6 BULLETIN 96, U. S. DEPARTMENT OF AGRICULTURE. 



and the top protection it was possible at any time of day or night 

 to look between the combs at the cluster. These protective cover- 

 ings were applied so as to be removed with the minimum jar. At 

 night, or even in the daytime, by means of a reflector, lantern light 

 could be thrown up between the frames. In this way the writer 

 was able to watch from day to day the shifting of the cluster and 

 the reaction of the bees to their environment and to compare this 

 with the readings of the thermometers in the hive on the scales. It 

 was necessary to maintain this check only during the whiter period. 



METHODS OF OBSERVATION AND RECORDING. 



Since none of the instruments recorded automatically, it was nec- 

 essary to make frequent readings of both the weights and tempera- 

 tures. The experiment proper lasted from October 22, 1907, to 

 September 26, 1908. The first colony, used to prepare the combs, 

 was also under close observation, so that the whole period of experi- 

 mentation was almost a year. Readings were taken at least every 

 hour throughout the working day. Whenever the hive was manipu- 

 lated, or when peculiar meteorological conditions prevailed, readings 

 were taken half hourly, or even quarter hourly. On the average of 

 about once in three weeks, by means of assistance, it was possible to 

 take consecutive hourly readings for a period of two or three days. 

 In this way practically the whole activity of the colony for a period 

 of a year was recorded. During the summer months the readings 

 usually covered a period of 14 hours daily. 



The temperatures were read to fifths of a degree. Weighings 

 were made to 10 grams. Every alteration or manipulation of the 

 colony was recorded. Hourly changes in the weather and activity 

 of the bees were also noted. 



The readings were recorded on 12.5 by 20 cm. cards, the size 

 standard to the office note. file. Later from these tables the curves 

 of the temperature and weights were plotted on millimeter cross- 

 section paper, one sheet to a month. The method of plotting is 

 obvious from examination of the several curves herein presented. 



THE CONSUMPTION OF STORES IN WINTER. 



At the outset of the investigations it was hoped by means of deli- 

 cate scales, which have been described, that sufficiently accurate 

 weighings could be made to show whether there is any correlation 

 between the loss in weight and the temperatures of the cluster in 

 winter. For instance, it was desirable to know whether there is any 

 relation or rhythm in the consumption of stores to changes in tem- 

 perature due to metabolism. It has not been possible to detect any 

 such relations. Nevertheless several significant facts concerning the 

 consumption of winter stores have been discovered. 



