42 TROWBRIDGE: THE THERMOMETRIC MOVEMENTS 4 
for Branches A and B, except that the days of observation were 
somewhat different. In Fic. 8, the relation of the variation in 
height of Branch A (same as Fics. 4 and 5) to the variation of 
temperature is shown. It is seen at once that there is a marked 
point of inflection of the curve at the freezing point of water. 
The fact that the observation points do not fall on the curve is 
effect of the temperature is below 32° F., there is evidently a 
minor thermometric effect just above 32°, which, however, is $0 
small as perhaps to be due to a warping of the branch due to in- 
ternal strains arising from the bending of the branch, caused by 
below freezing temperatures a short time before the observations 
at warmer temperatures were made. Thus the behavior of the 
branch depends somewhat on its immediately previous history. 
HOURLY RECORDS AND EXPLANATION OF CURVES IN FIGs. 9, 
IO AND II 
In ious curves, an hourly record was kept of several of the 
branches of Tree No. 1 to ascertain how soon the branch responded 
A + | 
Ic xe : Lt 
Bb bee 2 Ce BAG aS 
E =e ~y re HEIGHT OF BRA! A 
6 Op x ie aa . a B 
~ to! Hourly Record aI = 
. “a 
Tree No. 1 * ye WY | LAT Branch C 
Linden ss prs vA 
& y 
50° IN Z 
Be 7 
/ 
nt 
: \ : 
re A L ; 
PF | EIS ne err mcr abt ora on 4 on A : 
: — \ é se sone : 
OE: Baste AP a © es T. ca 
o AM Pr 
% 56 78 FT Oh 23% § 6 i 8.9 6 RS ee 6 6 
Fic. 11. Hourly record of Branches A, B,and C, Tree No. 1 (European linden), : 
the lag in the movements of the branches with fairly rapid changes of temperature 
and explain the variations among the observations shown in the curve in FIG. I. 
