WEATHER AND CHANGE IN WEIGHT OP BEE COLONY 5 



METHOD OF OBTAINING DATA 



From February to November, 1922, continuous records were made 

 of a colony of bees placed on platform scales sensitive to 10 grams. 

 Owing to the nature of the principal problem under investigation, 

 hive temperatures, the hive was left unprotected and certain of the 

 weight records were rendered useless, water remaining on the hive 

 and bottom-board after rains causing too great errors in the weight 

 records. During most of the month of June, 1922, alterations were 

 necessary in the apparatus for recording temperatures in this colony, 

 and weight records are" therefore not available for this month. 



A standard 10-frame Langstroth hive was used to house this colony. 

 Previous to the honey flow two brood chambers were in use, and a 

 few days before the beginning of the honey flow the colony was manip- 

 ulated to prevent swarming and additional supers were added. The 

 colony was left in this condition until, in June, alterations were made 

 in the temperature-recording apparatus. 



During May, 1923, two colonies placed side by side, both on equally 

 sensitive scales, were observed for additional data covering another 

 honey flow. Both colonies were in standard 10-frame hives, each 

 having two hive bodies for brood and three, above a queen excluder, 

 as supers. In this case the colonies were given ample room for brood 

 rearing and storage of nectar, and made no attempt to swarm. Pre- 

 cautions were taken to shield these hives from rain, by placing a 

 sloping board over the entrance of each colony in such a manner 

 /that rain would not fall on the bottom-board, while there would be 

 free access to the bees. A false roof was suspended over the outer 

 cover of the hive, which caught the rain and shed it beyond the hive 

 walls. A light wooden framework over the two colonies was covered 

 with canvas. In case of a light rain when the bees were still flying 

 the canvas cover was not used, but during periods of hard dashing 

 rains the canvas was dropped so low as to surround the hives com- 

 pletely and prevent any water from collecting on them. 



Weighings were made hourly throughout the 24 hours of each day, 

 including Sundays and holidays, three persons being assigned to the 

 work on 8-hour shifts. One of the unfortunate circumstances con- 

 nected with working on such a problem at Somerset, Md., is that the 

 main honey flow is exceedingly short (seldom over 10 days to 2 

 weeks), thus increasing the probable error in all calculations by re- 

 ducing the number of possible observations. The correlation coeffi- 

 cients for the spring honey flow are based on the one colony in 1922, 

 and on the two colonies in 1923. For conclusions relative to the fall 

 honey flow only the figures for the one colony of 1922 are available. 



In order to calculate the relationship between changes of weight 

 and weather conditions during the honey flow, only those days show- 

 ing a net gain of 980 grams or over were used for the spring honey 

 flow. The calculations for the fall period include all days from Sep- 

 tember 4 to October 5, irrespective of gains or losses; the two sets of 

 data are therefore not strictly comparable. 



Hourly thermograph and nygrograph records were made for use in 

 calculating the mean temperatures and relative humidity; the mean 

 of maximum and minimum temperatures was not used. Hartzell 

 (16) has pointed out the chances of error in endeavoring to correlate 



