The global eradication of malaria is a challenging, although perhaps not impossible, aim: Paraguay was recently declared malaria-free and, according to the World Health Organization, Algeria, Argentina and Uzbekistan are likely to be declared malaria free by the end of the year. Several tactics are being used to prevent infection, including preventing mosquito bites by sleeping under insecticide-soaked nets, spraying and draining standing water to control mosquito numbers, and CRISPR techniques to generate malaria-resistant mosquitos are being tested. A number of antimalarial medications are available for treating infection and it is advised that malaria infection is confirmed before starting treatment in order to limit the spread of drug resistance. Early and accurate diagnosis of malaria is essential for the best treatment outcomes and diagnosis can be achieved in several ways. Microscopy is the gold standard, but requires trained and experienced personnel, takes days and can still be inadequate and inaccessible in many remote/poor areas. PCR is another lab-based technique that allows detection and identification of the Plasmodium species and is useful where there is no access to microscopy or the microscopy results are unclear, as well as in cases of mixed infection. However, the cost of these tests is prohibitive in many regions with endemic malaria. Rapid diagnostic tests, using finger-prick blood samples, are specifically designed for use in remote, poor areas, detect malaria antigens and provide results within 30 minutes. One difficulty, however, is the identification and treatment of the small number of infected but symptomless carriers of the disease, who, if left untreated, can provide a continuing source of infection. According to research presented recently at the annual meeting of the American Association of Tropical Medicine & Hygiene, we might soon have a fast and non-invasive way of detecting these symptomless carriers. People infected with malaria give off an aroma that is imperceptible to humans but is very attractive to mosquitos. A pilot study by Steve Lindsay at Durham University and colleagues has shown that it is possible to train dogs to detect the same aroma and to discriminate between malaria-infected and -uninfected people with a reasonable degree of diagnostic accuracy, recognizing socks worn by children with malaria about 70% of the time and socks worn by uninfected children about 90% of the time. The next stage is to test the dogs with people instead of just socks. If successful, the potential exists to use information learnt from the dogs to create a bioelectronic nose for malaria, in the same way that medical detection dogs are already being used to aid the development of such a device for the detection of cancer.