the Branches in Shrubs and Trees . 639 
though in some individual cases this did appear to have some slight 
effect. - 
It occurred to me during the winter, especially when it became plain 
that the direct sunlight played little part, that perhaps the movement might 
be due to a warming, and hence swelling, of the inner faces of all the 
branches through a general warming up of the air among the branches of 
the shrub due to the reflection of the sun’s heat from one branch to another. 
To test this I placed very accurate thermometers, reading precisely alike 
and graduated to tenths of a degree, both near the centre of the shrub (but 
in the sun), and outside the shrub a few feet away. They showed that the 
temperature among the branches and that outside the shrub were not 
appreciably different, thus eliminating another possible cause of the 
movement. 
The idea that a direct action of the sun upon the plant produced the 
movement had therefore to be abandoned. 
The following winter, 1903-3, I continued the study, concentrating 
attention upon two plants, Lmdera Benzoin and a species of Salix , which 
had shown themselves particularly sensitive to temperature changes. In- 
cidentally I re-measured these two shrubs very carefully through the winter, 
and the results for Linder a are given on Fig. 57 , not because they bring 
out anything new, but because they show with particular clearness the 
correlation of movement with temperature. But the principal work during 
the winter was experimental, and directed to discover the precise physical 
basis of the movement. Its results were as follows. Certain observations 
made while measuring the shrubs seemed to render it probable that the 
outward movement was caused by the straightening of the curved branches 
due to the swelling of the air, and perhaps also the water, in the stems 
under the influence of the higher temperature. A marked swelling of this 
kind should produce a straightening of the branch upon precisely the same 
principle as it straightens the bulb of a Richard thermograph. This could 
be tested by bringing typical curved branches from the shrubs on very cold 
days directly into a warm greenhouse, and comparing the distance between 
the base and tip before and after the branch had time to warm up. I tried 
this in a variety of ways, even bringing them abruptly from a temperature 
much below o° (C.) directly into a large case kept at a temperature above 
30°. To make the conditions as to water supply as uniform as possible, 
I plunged the branches at once into water in some cases (cutting them 
under water higher up the stem in some instances), and immediately sealed 
the cut ends with shellac in others. The results in all cases were the same. 
A slight straightening could often be observed within a few minutes under 
the higher temperature, but this was always lost within an hour or there- 
abouts, and was then replaced by a gradually increasing curvature. It 
became plain, therefore, that while a rise in temperature might cause 
