50 



BULLETIN 934, U. S. DEPARTMENT OF AGRICULTURE. 



the horizontal line showing the location of the mean for experiment 

 68 are much larger below the means than above it in the left-hand 

 portion of the graph, while the reverse is true in the right-hand 

 portion. To this extent the relative activity of the strains in this 

 experiment agrees with the performance of the same strains in the 

 two jack-pine experiments, as shown by the solid line. It can not 

 be decided from an inspection of the graph whether there is a real 

 agreement, in view of the large accidental variation present. How- 

 ever, the correlation coefficient, 0.4460.079, five and one-half times 

 its probable error, indicates a considerable correlation, not as good 

 as was found for the Corticium strains, but sufficient to establish a 

 strong presumption that observed differences in activity of the dif- 



FIG. 16. Diagram showing the comparative virulence of 47 strains of PytMum debaryanum 

 in successive inoculation experiments on species of Pinus. The results in experiments 

 Nos. 66 and 67 (on Pinus bcmksiana) are shown by the solid line, the strains being 

 arranged from left to right in the order of descending virulence indicated by the number 

 of seedlings surviving in those experiments. The results from the use of the same 

 strains in experiment No. 68 (on Pinus resinosa) are shown by the broken line. Such 

 correlation as there is between the two curves < coefficient 0.45 0.08) goes to indicate 

 a real difference in virulence between the different strains. The strains indicated by 

 the underscored numbers are original strains, and those not underscored are reisolations 

 from the original strains in earlier inoculation experiments on pine seedlings. 



ferent strains in these inoculation experiments were in part actually 

 due to differences in the capacity of the strains. 



It has been suggested in the foregoing that the difficulty in demon- 

 strating constancy in the difference in virulence between the various 

 strains of PytMum debaryanum is due in part to the lack of such 

 extreme differences as were observed between the various Corticium 

 strains. Figure 13 shows the distribution of the different original 

 Pythium strains according to the virulence indicated in the three 

 inoculation experiments of figure 16 (application to autoclaved soil 

 at the time of sowing) . Each value plotted is based on the average 

 results in 15 pots. Of the strains used, 21 were from species of pine, 

 1 from spruce, 2 from potato tubers, 2 from fenugreek, 3 from sugar 

 beet, and 6 from soil direct. Despite the considerable number of 



