Virus diseases of plants are most often transmitted by vectors, typically herbivorous insects that spread the pathogen as they feed on different plants. That vectors of disease can be affected by host plant infection status has been known for a number of years. Indeed, it is known that vectors can exhibit orientation preference, in which they preferentially land on infected plants, or feeding preference, in which they preferentially settle to feed on infected plants after landing, or neither, or both. The preference of an individual vector can also depend on whether it is currently infected. Clearly these responses will affect the dynamics of disease spread. However these observations have only just started to be integrated into mathematical models. The epidemiological and evolutionary consequences are therefore not properly understood; neither is whether the responses can be manipulated to the detriment of the pathogen. Taking the recent emergence of cassava brown streak as the biggest threat to food security in East Africa as just one example, it is clear that anything offering hope of low cost control of viral disease would be a timely contribution.