80 
BOTANICAL GAZEFTE [JUNE 
fall was most rapid. Of much more importance to our 
problem is the retarding effect in point of time, shown 
in the fifth column. This increased very rapidly towards 
the close of the readings, but was for our purpose 
practically the same in both cases. It was greater in 
proportion to the slowness of heat penctration, and was 
also somewhat greater at first in Table I than in Table 
II. The greatest retardation capable of measurement 
with the thermometers used was about twenty minutes, 
while for most of the experiment it was only from one 
to nine minutes. It was found that decreasing the 
thickness of the scaly covering decreased this time 
difference very markedly; while the presence of air 
Oe i 
between the scales tended to make it greater. 
The mass of the thermometer bulb, or of a 
shoot in a normal bud, and the extent of the 
ai radiating surface, are important factors in deter- 
a mining the length of time required for such a 
structure to cool. While the mass of the mercury 
N in this case is much greater than that of 
ae 
\ a _- 
NJ 
a2 
c. 3 
S688 © Sa 2s 8 2 SS s ce = 
horsechestnut bulb; _........... naked bulb. Abscissas represent 
Fic. 8. 
5° F.; ordinates, 100 seconds. See Table III. 
the shoot, its specific heat being only one-thirtieth that of water 
would render the two not very dissimilar, so far as the present 
problem is concerned. In apparent volume they do not differ 
greatly, so that the radiation surfaces of the two would be nearly 
the same. 
I believe we are justified in saying that a normal horsechestnut 
bud would not behave in any essential way differently from the 
artificial one here used; and that the time for it to cool off would 
— 
