Field Experiments. 



133 



The percentage of bunt varied from 43 to 100 where the seed was infected, 

 and only in one instance did the uninfected seed show any trace of the disease. 

 Medeah, which was bunt-proof at Dookie in 1908, turned out to be quite 

 susceptible witli us in 1909, having 4(3.6 per cent. 



Results of Ixfectiox without Treatment. 



It has already been shown in Table VII. that wheat may be naturally 

 infected with the spores of bunt so as to yield 88 per cent, of infected plants, 

 but when artificially infected it may even reach 100 per cent. This severe 

 infection is employed both in testing the capacity of different varieties for 

 withstanding the bunt as well as the effect of different kinds of treatment, 

 both before and after infection. In the following table it is shown that both 

 T. levis and T. tr itici msij Tpxoduce complete infection, and that when a smut- 

 ball is sown along with each uninfected grain there may be as much as 88 per 

 cent, of diseased plants produced : — 



Table X. — Results op Infection without Treatment. 



In two of the plots of Federation wheat, where the grain was sown with- 

 out artiHcial infection, there was 8 per cent, of infected plants, so that the 

 wheat to begin wdth was not perfectly clean. 



Plots 70 and 71 were specially sown, and the grain infected witli T. levis 

 and T. tritiei respectively, in order to compare the resulting growth. It is 

 stated at p. 55 that Harwood found T. levis to have the effect of shortening 

 the straw and stunting the growth, while no such, effect was produced by T. 

 tritiei, but in this instance the very reverse was the case. The plants in- 

 fected with T. levis averaged 3 feet in height, but with T. tritiei they were 

 only 2J feet. No striking difference was generally noticeable, however, bet- 

 ween the infected and uninfected plots, as it is usually when the grain is 

 forming and not while the plant is growing tliat the effect of the parasite 

 begins to be outwardlv visible. 



