Department of Plant Pathology, University of Minnesota, St. Paul 55108
Two collections of urediniospores of Uromyces appendiculatus, each from a different commercial bean field, were characterized for associations of virulence among individuals within each collection. Four bean (Phaseolus vulgaris) lines with distinct, race-specific resistance to which virulence in each population was polymorphic were used to obtain measures of all six possible pairwise virulence associations for each collection. We inoculated one of the lines and collected urediniospores only from the segment of the population that was virulent on that line. This segment, when compared with nonselected collections from susceptible Pinto 111, gave a direct measure of degree of association as the change in frequency of virulence observed. Plants of the second bean line were inoculated in separate sets with both selected and unselected collections. Frequencies of virulence were estimated from the numbers of susceptible-type and resistant-type infections. Reciprocals of each pairing also were made. For collection P21, all virulences were significantly associated, either positively or negatively, except one pair (in one direction of selection only); whereas, for collection M5, all virulences were significantly associated. Virulence association in P21 was shown to be the result of predominance of phenotypes with certain combinations of virulence by inoculation of the four bean lines with 10 randomly chosen single-uredinial individuals. In support of this, a large random-mated F1 population derived from each collection showed much less virulence association, with the majority of pairs of virulences showing nonsignificant changes in virulence frequency after passage through the first line. Random mating also significantly changed virulence frequency from that of the original population in all instances. Changes were in both directions, suggesting either that virulences were not all recessive, or that heterozygote frequency was sometimes above and sometimes below the Hardy-Weinberg expectation in the field populations.