Kill the mosquito and you kill the disease. That is the usual approach to controlling malaria. And if done properly, it works. The problem is that the insecticides employed to do the killing destroy lots of other things as well. An old dream of those who seek to eliminate malaria is thus a way of selectively killing only what transmits the parasite: mosquitoes of the genus Anopheles, most notably Anopheles gambiae. And that, more or less, is what is proposed by Nikolai Windbichler and Andrea Crisanti of Imperial College, London, in a paper in Nature Communications. They think they have worked out how to stop A. gambiae females being created in the first place. That would break the chain of transmission in two ways: immediately, because it is only females that drink blood and so pass the parasite on; and in the longer term because without females a population cannot reproduce. The researchers' trick is to engineer into the mosquitoes a gene for a protein called a homing endonuclease. These genes are peculiar, and are probably a type of genetic parasite. They cut particular sequences of DNA in a way that damages the chromosome such DNA is found in. In extreme cases, that destroys the chromosome. (In less extreme cases, the process of repair often copies the endonuclease gene into the repair site; hence the name "homing," and also the suspicion of parasitism.) Dr. Windbichler and Dr. Crisanti have found a homing endonuclease in a species of slime mould that, by a strange coincidence, cuts a sequence of DNA found repeatedly in the X chromosome of Anopheles gambiae, destroying the chromosome completely.