164 



Pearl and Surface. 



tion. Ouh- oue plant was allowed to remain in each hill so that no 

 plant was crowded more than another. 



Finallj' it can be shown that the large and small plants were 

 evenl}^ distributed throughout the rows. There were no patches where 

 all the plants were very large and others where they were all small. 

 In tables 13, 14 and 15 the mean quintile position of each individual 

 plant is given together with the number of the plant. The plants were 

 numbered consecutivelj- from one end of the row to the other. From 

 these tables it is possible to follow the distribution of consecutive 

 plants. Some idea of the random nature of these distributions can be 

 obtained by making a frequency distribution of the mean quintile 

 positions of each group of ten consecutive plants. Such a distribution 

 from series A is given in table 22. The other series show similar 

 distributions. It will be remembered that all of the original sixty plants 

 did not complete the season so that some classes have less than ten 

 plants. 



Table 22. 



Frequency distributions of the mean quintile positions in 



each group of ten consecutive plants in series A. 



From this table it is seen that within each group of ten plants 

 the distribution is entirely random. We could hardly expect throws 

 of dice to show a more random distribution than that given in the 

 table if we take account of the difference in the probability of each 

 class. 



The fact that the position in the field has no effect upon the 

 size of the plant can also be seen by studying the distrilnition of the 

 plants as given in tables 13 to 15. A single example may be cited 



