44 BULLETIN 1339, U. S. DEPARTMENT OF AGRICULTURE 



In this case — 



a= -.27206 c= -.07350 

 1= .29408 d= .29642 



The solution of the equation — 



PltTlt +PltvTltv + fisTm + PluvHuv =l — o 2 



therefore gives — 



o 2 = . 56148 



It is interesting to note that the value of the unKnown nocturnal 

 factors during the fall is less than it is during the spring. With the 

 same variables in use for calculating the value of o 2 during these two 

 periods, the difference must be accounted for by variation in the 

 behavior of the bees themselves. 



The writer has calculated the value of o 2 from the coefficients of 

 correlation which he determined from data by Harrault (15), given 

 under the discussion of temperatures. In this case only two vari- 

 ables are available, namely, the average daily temperature and the 

 relative humidity. The solution of the two simultaneous equations — 



r G T = a + hr T H 

 Tan = ar T H + h 



gives a =.4779 and Z>=.1635. By substituting these values in the 

 formula — 



PotTgt + pcnToH = 1 — O 2 



we have — 



o 2 =.6998 



This figure shows the value of the unknown factors to be greater 

 than that found in this work for the spring period. 



A similar analysis of Bonnier's (1) data gives o 2 = .3093. This figure 

 is lower than that given for the spring period. One must take into 

 consideration that only temperature and relative humidity were con- 

 sidered by Bonnier and that activity of the bees does not enter into 

 consideration in his work. 



THEORETICALLY CHANGING WEATHER FACTORS AND PREDICTING 



RESULTING GAINS 



One of the primary objects in presenting the following data is to 

 demonstrate the value of accurately kept records of colony weight, 

 together with weather records. By knowing how changes of colony 

 weight vary with changes in the weather factors over a series of 

 years, it would not be an impossibility to predict whether or not, in 

 the long run, a certain locality would justify commercial beekeeping. 

 Such data could also be used to plan migratory beekeeping and to 

 learn perhaps whether a locality not suitable for honey production 

 would be suitable for the production of bees or for queen rearing. 



In order to give concrete examples in predicting gain under any 

 variation of weather conditions, the necessary data may be taken 



