I9I9] 



HENDRICKS— TORS 10 X 



431 



frame are given. This value, in seconds, is obtained in each case 

 by observing the length of time required for a certain number of 

 oscillations (usually 50 or 100, occasionally 200), making usually 

 4 or 5 determinations in each case. The average diameters are 

 given. In some cases these were not measured directly, but were 

 obtained from the diameter curve. The calculated values of 

 rigidity {L-^T') and of coefficient of rigidity (L^T'D^) are also 



given. 



TABLE I 

 Flowering bean 



* Value obtained from diameter curve: points a, b, and c, fig. 5. 



In fig. 5 these data are presented in curves. In all cases 

 abscissae represent lengths. It will be observed that in the twist 

 curve there is relatively little slope at first, but that this slope 

 increases later; and that at the end of the curve there is a slight 

 but sudden and distinct drop. In other words, at first the inter- 

 node twists only a little as it grows; later it twists faster; and, 

 when the full length has been attained, there is a slight reverse 

 twist. Note in table I that reading no. 8 shows a slight decrease 

 in length. This was frequently observed at or near the end of 

 growth in other experiments. It is probably due, for the most part 

 at least, to a temporary shortening of the silk thread on account 

 of changes in humidity. It may mean also a slight coiling of the 

 internode. As it probably does not mean an actual shortening, 

 this reading is not plotted in the curve. Reading no. 9 is also 

 omitted from the twist curve. 



