40 



this. Thus I found the following différences : A experiments 

 201— '212, 171 seedlings, of which 42 curved, i. e. 25% in 

 the second case, and 142 seedlings, of which 88 curved, 

 i. e. 58% in the flrst case. In the same way B experi- 

 ments 78—81, of 50 seedlings 19 curved i. e. 38 % in the 

 second case and of 50 seedlings 24 curved i. e. 48 o/o in 

 the first case. Probably this phenomenon is connected 

 with the much stronger nutation in the plane of the higher 

 percentage of curved seedlings. 



This différence could not be a source of error as ail 

 seedlings used in the experiments were in parallel positions. 



Preliminary treàtment of plants. 



Bach ') proved, that a stay of some hours at 4° C— 

 10° C. had a real influence on the magnitude of the pre- 

 sentation-time, even when the plants were stimulated 

 U< hou]s after being transferred to 20° C. On account of 

 this statement I at first attached great value to growing 

 the plants at as constant as possible a température. Later, 

 when it became évident, that the influence of small dif- 

 férences of température was quite lost during the first hour 

 of preliminary warming I did not exclude slight différen- 

 ces, taking care only, that the expérimental plants passed 

 the hours immediately preceding the experiment at a 

 température, differing from 20° C. by not more than a 

 few degrees. 



B. Sources of error, caused by expérimental conditions. 



PURITY OF THE ATMOSPHERE. 



Eichter'') has the merit of having traced in détail 

 the influence of impurities in the air upon physiological 

 processes and of having emphasised the fact that they 

 constitute a very dangerous source of error. 



1) Bach, 1907, p. 72, 76. 



2) 0. R i c h t e r, 1909 and 1906 and the literature quoted there. 



