W. F. Bewley 



161 



Probable error of the average percentage of diseased seedlings per box. 



Twelve boxes were each given 5 lbs. of infected soil per box, soaked 

 with water, sown with 100 seeds selected by sieving, and covered with 

 a layer of soil. The percentage germination of the seeds sown and the 

 percentage diseased of those which germinated was ascertained. The 

 seedlings were removed as soon as they became diseased in order to 

 eliminate the factor of lateral spread, and each experiment was brought 

 to a close at the end of 14 days. 



From the percentages set down below the probable error of the mean 

 of twelve results was calculated. 



Probable error of each result =0-67 



/29-88 



V lT = 



±1-102. 



Probable error of the mean = 



±1-102 

 V12 



= ±0-29. 



From the above results it will be seen that the percentage of diseased 

 seedlings per box varied from 42 to 48, while the mean was 44-8. This 

 makes the probable error of each result 1-1 and the probable error of the 

 mean 0-29. This result is typical of the results obtained from the whole 

 of the experiments in this paper. In no case was there a wide divergence 

 from the mean and the mean can therefore be taken as an adequate 

 measure of the effect of the particular "limiting factor" under in- 

 vestigation. 



Experiments were set up to test the effect of different methods of 

 making up the seed boxes upon the incidence of the disease. The standard 

 method was to weigh out 5 lbs. of soil into each box. Then the soil was 

 compacted by means of a builder's board, soaked with water, 100 seeds 



