COMPARISONS OF ISOLATES OF THE TAKE-ALL FUNGUS, GAEUMANNOMYCES-GRAMINIS VAR TRITICI, FROM DIFFERENT CEREAL SEQUENCES USING DNA PROBES AND NONMOLECULAR METHODS

A mitochondrial ribosomal DNA probe (pEG34) that distinguishes two mai n restriction fragment length polymorphism (RFLP) types of isolate of the take-all fungus Gaeumannomyces graminis var. tritici was used to c ompare isolates from different host species. RFLP analysis confirmed t hat one type (designated T2), unlike the other type (T1), was associat ed more with barley than with triticale or wheat. RFLP analysis and ot her methods were also used to compare isolates of the take-all fungus from wheat seedlings grown in soil from first and second wheat crops. from wheat crops grown in longer sequences, and from oat crops grown a s breaks from susceptible cereals. These provided samples from differe nt stages of disease build up and decline in the take-all epidemic. Di fferences in frequencies of some characteristics were found at convent ional significance levels (P less than or equal to 0.05): in a cereal sequence experiment at Rothamsted, U.K., isolates containing dsRNA wer e less frequent in T2 isolates than T1 isolates except in second wheat s; perithecia were produced more abundantly on roots of wheat seedling s inoculated with T1 isolates than with T2 isolates, regardless of the ir origin. A weaker association (P less than or equal to 0.1) was also found: T2 isolates were often more frequent from second wheat crops, before take-all peaked in wheat sequences, than from other wheat crops or oats break crops. Collectively, the findings add only modest suppo rt to the controversial hypothesis that changes in the population of t he take-all fungus occur during repeated cropping of wheat. The limita tions of traditional sampling and assessment methods in such studies a re discussed. (C) 1997 Elsevier Science Ltd.