WEATHER AND CHANGE IN WEIGHT OP BEE COLONY 



45 



from Table 6 and given, decoded, in Table 7. The following path 

 coe3icients are also used in the computations (see p. 42) for the 

 spring period : 



a = p T = .6898 ' d = p Q Hv = -.2075 

 h = p aTV = .0630 e = p s = .1004 

 c = p 0H = -.3217 



The formula for multiple regression, as given below, is used in 

 arriving at the value of the predicted gain G : 



n , - T-T.i Tv'-Tv^ H'-H^-j Hv'-Hv , e S'-S 

 G = G + aao — + ba a +c<tq \-da a + ea a - 



<Jt 



0~tv 



GH 



ffnv 



<Ts 



This equation, in turn, is simplified to — 



G> = G + ~k T (T' - T) +~k TV . (Tv'-Tv)+]ch .. (H'-H)+~k B v. 



(Hv'-Hv) +Jc* . (S'-S) 

 where — 



~k T = 134.5550 fc fl *= -18.3968 

 Tctv=* 11.3678 Tc s = 30.6658 

 Jc B = - 38.0266 



Table 6. — Mean and standard deviations of weather factors and changes of colony 

 weight for the spring period 





Net 

 gain 



Noc- 

 turnal 

 loss 



Diur- 

 nal 

 aver- 

 age 

 tem- 

 pera- 

 ture 



Diur- 

 nal 



vari- 

 ation 

 of tem- 

 pera- 

 ture 



Diur- 

 nal 



aver- 

 age 

 hu- 

 midity 



Diur- 

 nal 

 vari- 

 ation 

 of hu- 

 midity 



Hours 

 of sun- 

 shine 



Solar 

 Radi- 

 ation 



Noc- 

 turnal 

 aver- 

 age of 

 tem- 

 pera- 

 ture 



Noc- 

 turnal 

 vari- 

 ation 

 of tem- 

 pera- 

 ture 



Noc- 

 turnal 



aver- 

 age hu- 

 midity 



Noc- 

 turnal 

 vari- 

 ation 

 of hu- 

 midity 



Mean 



3. 9209 



4. 2630 



4.6314 

 2. 5385 



4. 9472; 3. 5262 

 2.4705 2.2448 



4. 6840 

 2. 5140 



5. 5787 

 3. 0406 



6. 1577 

 2. 6808 



5. 4998 

 2. 8169 



4. 7103 

 1. 5717 



5. 8419 

 2. 0333 



3. 8946 





2.2289 2.2908 



2. 2687 

















Table 7. — Decoded values of the means and standard deviations of variables used in 



determining G' 



Variable 



Mean 



a 



G 



3023. 1600 



1114.4500 



T 



75. 8(100 



5. 7110 



Tv 



25. 7300 



6. 1702 



H 



55. 0700 



9.4281 



Hv 



57. 0200 



12. 5700 



S 



10. 5580 



3. G487 



By substituting in cither of the foregoing equations the various 

 weather factors for each day, as given in Table 8, we obtain the effect, 

 measured in grams, of the individual weather factors upon the gain 

 of any particular day. Table 9 also j^ives these individual values for 

 all days shown in Table 8. 



Using this method, calculations have been made from the weather 

 data shown in Tabic 8 to determine the predicted gain for each day 



