128 Pearl and Surface. 



the general form of these polygons is W- shaped with the outer wdngs 

 very much elongated. The fact that the values for the end quintiles 

 (i. e., for verj- large and very small plants) are so much greater than 

 the values for the middle quintiles is of some significance. As shown 

 in the figures, there is nothing in the theory of probability which would 

 tend to make the standard deviations for the ends of the series larger 

 than for the middle. Hence it is clear that the nou- random influence 

 affecting plants which are very small or very large at the beginning, 

 is much greater than that affecting the medium sized plants. 



Having established this fact it may next be inquired as to what 

 tins influence is which affects the extremely large or extremely 

 small plants. The first explanation that suggests itself is that of 

 nutrition. Thus one plant may lie in an exceptionally fertile spot of 

 soil. Its root system becomes developed early, and this extra amount 

 of food supply is continued during the whole season. Naturally under 

 such conditions the plant would continue to be large during the entire 

 season. Again plants lying in exceptionally poor spots of soil would 

 tend to remain small during the entire season. 



Opposed to this explanation is the fact that the field on which 

 this corn was gi'own is an extremely even piece of intervale ground. 

 There were no spots in which the soil appeared to be better or worse 

 than other places. The manure and fertilizer were evenly distributed 

 over the field and well mixed with the soil. Further, considering the 

 nature of the root system of a corn plant, it does not seem probable 

 that the above explanation would account for the gi'eat differences 

 observed in adjacent plants. Thus, as will be pointed out later, it often 

 happened that of adjacent plants only eighteen inches apart, one 

 remained very small throughout the season and the other remained very 

 large. Such constant differences can hardly be due entirely to tlic food 

 supply or to other environmental conditions. 



A second explanation might l)e found in genotypic or hereditary 

 differences in the growth factors of the different plants. Within the 

 past few years students of genetics have clearly shown that the 

 inheritance of quantitative morphological characters may be explained 

 upon a Mendelian factor basis. For example, Emerson and East ('13) 

 have shown that the final height of corn plants is in a large measure 

 determined by their germinal constitution. It is not inconceivable, 

 therefore, that the steps by which a plant reaches its final heigiit is 

 also determined in the same manner. It seems probable that the greater 



