398 BOTANICAL GAZETTE [JUNE 
Such substances can only retard, not prevent, the escape of heat. 
As a final argument we may return to the fact that observation 
shows that buds are always filled with ice during cold periods, which 
of course could not occur if they were kept warm. 
It is a more difficult matter to demonstrate whether the non- 
conductivity of the bud-scales is of importance to the bud in any 
other way. Recently Griiss'® has quite firmly upheld the theory 
that one of their chief functions is to modify the temperatures reach- 
ing the interior of the bud. We can conceive of several ways in 
which such protective service might occur. First, poor conduc- 
tivity might prevent injury from too rapid thawing. Second, bud- 
scales might prevent extreme fall of temperature by preventing 
excessive radiation. Third, they might prevent too frequent rapid 
thawing and freezing due to fluctuating sunlight, and thus prevent 
excessive water evaporation. 
Before answering any of these questions let us try to understand 
a little more fully the actual relation of bud-scales to heat. This 
problem resolves itself into two parts, namely, a consideration of 
the conductivity simply, and a consideration of the relation to normal 
atmospheric heat changes in the open. 
On the question of conductivity the following experiments seem 
to throw some light: Two thermometers, previously tested as to 
their readings, were selected, and the bulb of one was covered with 
the imbricated scales of a fresh horsechestnut bud, as in the previous 
experiment to determine the effect of color, thus forming an arti- 
ficial bud with the thermometer bulb in place of the young shoot. 
The other bulb was left naked. The experiments were all conducted 
within the building where the conditions were more constant and 
presented fewer uncontrollable factors than outside. The room 
rapid growth. During the dormant winter period it must be very slight in all trees. 
An ordinary thermometer probably could not measure it. It may also be suggested 
that since the large size and mass of the trunk would retard heat changes, being 
warmer than the air meee the temperature is falling, and cooler when the latter is 
this, by conduction along the gp hae modify the temperatures in the 
shoots and buds. UIRES has shown (M Bot. Stud. 1:453) that the average 
empertr in a box elder tree was in ace 1.3°C. higher than the air, in Fe 
ry the same as the air, and in March 1° lower. The differences between internal 
‘esd i deena temperatures during the day was in all cases only a few degrees. he 
idea that the branches can conduct such slight modifications so long a distance 
without loss is so evidently unreasonable as to require no more discussion here. 
16 Griiss, J., Beitrage zur Biologie der Knospen. Jahrb. Wiss. Bot. 23:651. 1892. 
