BEHAVIOUR OF PLANTS IN MASSES 315 



Outside row 

 Second row 

 Third row 

 Centre plants . 



Though branch-interference has so small an effect on the 

 total growth of the plant, it has a very great effect on the 

 character of that growth : the mustard plants in the above 

 experiments (p. 314) gave relative weights of 



100 114 and 124 



where the distances were 4, 6 and 9 inches, respectively; yet 

 the heights of the plants, as determined by measurements 

 throughout their life, were in the reverse order 



100 80 and 70, 



indicating a considerable drawing up of the plant by crowding : 

 but this drawing up was not accompanied by any appearance 

 of weediness, as may be judged from Fig. 45, so that any one 

 estimating the results by appearance only would not have 

 hesitated in saying that the more crowded plants were the 

 most flourishing a good instance of the impossibility of esti- 

 mating results by appearance : and such instances are con- 

 tinually occurring in the course of pot experiments. 



Having proved that root-interference produces no effect 

 when the same number of plants are grown in the same weight 

 of soil, the next point to be examined was the question of root- 

 competition when the weight of soil is the same, but the number 

 of plants in it is altered. For this purpose the undivided 

 pots were taken, and various numbers of tobacco plants up to 

 16, and of mustard plants up to 64, were grown in different 

 pots. The results showed that the total growth obtained in a 

 limited and insufficient weight of soil was nearly independent of 

 the number of plants : with tobacco constancy prevailed with 

 from about 3 plants (nearly so from i plant) up to 16 plants 



1 The outside rows in these series were grown in ordinary flowerjpots, 

 and were not included in the experiment. 



