1910] PENNINGTON—LONGITUDINAL COMPRESSION 259 
vulgaris). The experiments upon woody plants were conducted 
in the garden during the spring and summer; those upon her- 
baceous plants were conducted in the greenhouse during the winter 
and in the garden during the summer. 
METHODS OF EXPERIMENTATION 
After many preliminary trials, the methods given below were 
found to be most suitable. For methods other than these, descrip- 
tions are given in the account of the different experiments in which 
they were used. In general, a large series of plants of the same size 
and age and growing under the same external conditions was used 
for each experiment. The plants were numbered consecutively. 
The height and two diameters of each plant were measured, one 
diameter of the part to be under compression and the other of the 
part above the compressed region. In case a stem was not cylin- 
drical, the average of the greatest and the least diameter was taken 
as the diameter of the stem at that place. Strong cords were then 
looped about each stem at a suitable distance above the ground 
to serve as an attachment for the weights. The stems were pro- 
tected from mechanical injury by the cords by first wrapping 
pieces of soft cloth, commonly known as table matting, around 
each stem. Each stem was then firmly fastened in a vertical 
Position by tying it to stakes, not in contact with the plant, so 
that it could not bend or sway, and yet could be restrained in no 
way except by the downward pull of the weights. Weights were 
then suspended upon half of the plants, while the other half were left 
to serve as controls. All plants in a series thus treated were under 
exactly the same conditions, except that the lower part of each 
experimental plant was subjected to a compression caused by the 
downward pull of the weights. If the plants were practically 
€qual in size, alternate stems were weighted; but if they happened 
» be somewhat unequal, care was taken to see that the average 
Size of the experimental plants was equal to the average size of 
_the control plants. For every large experimental plant, an equally 
large control plant was chosen, and for every small experimental 
Plant, an equally small control plant was chosen. 
More weights were added from time to time according to the 
