Mayaro district, island of Trinidad, at the beginning of the rainy season. Along the coast, which faces nearby Venezuela, a tender breeze is blowing from the ocean. Here, a dozen miles inland, there is no such refreshment. It's mid-morning, and already the humid air is a giant thumb pressed on this ragged clearing. The surrounding foliage seems to be advancing almost visibly. Every once in a while a crowded van, a Maxi-Taxi, hurtles past, upending the silence with blaring reggae gospel or jangly Hindu pop. Across the road, a cow tethered in high grass stands statue-like, dreaming of shade. On the near side, behind Roopchand Baschk's plywood-sided house, a team of laboratory-pale Americans is toiling in the sun.

Blame it on Candi Thomas, the grinning young woman with the reddening shoulders and the ten-pound toolbelt strapped to her waist. Thomas, a junior biology major at Penn State, was doing a work-study project with Mark Guiltinan, associate professor of plant molecular biology, when she heard about Roopchand Baschk. Her church back home in Coudersport, Pennsylvania, was trying to help Baschk find medical care for his ailing young son. Baschk, meanwhile, was struggling to tame a long-neglected seven-acre plot of cocoa trees, and make it yield enough cocoa beans to support his family.

Guiltinan, as luck would have it, is director of Penn State's cocoa biotechnology program. "We should be able to do something for this guy," Thomas remembers telling her boss.

In a way, they already were. For three years, Guiltinan had been perfecting a system for making better cocoa plants: lots of them, and quickly. But that was labwork. Guiltinan, a stocky, fresh-faced Californian with a gene for tinkering, liked the idea of helping a farmer directly. "Most of the stuff we do is so far from practical application," he says, "that this was a nice change."

It could also, he thought, be a good way to introduce Thomas and other undergraduates to international development work, "to get their feet wet." Too, Guiltinan, who studies cocoa in a place where the stuff grows only indoors, is always looking for field partners. And Trinidad is a place rich in cocoa history, home to one of the largest collections of cocoa varieties in the world.

There followed a series of lab meetings, and questions: What are we really good at? And, What does this farmer really need? The answers converged on a plan that was now, after six months of long-distance groundwork, being realized. "Plants Without Borders," Guiltinan dubbed the project. He, Thomas, and three other researchers from the Penn State cocoa program—research associate Siela Maximova, doctoral student Carter Miller, and undergrad Nick Willis—were helping Roopchand Baschk put up a greenhouse.

Theobroma cacao, the "food of the gods" in Linnaeus' 18th-century description, is a broad-leafed understory tree that originated, according to ethnobotanists, in the rainforest of the Peruvian Amazon. At maturity, three to four years, the tree produces bright football-shaped pods up to eight inches long—yellow and orange and many rich shades of red—which bear, tight-packed in a milky pulp, the knobby, bitter seeds that are the raw material for chocolate, one of the West's most widely craved food substances and, for Pennsylvania, a $5 billion industry.

Grown for production, as it is today across the humid tropics, T. cacao is a finicky species. Tender seedlings need adequate shade, and protection from wind and pounding rains. Outside its native habitat, the plant is prone to attack: Its enemies include the vividly named fungal diseases Witches' Broom, black pod, and frosty pod; the virus called swollen-shoot ("Get this one and they come in with chainsaws and gasoline," Guiltinan says); and local insect pests: mirids, which West African farmers are forced to battle with heavy agrochemicals, and the cocoa pod borer, bane of growers in Malaysia and Indonesia. Altogether, Guiltinan estimates, these varied scourges destroy some 40 percent of the global crop in a given year.

The fragility of the yield, and fluctuating demand, make for a volatile economy, a boom-and-bust pattern that is particularly hard on the small farmers—like Roopchand Baschk—who grow 70 percent of the world's cocoa. The result is evident in the landscape: There are abandoned cocoa farms scattered across Trinidad. "When cocoa prices rise people want to grow it," Guiltinan says. "Then the market gets flooded, prices drop, and people abandon the farms. Then demand increases again, and people want to grow it again."

In Trinidad, actually, things are comparatively stable. The island's cocoa has long been prized for its flavor, apparently the fortuitous outcome of a disastrous epidemic way back in 1727. At the time, most of Trinidad's cocoa trees were criollo, a delicate variety known for fine flavor. When these trees died back, specimens of the less distinctive but hardier forastero type were brought from South America. The two strains eventually crossed, resulting in the hybrid known as trinitario, which fetches a consistently high price on the world market.

Elsewhere, cocoa farmers have a harder time of it. In the Bahia region of Brazil, until recently one of the world's leading cocoa-producing areas, an epidemic of Witches Broom during the early 1990s caused a 60 percent drop in yields, turning Brazil into a net importer of cocoa. A similar disaster now looms in West Africa, where some 70 percent of the world's cocoa is currently grown. Pierre Tondje, a scientist from Cameroon who visited Guiltinan's lab in the fall of 1999, reported that a new strain of black pod disease from the Congo basin was destroying up to 80 percent of his country's crop. The disease had already spread into Ghana, Tondje said, and was moving west, toward Ivory Coast, at a rate of 200 kilometers per year.

"In Africa, our only control is chemicals," Tondje said. "But farmers can't afford them," especially since 1985, when the price of cocoa bottomed out on the world market, and African governments, urged to policies of liberalization by the World Bank and International Monetary Fund, dispatched with subsidies. "The problem is acute. Even though cocoa cultivation is well-established culturally—giving farmers long-term title to land that other crops do not—many farmers are now abandoning cocoa for other crops. Cocoa is no longer profitable."

By 1998, projections for cocoa futures were disturbing enough to occasion a high-level summit of chocolate-industry executives and researchers in Panama. There, concerns were raised about the possibility of a worldwide shortage, perhaps within the next decade. When the story hit the New York Times ("Chocoholics Take Note: Beloved Bean in Peril," read the front-page headline), "the emotional response was incredible," Guiltinan remembers. "I have friends in the industry who said they were shut down for a couple of weeks, just trying to calm fears." Guiltinan's program, aimed at speeding up propagation of improved plants, is part of the proposed solution.

"Cocoa is an old plant, but a new crop," says Carter Miller, an intense, fair-haired Ph.D. candidate in Guiltinan's lab. We're standing in a state-of-the-art greenhouse on the University Park campus, surrounded by potted six-foot-tall cocoa trees. The air in here is a balmy 80 degrees; the computer-controlled fogging system deposits a fine mist indiscriminately across leaves, forearms, and eyeglasses.

Cocoa was domesticated by the Olmec people of southern Mexico sometime before 1000 B.C. The Olmecs, and later the Mayans and Aztecs, used the beans both as currency and as the basis for a frothy ceremonial beverage—the first prepared chocolate. The initial European encounter took place when Columbus, on his last voyage, spotted a cargo of what he called "almonds" in the belly of a Mayan canoe. By the early 17th century, chocolate libations were being sipped and savored in the Baroque courts of Spain, Italy, and France. Soon the stuff caught on in solid form, and conquered the masses. The world's taste for chocolate has been growing ever since.

Apparently, however, there has not been much selective breeding done over the centuries. As a result, Guiltinan says, today's world population of cocoa trees, estimated at five to six billion, is dominated by low-yielding, disease-prone specimens. The variation in productivity is an agriculturalist's nightmare: a mere five percent of trees brings forth some 50 percent of the total annual crop.

Guiltinan's plan calls for finding that five percent, what he calls the "elite genotypes," and multiplying it. The finding part requires local knowledge: at Union Vale estate on the Caribbean island of St. Lucia, where he went in 1997 to gather specimens for his early trials, he spent a lot of time in the company of Eustis George, a 74-year-old field manager with 60 years experience on the farm. "He knows every tree," Guiltinan says. Taking flowers from the most productive specimens, Guiltinan brought them back to University Park and practiced cloning copies of those trees in the lab.

Cloning in its simplest form—cutting and replanting snippets from an established plant—has been a standard practice of plant breeders for hundreds of years. The rooted cutting, if it takes, results in a plant that is genetically identical to the parent. But tree species show up a weakness in this method. "Let me give you a short botany lesson," Miller says. He reaches in to grasp the slender trunk of the nearest tree. This vertical growth, he notes, is juvenile—the young plant's upward thrust, known as orthotropic. (Ortho = upright; tropic = turning.) "At a certain point, orthotropic growth will branch" into plagiotropic (plagio = side) growth: the horizontal, leaf-bearing limbs that fill out the tree. He runs his hand over the crown of broad leaves. "Ninety-eight percent of what you see in a mature tree is plagiotropic," he says.

Cuttings from established trees are taken from leaf-bearing—plagiotropic—branches. And the resulting plants, Miller says, are essentially plagiotropic: they lack the strong vertical spine and branching architecture of trees grown from seed. If not trained upward, these cuttings will grow without form, as bushes. More importantly, plagiotropic cuttings will not sink a taproot; and without this deep anchor they are susceptible to drought, wind, and other disturbances.

Guiltinan and his team employ an advanced form of cloning that eliminates this problem. Instead of cuttings, they take tissue samples from the tiny white cocoa flower, which buds directly from the tree's trunk. By removing the sterile stamens from immature flowers and placing them in culture, then "pulsing" them with hormones, they can induce the formation of embryos—without the business of pollination. Moved to the greenhouse, these cloned embryos are converted (not germinated, technically) into plantlets: genetic copies of the original plant which also retain the original's orthotropic, tree-like, structure.

When it works well, tissue-culturing produces large numbers of identical plants quickly. The culturing process may also "cleanse" the new generation of disease. "It isn't clear yet exactly how this works," Guiltinan says, "but our preliminary research suggests that when flower cells with fungal disease are put through this process, the fungus doesn't show up in the embryos. We think this may work for viruses, too." If it checks out, this curative function could be of great importance in places like Africa, where plant viruses run rampant and disease-free replacement stocks are hard to find.

Tissue-culturing could also serve as a vehicle for genetic enhancement of cocoa plants, Guiltinan notes. The tissue-culturing technique that Guiltinan uses, called somatic embryogenesis, isn't new. It was developed in the 1970s, he says. Since the early 1980s, it has been tried many times on cocoa. Until now, however, no one has been able to make it work. "People were getting a few embryos here and there," Guiltinan says, "but not much. And for some genotypes none at all." That was where things stood when he agreed to take the directorship of the Penn State cocoa program, switching over from research on apples. Coming in, Guiltinan looked closely at the experimental protocols—the step-by-step directions—that had been used for cocoa tissue-culture up to that point. Convinced that the fundamental approach was sound, he designed a series of experiments to tinker with the recipe.

"What we did was simple," Guiltinan says. "Very simple, but not easy to get to. There are so many components in this kind of protocol, so many different steps—the possible combinations are almost infinite. It would take forever to systematically check every variable." So he and his students homed in on the basics. "We did an experiment comparing the growth of cocoa cells on different media. What we found was that the medium people were using, something called MS—which is used for 99 percent of all tissue culturing— is almost toxic to cocoa cells. To get them to survive at all you have to keep transferring them from one dish to the next.

"So we changed media. We also changed hormones. And we tried lots of different cocoa genotypes, to see if some were better-suited than others." In time, they found the proper combination, one that would consistently yield a respectable number of embryos. The next step was conversion: turning those embryos to plantlets. Instead of being a trivial matter, as earlier researchers had assumed, conversion "turned out to be the hardest step. It took another whole series of experiments," Guiltinan says, shifting types and amounts and concentrations of ingredients.

At last, armed with a sturdy new protocol, Guiltinan and his team looked for ways to bump up their production efficiency. One way was by adding a second round of embryogenesis to the process. Once primary embryos are viable, Guiltinan explains, these embryos are redivided; their cotyledons, or seed leaves, each having cells with "embryogenic potenial," are removed and chopped up, and these fragments are again placed into culture. This second round, the researchers discovered, had an even higher rate of success than the first one.

Last May, at a meeting of the scientific advisory committee of the American Cocoa Research Institute, the chocolate-industry consortium that endows the Penn State program, research associate Siela Maximova presented the results of an experiment that demonstrated the effectiveness of the lab's efforts. In a year-long trial, she reported, senior research technologist Ann Young had successfully produced 4,000 cloned plants from a single flower.

Even honed to maximum efficiency, however, somatic embryogenesis by itself remains too expensive for commercial use, Guiltinan says. Tissue-culturing requires sophisticated equipment, high-priced chemicals, and a lot of painstaking labor. So he and his team started searching for a way to lower the cost, for "another, cheaper propagation technique that we could couple with ours." They found their answer in something called bentwood gardening. In this simple technique, a young sapling is bent at a 90-degree angle, so that its top half grows parallel to the ground; in other words, it is forced to grow plagiotropically. Defiantly, the tree sprouts new vertical shoots from the bend. These shoots can be harvested every few months and rooted. By this process, theoretically, a sapling can produce up to 200 new plants a year, Guiltinan estimates. And unlike standard cuttings, these shoots are orthotropic.

Again, "this was no great intellectual leap," he says. "Similar systems are used with coffee and apple. It's a natural thing for cocoa to propagate this way." But by joining bentwood gardening with somatic embryogenesis—producing tissue-culture clones in the lab, say, and then teaching farmers to take and multiply them via the bentwood technique—propagation rates could be amplified dramatically. A government or other controlling agency, Guiltinan estimates, could effect a large-scale tree-replacement program for pennies per plant.

First, however, the tissue-culture method had to be tested in the field. On the Union Vale estate in St. Lucia, the first set of clones cultured from the elite trees that Eustis George identified are now producing cocoa pods. Every few months, Guiltinan and his students return to check on these pioneers and to carry down new specimens. They plan to track these trees through full maturity, monitoring growth rates, root structures, and yields.

A similar field study is being conducted in Brazil, and Guiltinan and Maximova are simultaneously working to spread the word, presenting their protocol at meetings in Malaysia and Ghana, holding workshops in Costa Rica, Brazil, and Ecuador. Over the years the lab has hosted a stream of students and visiting scholars from cocoa-producing regions around the world.

One of these students, Abdoulaye Traore, is an Ivory Coast native who finished his Ph.D. in Guiltinan's lab in 1999. With funding from the American Cocoa Research Institute, Penn State's Office of International Programs, and USAID, Traore returned to West Africa with the goal of promoting modern propagation techniques in his country and Ghana, the world's top two cocoa producers. In less than a year, he told the ACRI committee at last May's meeting, he had made significant progress, scouting the resources available in both countries, gauging the political winds, and initiating partnerships. "I have found some facilities that are ready for tissue culture," he said, "and I have already identified and started tissue culture on some elite clones." But Traore also noted the persistence of some very basic obstacles, from the frequent power outages which contaminate whole trials to aging or absent equipment—"There are no dissecting scopes in Ghana"—to a shortage of running water.

The situation in Trinidad is not nearly that bad. Still, Roopchand Baschk's farm, 10 miles from the greenhouse site, looks like it could stand all the help it can get. The muddy access road to the leased plot is lined with huge hardwood logs, tagged for transport. The main section, roughly five acres, is densely forested, and so choked with vines and undergrowth (and snakes) that it is difficult to enter. Once you step through the first line of trees and your eyes adjust to the shade, your gaze quickly fixes on the orange and yellow cocoa pods hanging like lanterns. It takes a while longer, however, to make out their gnarled and overgrown trees.

"Here you see why so many farmers quit cocoa and start growing bananas," Siela Maximoxa says quietly, peering up at the close-knit canopy. There's no telling how long this place lay untended before Baschk took it over two years ago. But Baschk, a strapping man with a clear eye and a ready smile, seems undaunted. Back in the sunlight, squatting on his heels, he points down the road toward the front of his plot, a two-acre stand of corn. Here, he says, he will plant his first clones.

The greenhouse will give him an important advantage: the means to raise all the plants he needs. A visit to the government agricultural station at La Reunion, in the north central part of the island, had underscored the value of such independence. At La Reunion, a large, sprawling compound bordered with lush hedges and brilliant bougainvilleas, the agricultural ministry propagates citrus, avocado, and exotic fruits like pommerac, as well as cocoa. But the extensive, heavily shaded cocoa greenhouse area, except for a few dozen forlorn-looking plants in black plastic sacks, was curiously empty.

Clones are raised here, from rooted cuttings. The approach, as an officer explained, is somewhat different from what Guiltinan and Miller use back in University Park. Here, the starter bins are tightly closed under plastic and burlap in order to retain moisture. To keep things from overheating, the shade cover has to be extra-heavy. The combination makes for slow growth, and problems with mold. Forty percent of the cuttings are lost.

Those that survive are sold to farmers at a heavily subsidized price: one Trinidad dollar (TT), or about 15 cents American, although they cost about seven TT each to make. And the demand, from the looks of it, far exceeds the supply. In the course of an hour, three farmers came looking for clones, and were turned away.

At the smaller Rio Claro station, back in Mayaro, another officer explained that the government has purposely curtailed its propagation operation in order to encourage farmers to build their own facilities. "But the response has not been good," he said. In fact, there is little incentive for farmers to take this step, since they are paid a subsidy simply for working land planted in cocoa, no matter the type or quality. The result of this tangled policy, he admitted, is that Trinidad, once a world leader, no longer produces enough cocoa to satisfy its markets.

It is Guiltinan's hope that Baschk's small greenhouse might stir a change. Its design is open-air, allowing more light and requiring more water, the latter provided by a simple drip-irrigation system. The overall effect should mimic the climate in the highly mechanized University Park greenhouse, where Miller achieves propagation rates of 90 percent. If Baschk, using this facility and combining his own knowledge with Guiltinan's advice, can make a success of his farm, that success will undoubtedly be noticed, and perhaps imitated.

Indeed, the island's leading cocoa growers—they thus distinguish themselves from farmers—seemed receptive to Guiltinan's ideas. One of them, Philippe Agostini, runs a 400-acre estate in the north-central hills near Gran Couva that has been in his family for four generations. His premium cocoa—100 to 120 tons a year—is used exclusively in French Valhrona chocolate, widely recognized as one of the finest in the world. Agostini, whose grandfather completely replanted the estate in the 1930s and '40s to ward off Witches' Broom, and whose father did so in the 1960s after an outbreak of frosty pod, is replanting 10 to 20 acres per year to protect against Black Pod disease, the current nemesis. He propagates his own plants, and is interested in cloning technologies, he said, because he wants to preserve the quality his family has long been known for. "That's what we're selling here," he said. "A certain quality of flavor."

Agostini's friendly rival Paul Manikchand, an American-trained engineer, grows cocoa intensively on an orchard-like farm near Sangre Grande. On 33 acres, Manikchand has planted 75,000 trees; they flank his long driveway in perfect close-ordered rows. He claims yields of 1,600 to 1,800 pounds per acre, seven or eight times the national average.

Manikchand at first said he prefers ordinary seedlings to clones—but his reasons turned out to be pragmatic. Clones, he said, tend to be larger at planting time than seedlings, and so they cost him more in transport and labor than seedlings do. The clones he's seen, he added, are an uncertain improvement over seed plants. "I am a businessman," he concluded. "I'm not interested in anything that doesn't work. But if you could prove to me that you had a plant that was orthotropic and disease-resistant, and retained good flavor, I'd pay five dollars [TT] for it."

Back at the greenhouse site, there were delays. A check valve in the drip-irrigation line refused to function properly. Without it, the system's water source—a raised cistern into which Baschk funnels rainwater from his roof—would be drained off in a few days. To eliminate the now-troublesome force of gravity, Carter Miller spent most of a day digging a great round hole in the heavy clay, deep enough to fit the tank. Water would be coaxed forth, when needed, with a pump.

On another day, a hot, dusty scavenger hunt in the capital city of Port of Spain failed to produce any clamps suitable for anchoring the greenhouse's shade-cloth roof. That evening Guiltinan, Miller, Baschk, and others worked until midnight sawing and fashioning clamps from scrap aluminum.

In a week's time, however, the job was finished—just in time to avoid a general strike which shut down Mayaro district, closing all roads. The completed greenhouse included open propagation bins with misting sprinklers; cast-iron nursery tables; and a starter set of seven cloned plants, donated by the station at La Reunion. With the addition of a single plastic T-connector and a length of extra tubing, it even brought running water into the Baschk house. Importantly, the whole structure was built for less than a thousand dollars U.S.

On their last day in Trinidad, Guiltinan and his team toured the International Cocoa Genebank, a living collection of over 2,500 varieties of cocoa planted on a vast tract near the Port of Spain international airport. From the heat of mid-afternoon through cooling dusk and the onset of mosquitos, Miller and Guiltinan went from tree to tree, snipping and bagging and labeling tiny white flowers, and packing them in ice for the long trip back home. Miller and Maximova would spend the days after arrival dissecting these fragile samples and preparing them for cloning.

On the plane, Guiltinan, pleased with the trip's success, speculated on the possiblity of future Plants Without Borders projects—building greenhoues in Africa, say, or South America. "I could see doing two or three of these a year," he mused. "It would be a great thing to look back on some day."

Now, however, it was time to get back to the lab.

Mark J. Guiltinan, Ph.D., is associate professor of plant molecular biology in the College of Agricultural Sciences, 306 Wartik Laboratory, University Park, PA 16802; 814-863-7957; mjg9@psu.edu. Siela Maximova, Ph.D., is a research associate in horticulture. Carter Miller is a Ph.D. student in horticulture. Candice M. Thomas is a senior majoring in biology. For archived dispatches from "Plants Without Borders" in Trinidad, visit the Web at http://www.rps.psu.edu/trinidad/.