A quantitative assessment of the susceptibility of 131 cultivars of Antirrhinum majus L. to the rust fungus Puccinia antirrhini Dietel and Holway by means of replicated trials over two years revealed a wide range of general resistance; there is scope for improvement of rust-resistance by breeding. The loss of immunity was due to genetic change in the fungus because variations in pathogenicity occur among geographical isolates. An investigation of the epidemiology of the rust revealed that during the summer in Britain, uredospores are liberated through the day to germinate and establish infection after dew has fallen in the evening. The local dispersal of the pathogen is largely by wind but the spread between continents throughout the world is more likely to be due to human activity. A breeding programme is suggested for the improvement of A. majus by hybridizing the more resistant cultivars. Meanwhile, some horticulturally acceptable varieties with rate-reducing resistance to the fungus are recommended. These varieties should not become disfigured by the disease provided they are grown away from susceptible varieties.