Abstract: Four sets of combinations of two races and three races of p. recondita f sp. tritici were made by combining different races of 261, 362 and 370 an successively cultured respectively on resistant, susceptible and slow rusting varieties for seven generations to see if any variety has stabilizing effect.The results showed that the highly resistant varieties Lovrin 10 and Lov rin 13 had no stabilizing effect. Susceptible varieties 5389 and Dongfanghong 3 and slow rusting variety Jimai 3 had stabilizing effect on certain combinations of races, but had no such an effect on other combinations. This fact indicates that whether a variety has effect of stabilizing selection or no depends on specificity of interaction between host and pathogen.The competitive ability comparison between races in the mixture interacting with slow rusting and susceptible varieties showed that the aggressiveness of race 261 was the strongest, that of race 362 was the next andthat of race 370 was the weakest. Generally speaking, any variety will not have the stabilizing effect if it interacts with a mixture of races which has a great difference in aggressiveness between theraces.The results of using two differential hosts (6068 and Predgornar 2) to detect the frequency of occurrence of the relevant race in each generation of subculturing were different. This was due to the difference in aggressiveness of races. This case indicates that the apecificity of interaction between the host and pathogen is due to not the virulence but the aggressiveness of thepathogen.Curbing the shift of population of pathogen through diversification of slow rusting and resistant varieties and the effect of environmental factors on the study of stabilizing selection were "discussed.

Key words: Wheat, Puccinia recondita, Stabilizing selection