infected.


Forsline clipped 25-centimeter-long twigs from the scab-free tree, wrapped them in paper and kept them cool all the way back to the United States.


Along with them came cuttings, or scions, from six other "super elite" trees. Among them were apple trees that grew in a semi-arid region, plants that could give rise to a new apple that can grow in near desert conditions.


Because apple twigs can harbor viruses and other agents of apple disease, they cannot immediately be planted anywhere. So the Kazakhstan scions are in quarantine at the Agricultural Research Service's facility in Beltsville, Maryland.


"They'll stay here for maybe three or four years while we test them to be sure we're not bringing anything into the country that we might regret," said Allan Stoner, a research leader at the National Germ Plasm Research Laboratory.


One day the Kazakhstan twigs may become trees in Forsline's Geneva, N.Y. collection. Most apple types in his collection are growing in orchards. Others are kept as seeds in a freezer, where they can survive for decades, and a few are stored as buds frozen in liquid nitrogen at minus 150 degrees Celsius.


Each specimen represents a unique combination of thousands of genes conferring traits such as flavor, fruit size, tree height, color, disease resistance and environmental tolerance.


Such gene banks or, more properly, germ plasm collections, are mainstays of modern agricultural research. As new plant diseases crop up or as breeders seek new varieties with better characteristics, they turn first to gene banks, looking for specimens that possess the desired trait. If a sample in the collection has the needed trait, however, it will usually be combined with undesirable traits. If it were possible to extract the gene or genes responsible for a single trait, it would be relatively easy to transfer them into a good commercial apple variety. But this is not yet feasible.


So researchers crossbreed the banked specimen with an existing variety with good traits and hope that chance regroupings of genes will produce a seed with the desired combination. To find out, they must plant the seeds and wait for the trees to mature.


Differences among trees arise because apples do not "breed true." This is why you can't plant seeds from a Red Delicious apple, for example, and expect to get a Red Delicious tree. Each seed contains a different combination of genes that almost certainly won't produce as good an apple.


The only way to get another Red Delicious is to clone an existing one -- to take a scion and root it. All the Red Delicious apples in the world were produced by trees cloned from a weed tree that sprouted on an Iowa farm in 1872, a random combination of genes from an unknown ancestor that the farmer discovered to be tasty.