SUMMARYIncidence of eyespot was assessed in different‐length sequences of spring‐sown wheat and barley in field experiments. Disease levels in the fourth crop were greater with single cereal systems than when the alternative crop was grown in the fourth year. The existence of host specificity as a possible cause of this phenomenon was examined in comparative pathogenicity studies of isolate collections obtained from differentially‐cropped plots from the same field.In two experiments, isolates from the third, fourth or fifth consecutive barley crops produced more disease on barley than on wheat and isolates from wheat caused more disease on wheat than did isolates from barley. Isolates from 4‐yr wheat and barley sequences showed significant specificity of isolates for their original hosts but in another experiment such specificity was not detected in isolates from 3‐ and 7‐yr sequences of wheat and barley. However, there was a significant reduction in virulence of isolates in the seventh compared with the third crop, when comparisons were done on the alternative rather than on the original host species. This suggests that extreme selection pressure for pathogen virulence on a particular host species may lead to impairment of patho