6 
ter of the cluster gradually rose from an average of 16° C. on 
December 13 to an average of 30° C. on the 22cl, though the air outside 
the hive kept in the range of temperature from 6° C. to 9° C. This 
is in agreement with results obtained by Phillips and Demuth (loc. 
cit.) with bees wintered in a cellar which were interpreted as indi- 
cating that such an upward drift of temperature of the colony dur- 
ing confinement is the result of irritation because of an accumulation 
of feces. In the case of the colonies recorded in an earlier publica- 
tion, 6 one colony showed a slower rise than was f ound in this colony, 
while another, wintered on honeydew stores, showed a more rapid 
rise. Since it has been shown that disturbance of any sort causes a 
rise in cluster temperatures, it is not entirely clear to which disturb- 
ance the rise of this colony should be attributed. Of course, as this 
colony was located in a respiration chamber in a busy laboratory, 
it was exposed to greater disturbance than would have been the case 
in some other experiments or in the average bee cellar, although all 
practicable precautions were taken to avoid jar and the apparatus 
was cushioned. It is not improbable that the sudden and temporary 
increases in temperature may have been due to physical disturbance 
and that the cause of the continued rise was physiological dis- 
turbance. 
It will be noted that beginning at 6.30 p. m. on December 22 the 
temperatures of the cluster began to drop. At this time the carbon- 
dioxid content of the air in the chamber was high and the oxygen 
deficient, as will be explained later. Under these conditions the bees 
were more quiet (generated less heat) than when under conditions 
which would usually be considered more favorable. The temperature 
of the center of the cluster dropped until it reached 23° C. The 
reason for the decrease in activity at this time has not been dis- 
covered. It was thought that the bees were dying because of un- 
favorable atmospheric conditions, but at 5 a. m. on the 23d the tem- 
perature again began to rise and continued until it again reached 
34° C. Whether this increase in activity was a reaction in response 
to physical disturbance or to change in atmospheric conditions made 
at this time (see p. 13) is not clear. 
METHOD OF MEASURING THE WORK DONE BY THE CLUSTER. 
At noon, December 12, measurement of the metabolic activity of the 
bees was begun. The respiration calorimeter used for this experi- 
ment has been described in a publication of the department, 7 but to 
aid in explaining the conditions of the experiment the principles of 
6 U. S. Dept. Agr. Bui. 93. The Temperature of the Honeybee Cluster in Winter. 
7 Jour. Agr. Research [U. S.], 6 (1916), No. 18, pp. 703-720. 
