Barrier crops have been used since the early 1950s as a cultural control strategy for reducing the spread of non-persistently transmitted aphid-borne viruses. Since then, this strategy has been investigated by several authors, resulting in a wide range of divergent conclusions on its effectiveness. In a series of field and laboratory experiments carried out in central Spain during 1995-1998, we investigated the efficacy and mode of action of various barrier crops for reducing the spread of Potato virus Y and Cucumber mosaic virus in pepper. The barriers acted as natural 'sinks' for non-persistent viruses and did not reduce the number of aphids landing in the protected crop, as suggested by some authors. There was a significant reduction in virus spread and an increase in yield in two of the four years of trials. Such differences in the effectiveness of the barrier crop strategy show that its efficacy depends on a series of factors such as the kind of virus spread pattern (monocyclic or polycyclic), the height of the barrier crop at the time of maximum risk of infection and the extent of competition between the barrier and the protected crop. Moreover, the barrier crop should not host any potential insect pest or pathogen able to damage the protected crop. It is concluded that use of barrier crops can be an effective crop management strategy to protect against virus infection, but only under specific circumstances.