OF TREE BRANCHES AT FREEZING TEMPERATURES 49 
In Fic. 18 the frost crack is shown closing rapidly when the 
“Bark” (one inch deep) and the “‘ Medium” (three inches deep) rise 
abruptly, while the‘ Deep’ thermometer (fiveirnichesdeep) remained 
nearly constant. In Fic. 19 the Deep thermometer is still below 
19° F. while the frost crack is nearly closed. Moreover, in both 
sets of observations, the frost crack appears to close with the rise 
of temperature of the surface layers of the trunk. 
THE ORIGIN OF THE THERMOMETRIC MOVEMENT 
While there may not yet be sufficient data on which to formu- 
late a theory for the actual mechanism of the thermometric 
bending of the branches, there are some salient facts which 
indicate the origin of the phenomenon. It has been shown, at 
_ least so far as the linden is concerned, that the movement begins 
_when the air temperature is at the freezing point of water. The 
internal temperature of the small branches could not be far from 
that degree of cold. 
In the curves published by Caspary* on the European linden 
showing the branch movements with temperature, there is corre- 
spondence between the temperature and branch positions, which 
indicates that the depressions of the branches began at the freezing 
point of water, but there is no mention of this in his text. 
Owing to the fact that a number of trees were studied, which 
showed various behavior with change of temperature, the matter 
might well have been overlooked. This opinion seems to be 
substantiated by the following quotation from Caspary’s paper: 
“Der Grad der Kalte, bei welcher ein Sinken eintritt, kann 
wegen Mangel stiindlichen Beobachtungen und Beobachtungen der 
Kalte des Holzes selbst im Innern nicht sicher angegeben werden.”’ 
Besides the linden, measurements of trees of various other 
species by Caspary show that the movement begins at approxi- 
mately 32° F. By making hourly observation, not made hitherto, 
the writer has brought out some important facts; namely, that 
below freezing the branches respond to the changes in temperature 
almost immediately. Thus, as shown by the curves in Fic. 11, 
when the temperature falls to 20° F., and then rises, the branches 
at once slow up in their downward movement, and shortly after- 
* Loc. cit. 
