Relevance of randomizing host genotypes of wheat that are tested for leaf rust resistance in race nurseries

Summary

Ten wheat leaf rust (Puccinia recondita f.sp.tritici) race nurseries were planted in 1984. In five, each with a different race, no randomization of the 20 wheat genotypes was carried out. The remaining five race nurseries, also each with one of the same five races, were completely randomized. Comparison of the wheat genotype assesments was done using three procedures:

• i) Correlating (Pearson's and Spearman's r) the leaf area affected of the twenty wheat genotypes in the randomized and the non-randomized nurseries;

• ii) comparing the best five wheat genotypes in each race nursery;

• iii) comparing the W_i-indeces of wheat genotypes in the randomized and non-randomized nurseries. All three procedures showed excellent agreement between the randomized and the non-randomized nurseries

. The use of a spreader row was most likely the reason for the large similarity. The consequence is that randomization is not sufficiently relevant to justify the increase in risk of mistakes and the extra administration needed. The host genotype-pathogen race interaction was significant and rather large. All but two genotype-race combinations showed a susceptible infection type, which indicates that partial resistance of wheat to wheat leaf rust caused the major part of this interaction. The results which suggest a gene-for-gene relationship, support the \lsintegrated concept\rs of Parlevliet and Zadoks.