ALTHOUGH YOU WOULDN'T know it from a day on Little Pine, human beings are an acquired taste for mosquitoes. Of 2,500-plus species, only the females of a few dozen varieties feed on people. And even that is a recent development. Five thousand years ago, when humans were still scattered, mosquitoes fed primarily on deer and cattle. Over the millennia, however, people multiplied and migrated. We made jugs to store water and canals to carry it. Mosquitoes, which need still water to lay eggs, naturally followed.

A handful of lethal pathogens, meanwhile, learned to hitchhike, infiltrating the mosquito's salivary glands and multiplying in its gut. It was a brilliant move. Furnished with door-to-door transport between bloodstreams, diseases traveled far and wide, crossing species and taking hold wherever man's footprint gave them access. West Nile, a bird affliction first discovered in 1937, in an area of Uganda near the Nile Valley, soon colonized people—picked up and passed on in two bites of a mosquito. In Africa, dam building helped spread malaria and Rift Valley fever, while rainforest logging in South America brought species that once bred only in sunny canopies down to the forest floor, to bite people's ankles.

The effects have been staggering, particularly where malaria is concerned. The species that carries the parasite, Anopheles gambiae, is a prolific breeder, capable of laying hundreds of eggs in places as hard to target as the muddy water filling a hoofprint or the ounce of rainfall left cupped in a plant. As a result, they also adapt quickly to new threats. Even DDT has begun to lose its punch. Malaria parasites, meanwhile, have become drug-resistant in areas throughout Africa, Asia, and South America, rendering chloroquine and other preventives less effective. Faced with these facts, the mosquito wars recently underwent a profound ideological shift: Instead of trying to eradicate the insects, researchers began trying to cure them.

This strategy is driven by a strange paradox of mosquito-borne illnesses: that for all their virulence, diseases like malaria and dengue fever rely on a system of transmission so improbable and Rube Goldbergian, it's amazing they exist at all. Of the millions of anopheles exposed to malaria, for instance, the vast majority successfully fight off the parasite, the same way a person might shrug off the flu. "If you go to a place with a high malaria rate, the number of infected mosquitoes is surprisingly low—a few percent," says Marcelo Jacobs-Lorena, a geneticist who studies malaria at the new Johns Hopkins Malaria Research Institute, in Baltimore. Those insects—the tiny percentage whose immune systems aren't up to snuff—are the ones that spread the disease.

And even that transmission is a crapshoot. The malaria parasite needs a full ten days to reproduce inside a mosquito's stomach—a time frame that falls uncomfortably close to the anopheles's life span. The vast majority of mosquitoes, moreover, die well before then: eaten, drowned, swatted, or crushed by spiders, fish, carnivorous plants, and people. Overall, just three or four infected mosquitoes out of a hundred live long enough to bite a new victim.

Given these odds, Jacobs-Lorena theorized that even a small tweak might be enough to interrupt the infectious cycle. "You have to realize that thousands of different parasites have probably tried to evolve to be carried by mosquitoes," he notes. "Most failed. Finally, after millions of tries, malaria and a few others managed it."

A courtly and patient researcher, Jacobs-Lorena has spent 17 years studying mosquito diseases, 14 of them in relative obscurity at Case Western Reserve University, in Cleveland. Then, in 2003, he rather suddenly found himself being wooed by Johns Hopkins, which had just received a large, anonymous donation earmarked for malaria research. (Prospects in this line of work have also been enormously brightened by another donor, the Bill and Melinda Gates Foundation, which has given hundreds of millions of dollars to malaria research over the past seven years.)

For Jacobs-Lorena, meanwhile, the work at Johns Hopkins intensified. He'd already identified the peptide that protects mosquitoes against malaria, and also synthesized the gene that produces the peptide. But then things turned complicated. While it's easy to endow mosquitoes with a protective peptide gene in the lab, it's far harder to do in the wild. "How do you spread the gene in nature?" Jacobs-Lorena asks.

The problem proved intractable enough for him to switch tacks. Rather than trying to cure mosquitoes with gene therapy, he's experimenting with something like antimalarial medicine for them: an innocuous transgenic bacteria that they can feed on in the wild.

Whether or not it works, interrupting a complex system like malaria with genetics is a bit like damming a river with pebbles. There are a lot of points where something can slip through and perhaps make things even worse. According to David Smith, a mathematical epidemiologist with the National Institutes of Health, reducing the number of infected mosquitoes, while good in the short term, can actually be dangerous, because it causes people to lose the immunity they develop from constant exposure. "As long as the reduction holds, death rates will go down," Smith says. "But if malarial mosquitoes rebound, the number of people dying from malaria could skyrocket."

The upshot, Smith believes, is that "once we start to intervene, we may have to be truly committed. Forever."