S64 
Journal of A gricultural Research 
Vol. XXIV, No. 7 
case which can well illustrate this point. The two biologic forms differ by 
exactly 1.00^0.51 fx in the means of length and by 0.99 io.i6 yu in the 
means of width. The difference in lenglh was probably due to random 
sampling, as it was only 1.96 times greater than its probable error, 
whereas the difference in width appeared to be rather significant, because 
it exceeded its probable error by 6.18 times. 
Two numerical differences of unequal magnitude, on the other hand, 
may sometimes be of identical or nearly identical significance. Thus, 
for instance, the urediniospores of P. graminis tritici exceed those of P. 
graminis phleipraiensis by 8.45 ±0.22 tx in the means of length, but the 
difference in the means of width is only 2.91 ±0.08 ix. The difference in 
the means of length divided by its probable error is 38.38, and that of 
the width is 36.38. In general, the greater the numerical difference the 
greater its significance, although in certain cases the situation may be 
reversed. An excellent illustration of this phenomenon is furnished by 
the difference in the means of urediniospores of P. graminis tritici and P. 
graminis secalis. The means of the w'heat-rust spores were 5.27±o.24 fx 
longer and 2.60±0.08 fx wider than those of the urediniospores of the 
rye rust. But the difference in the means of length exceeded its prob¬ 
able e^or 21.92 times as compared with 32.50 times for the excess in the 
case of the difference in width. Thus the important factor in determin¬ 
ing the value and significance of differences is not so much the numerical 
deviation as the probable error of these differences. 
The comparative differences in spore dimensions of the various biologic 
forms are illustrated by photomicrographs and schematic drawings in 
Plates I and II. Figure 13 represents diagrammatically the apparent 
correlation between the means of the different forms, while figure 14 
shows the rather consistent parallelism in the means of the various spore 
types. 
Although the size of spores is specific for each biologic form, the spores 
are not rigidly fixed either in size or in shape. Resistant host plants 
and other adverse environmental conditions, affecting the normal devel¬ 
opment and vigor of the rust fungus, also affect the size and shape of the 
