The lack of refrigeration remains a vexing problem in vaccinating the world's poor, because drugs often lose their efficacy in the heat. One solution 200 years ago was to propagate the vaccine in orphan children. "It was a really fascinating idea," remarks Charles Arntzen, founder of the Biodesign Institute at Arizona State University--but not one to be implemented today. Instead Arntzen has hopes for vaccines that exist inside crops and could just be eaten.

Arntzen conceived the idea of edible vaccines in the 1990s and has since tried to realize it. He genetically engineered potatoes to produce a vaccine against the hepatitis B virus, which kills one million people every year. This past February he reported that in a trial of an edible vaccine, up to 60 percent of volunteers who ate raw chunks of the potato developed antibodies against the virus. The signs of immunity are "an excellent start," Arntzen says.

Even so, Arntzen and others in the field are abandoning food vaccines. Consumer fear that the modified fruits and vegetables could end up in grocery stores is one issue. The medical worry, though, is the dosage: as public health expert Jurrien Toonen of the Royal Tropical Institute in Amsterdam points out, a tomato or banana is never the same size, so the quantity of the vaccine could vary from one piece to the next.

Because of the dosage concern, Hilary Koprowski of Thomas Jefferson University, the discoverer of the live polio vaccine, discounts the use of raw plants for massive immunization, even for farm animals. "The edible vaccine should be given in capsules containing desiccated leaf extract," he says.

Others agree that shifting the strategy for edible vaccines from food to processed pills makes sense. Achieving uniform doses would be easier, whether the extracts come from potatoes, lettuce, corn or even tobacco leaves (if the nicotine and other alkaloids could be removed). "We can freeze and dry potato, pack it in gelatin capsules and make uniform dosages of the vaccine," Arntzen notes. And pills would be cheaper. The Biodesign Institute projects that 200 acres would produce enough hepatitis B antigen to immunize all the babies in the world, at a cost of $0.05 per dose, compared with $0.30, the lowest price of the current vaccine. Trials for edible vaccine pills will probably begin in four to five years.

"Edible vaccines offer great advantages, as they do not need cold for conservation," Toonen says. "But they are not enough to resolve the whole situation." Even if they prove their efficacy, he states, they would face the logistical problems of some countries that make it difficult to deliver even the simplest pharmaceuticals to places far from urban centers, where most children are vaccinated. And Arntzen is aware of groups that systematically oppose genetic engineering. But he is optimistic: "It is going to be hard to justify blocking genetically modified plants if we can document we are reducing infant mortality."