THE RHEOTROPISM OF ROOTS 181 
asuitable amount of moisture, if it is allowed to touch the sur- 
face of the water. The wooden bar is wholly unobjectionable if 
of white pine. I have always taken the precaution of immersing 
these wooden bars for twelve or more hours in running water 
before using. I have tested a score of species of seedlings, and 
_ have never been able to discover any disturbance of growth hav- 
_ ingits source in the wood. A heavy bar of glass, with a plane 
_ Surface for the seedlings might be substituted for the wood. 
: For many kinds of experiments in which roots are to be 
immersed in water, it is advisable, as here practiced, to allow 
the preparation to stand with the roots in water for twelve to 
twenty-four hours before beginning the experiment; for, with 
a the seedlings of many plants, a few or many roots undergo curves 
OF distortions when placed in water. In my experiments all roots 
sp ch had deviated from the vertical direction were removed 
_ before the Streaming of the water was begun. 
aS In most of the experiments here recorded, a flow of water was 
Stared: by revolving glass basins full of water on horizontal 
Kl ane ats and centrifuges. The centrifuges were turned for some 
“xperiments by water motors, and for others by an electric motor. 
engl nde a constant rate of revolution, so that with itthe — 
gee tof different velocities could be studied. The apparatus, 
“xcept the motor, is shown in fig. 2. Juel states that in a similar 
ms wing basin which he used, he found the current deviating from _ 
‘angential direction, due to the disturbing effect of the oppos- 
‘7k bare = To Overcome this difficulty he divided his large basin 
Within 2 Seine compartments by inserting glass dishes, one 
PS tec In the most of my experiments, I suspended 
ae alings along 
three diamete 
3 
os by observing the movement of particles 
non-divided revolving basins, and assured myself 
€ flow as ;. een 
ie flow as it met the roots was practically uniformly: 
but one diameter, while Juel suspended his 
: rs. There was more reason for irregular or 
ee ents in his basin, therefore, than in mine. I have use 
ubdivided into three concentric com partments, and others 2 
ee and obtained uniformly the same results with both. 
