Oocroser 4, 1918] 
for special types of investigation. Together 
with each measurement a reading was taken 
of the air temperature as given by a mercury 
thermometer attached to the tree, as well as 
the reading of one or more thermometers in- 
serted into the tree to various depths. 
_ Observations were made on a European 
linden tree (Tilia europea), and a plane tree 
(Platanus orientalis), both on the campus of 
Columbia University. The observations ex- 
tended from February 2 to May 19, 1917, and 
from December 22, 1917, to April 25, 1918. 
During the first winter, observations were made 
on both of these trees, but attention was con- 
fined to the linden tree alone during the sec- 
ond winter, as the same effects were observable 
here to a far more marked degree. During 
both winters, longitudinal and transverse meas- 
urements were made, a separate point and 
socket being used for each, and a longer rod 
being used for the longitudinal observations, 
as the longitudinal changes were, as a rule, 
much smaller in amount. An extended series 
of measurements was also made on the changes 
in the circumference of the tree and on frost 
eracks, during the second winter. Three in- 
terior thermometers were used in the first win- 
ter’s observations, four in the second, one ex- 
tending to a somewhat greater depth than the 
deepest of the previous winter. No observa- 
tions were made during the summer, as it was 
found that at ordinary ‘and high temperatures, 
the changes in dimensions were extremely 
slight. Observations were made from one to 
four times a day, and readings of the various 
thermometers were sometimes taken more fre- 
quently. During the winter of 1917-18, the 
writer made 109 sets of measurements, and 
about the same number during the preceding 
winter. 
The second winter’s observations fully con- 
firmed the earlier series, and added some new 
results. In regard to the transverse measure- 
ments, it was found that above 32° Fahren- 
heit there is a slight expansion with rise in 
temperature, while below that temperature the 
changes are far more marked. As the tem- 
perature falls below 32° Fahrenheit there is a 
very marked transverse contraction. The dif- 
SCIENCE 
349 
ference in the changes above and below 
freezing may best be illustrated by stating, 
in the case of the linden, that above the freez- 
ing temperature, the coefficient of expansion is 
nearly the same as that of dead wood, 2%. e., 
of the order of 5 & 10“, while below freezing 
it is some fifty times as great. 
The transverse change in dimensions of the 
tree, below freezing, usually lags behind the 
change in temperature of the bark by several 
hours at least, often as much as twenty-four 
hours. When there is a sudden change in the 
temperature of the bark, the contraction is 
rapid, but not synchronous. With a rise in 
temperature, the lag, as a rule, is relatively 
greater. It is probable that the temperature 
at a depth of four or five inches has little or no 
influence on the changes in transverse dimen- 
sions. 
In the case of longitudinal measurements 
the fact was revealed that below the freezing 
temperature there is a minute but extremely 
definite increase in length with fall of tem- 
perature, and that above freezing, there is an 
equally minute increase with rise of tempera- 
ture. At extremely low temperatures, near 
zero, Fahrenheit, however, there is a small 
contraction with fall of temperature, but when 
the temperature rises again, the expansion is 
extremely rapid, and by the time the tempera- 
ture is again the same as before the drop, the 
tree is very much longer than previously. 
In this series of measurements at very low 
temperatures, there is distinct evidence of 
two changes—thermal and physiological, ap- 
parently acting in opposite directions. At 
slightly higher temperatures the thermal 
change is not so much in evidence, and so, as 
a tule, only the physiological expansion with 
drop of temperature is observed. There is 
evidence of a lag of longitudinal expansion 
and contraction behind the temperature of the 
bark of the tree, but excepting at the lowest 
temperatures, the phenomenon is not clear 
cut, as in the case of the transverse measure- 
ments, and the details have not as yet been 
worked out. 
A very extended series of measurements was 
made on the circumference of the linden tree, 
