Humans could be protected from dengue fever by infecting the mosquitoes carrying it with a parasite which halves their lifespan, say researchers.
Australian scientists, writing in the journal Science, found that Wolbachia bacteria spread well through laboratory-bred mosquitoes.
Only older mosquitoes pass on dengue - so killing them could cut disease.
Experts said it remained to be seen how well the bacteria would spread outside the laboratory.
The virus might also adapt to survive, they added.
Female Aedes aegypti mosquito, which spreads dengue and yellow fever
Many thousands of cases of dengue fever occur worldwide each year, mainly in warmer tropical countries.
The virus is passed to humans when mosquitoes carrying it feed on their blood, and while there have been efforts to eradicate them using insecticides, these have been fraught with problems, including the ability of the mosquito to become resistant to the chemicals used.
The potential of Wolbachia as a way of controlling mosquito populations has been suggested for some time, but the latest study offers hope - albeit under laboratory conditions - that it might work.
The researchers from the University of Queensland in Brisbane picked a strain of Wolbachia known to halve the lifespan of its host.
The mosquito which carries the dengue virus is not naturally susceptible to the bacteria, so the researchers adapted it to create a successful infection.
The bacteria can be passed from infected female to offspring, and even though the cost in terms of lifespan should mean that infected insects should die out, Wolbachia has another trick up its sleeve.
It makes subtle changes to infected males which mean they can only produce offspring with infected females.
As expected, the infection thrived in the laboratory population of mosquitoes, and halved their lifespan to just a few weeks.
This is potentially significant because, after a mosquito acquires the dengue virus by biting an infected animal or human, there is a period of incubation lasting from a week to three weeks before it can pass on the infection when biting.
This means that only mosquitoes older than this are likely to be dangerous to humans and even these are likely to die swiftly, reducing their ability to infect.
The researchers suggested that the parasite represented a potentially inexpensive way to tackle the problem, particularly in urban areas, where other methods of control were difficult.
Dr Andrew Read and Dr Matthew Thomas, specialists in infectious disease dynamics from the Pennsylvania State University in the US, said "substantial" reductions in disease transmission could occur, but there were still obstacles to success.
"Determining whether it can remove enough infectious mosquitoes will be a challenge," they wrote.
If the bacterial strain chosen was too virulent it would spread very slowly and large numbers of infected mosquitoes might need to be released, they said.
It was also possible that dengue virus strains would adapt to require a shorter incubation period, they said.