642 Ganong. — An Undescribed Thermometric Movement of 
Fig. 57. In Linder a, however, while there is agreement in some places, 
there is a wide deviation in others, so it becomes plain that either my 
figures are in error or else this method is worthless. There is, however, 
this difference between the two plants, that as the Linder a dries it loses 
some of its buds and bud-scales, while the Salix does not, and my method 
of drying the stems did not originally allow for this possible source of error. 
Despite the lack of agreement in the Linder a, however, I believe that the 
testimony of the Salix , and also that afforded by the gradually increasing 
curvature of all branches as they lose water, indicates a fundamental fact, 
namely, that the movement is connected with the amount of water in the 
stem, and that this amount of water is dependent upon temperature. 
This conclusion involves two further questions: (1) by what mechanical 
method does the increased amount of water produce the movement ; and 
(2) by what method does the variation in temperature produce a variation 
in the amount of water? We consider first the former, for which there are 
two possible explanations : ( a ) the weight of an added quantity of the 
water will tend to depress the obliquely-ascending branches and may thus 
produce the outward movement ; (b) the added water may permit of the 
larger absorption by the various cells of the younger branches and their 
consequent swelling, whereby the straightening of the stem must result, 
precisely as any flaccid tissue straightens with more abundant water- 
content. 
I have carefully tested both of these possibilities. As to the first 
I have repeatedly placed branches horizontal, and forced water into them 
both under an atmosphere of mercury, and also under the greater pressure 
of the water directly from a water tap. In such cases the water would 
be forced out in a few minutes from any injury incidentally or purposely 
made near the tip of the stem, showing that the water penetrated to the 
end. In such a case a distinct depression of the branch can usually be 
measured, but it is never of an amount as great as the natural amplitude 
of the movement in the branch attached to the plant. Furthermore, this 
small amount of movement occurs in the most favourable possible position 
(horizontal) of the branch, and would be much less when the branch is 
partially upright upon the tree. When the branches are placed upright, 
and the water is then forced into them, there is very little, if any, measurable 
movement. There is yet another consideration which shows that it cannot 
be the weight of the water which causes the outward movement, namely, 
that in many of the shrubs which showed marked movement of the branches 
the latter are nearly vertical as a whole, and hence the weight of the water 
cannot act to move them outward. The weight of the water, therefore, may 
aid the movement somewhat, but it cannot be the principal factor in 
causing it. 
We turn now to the other explanation, that an added supply of water 
