What precipitates malaria illness in an area of intense transmission? Greenwood, Marsh and Snow1 and Marsh2 hove discussed the hypothesis that the severity of malaria depends on the size of the inoculum, that is, on the number of sporozoites inoculated at one time3. According to these authors, this is suggested by the results of vector control trials in which parasite prevalence remained the some but episodes of disease were reduced. This is clearly on important observation, but Jo Lines and Jo Armstrong interpret it differently. They see the primary implication of these results as being that natural immunity to malaria must be largely strain specific, and only if there is strain specificity can inoculum size be important. Here they present arguments to support this view, and point out that, if there is strain specific immunity, then short-term evaluation of vector control is likely to overestimate its long term benefit.