10 



Harold Thurston Cook 



field, the circumference of the bulbs of twelve diseased and twelve 

 healthy plants was measured. This was accomplished without disturb- 

 ing the plants, by carefully scraping a little of the soil away from the 

 bulbs, which were already half exposed, a natural condition on muck 

 soil. The same bulbs were measured a second time on September 1, 

 four weeks later. Unfortunately, by that time even the plants which 

 were healthy at the first measuring had become diseased, and so it was 

 possible only to compare early-infected plants with those which became 

 infected late. The results of these measurements are given in table 2. 



TABLE 2. Effect of Downy Mildew of Onions on Increase in Size of the 



Bulbs* 



The increase in the circumference of the bulbs of late-infected plants over that of early-infected plants 

 is 1.386 ±0.15. The odds that the difference is significant are greater than 9999 to 1. 



* The probable error was calculated by Bessel's formula, P. E. = — — ■ — '- — . This formula 



Vn-1 

 may not give the true probable error, since the average size of the bulbs in the 

 late-infected group was slightly larger on August 4 than that of the early-infected 

 group. Therefore the probable error of the first four pairs in the table, which were_ of 

 the same size on August 4, was calculated according to Student's method for interpreting 

 paired experiments described by Love and Brunson (1924). By this method the mean 

 difference of the increase in size of the bulbs of the late-infected plants over that of the 

 early-infected plants is 1.75 inches. The standard of deviation is 0.77, and the odds that 

 the difference is significant are 65.2 to 1. 



t This plant was missing on September 1. 



The figures given in table 2 show that there is a very marked dif- 

 ference in the rate of growth between the bulbs of plants infected early 

 and those of plants infected late. It is reasonable to suppose that the 

 difference would have been even greater if the bulbs of the diseased 

 plants had been measured earlier and the healthy ones had not become 

 infected. 



