Growth ami variation in maize. 169 



Summary. 



Part I of this paper gives a description of the data used and 

 deals with the general growth curves and the constants of variation 

 for each series as a whole. The reader is refeired to pages IIH — 117 

 for a summar}' of the results of this portion of the paper. 



In Part II we have attempted, by the study of the growth of in- 

 dividual plants, to analyze the adult variation cun'e into its component 

 elements. Specifically, we have attempted to follow individual plants 

 and groups of plants, having the same relative size at one stage of 

 their' growth, through the remaining growth stages. We have en- 

 deavored to ascertain how such plants are distributed as to relative 

 size in the successive growth stages and to discover some reasons for, 

 or the laws governing the distributions. 



To study these questions it is necessary to have a measure of the 

 relative size of the plants at each growth stage. For this, each 

 distribution was divided into five equal parts or quintiles. In any 

 distribution the relatively small plants are in quintile I and the relati- 

 vely large ones in quintile V. Fig. .5 illustrates the exact meaning of 

 a quintile. 



The problem was first approached by studying the quintile distri- 

 bution of all the measurements, throughout the season, of a group of 

 plants starting in a given quintile. 



From this it has been shown that there is a strong tendency for 

 the plants to remain in or near the quintile in which they started. 

 As a measure of this tendency we used the root-mean-square deviation 

 of these observed distributions from the most probable distribution of 

 such measurements on the theory of chance. These root-mean-square 

 constants are given in tables 7 to 9 and are shown graphically in Fig. 8. 



From these constants it is seen that in every case (except one) 

 the deviation from the theoretical mean is very much greater than 

 would occur on the basis of chance. 



The deviation of the very small and the very large plants (quin- 

 tiles I and V^) are much greater than any of the others. Thus there 

 is a much more marked tendency for the extreme plants to remain in 

 the extreme classes than for the medium sized plants to remain in a 

 particular medium sized class. 



Tables 10 to 12 and Fig. 9 show, by means of the same con- 

 stants, that there is a similar, though less marked, tendency for the 



