638 Gcmong. — An Undescribed Thermometric Movement of 
probability, and in order to obtain a definite basis for experiment we 
assumed that the secondary fluctuations were simply outward movements 
from the seasonal position. As a physical (or mechanical) cause of this 
outward movement under higher temperature, it seemed to us likely that 
the warming up, and consequent swelling, of the inner faces of the long 
slender branches under the influence of the sunlight on the warmer days 
was sufficient. Evidently this hypothesis could be submitted to experi- 
ment, for not only ought the outward movement to be greater on a sunny 
than upon a cloudy day of approximately the same temperature, but the 
movement in branches illuminated at the time of measurement on their 
inner faces should show more movement than those at that time shaded, or, 
still better, than those illuminated upon their outer faces. Simple as such a 
test appears the weather never allowed us to put it to satisfactory use, and 
the season closed without its accomplishment. 
The following winter, 1 900-1, I was occupied with other matters and 
did nothing with this subject ; but the next year, 1901-2, I resumed the 
study. Influenced by the theory above mentioned, 
I prepared to make more exact measurements than 
before of the respective movements of the four 
branches, for it was evident the theory could be 
tested by observing whether, as it requires, the 
greatest amplitude of movement occurs in the north 
branches, the next greatest in that east or west 
branch which happened to be illuminated on its inner 
face, and the least in the south branches. I made 
an improvement in Miss Persons’s method by re- 
placing the single gas pipe, which would yield a little 
under tension when the tape was drawn tight, by 
a perfectly firm tripod, formed of three gas pipes 
driven deeply into the ground, and bound immovably 
at their tops by twisted copper wire into which the 
brass screw was set, an arrangement illustrated 
diagrammatically in Fig. 56. Throughout the winter 
very careful measurements were made of six shrubs, 
including the Linder a and Cercidiphyllum used the previous winter, together 
with two species of Salix and two species of Populus (young trees). The 
results need not here be given in detail, since in general they are simply 
confirmatory of those earlier obtained. As to the two main points at issue 
they were as follows : — 
1. There was no such regularity or order in the amplitudes of move- 
ment of the respective branches as the theory required. 
2. There was no regular influence produced upon the movements by 
the presence or absence of direct sunlight upon the faces of the branches, 
