6 BULLETIN 1339, V. S. DEPARTMENT OF AGRICULTURE 



the average of maximum and minimum temperatures with various 

 biological activities. 



A rather rough record was kept of sunshine and cloudiness, but in 

 all calculations the records of the United States Weather Bureau for 

 Washington, D. C, were employed. There is a distance of 6 miles 

 between the Weather Bureau Station in Washington and the Bee 

 Culture Laboratory, at Somerset, Md., and it was found that only 

 negligible differences existed beween the recorded hours of sunshine for 

 the two places. Records of solar radiation were also available through 

 the kindness of Prof. H. H. Kimball, of the Weather Bureau, these 

 being taken at American University, in Washington, about 1 mile 

 from the Bee Culture Laboratory. 



Calculations relating to the effect of wind upon honey produc- 

 tion were not made. The anemometer maintained at the Bee Cul- 

 ture Laboratory was so located for another purpose that its records 

 were not applicable to this problem. 



The principal source of nectar available to the bees in this experi- 

 ment was that from tulip tree (Liriodendron tulipifera). It is an 

 excellent honey plant in this locality, yielding abundantly for about 

 two weeks in early May. Since it blooms so early in the season, it 

 is of the utmost importance that the beekeeper have his colonies 

 strong, otherwise its nectar is wasted with the exception of what little 

 is used for brood rearing. Occasionally black locust (Robinia pseu- 

 dacacia) furnishes considerable nectar, but this plant is not depend- 

 able in this region. The period of the secretion of black locust coin- 

 cides closely with that of tulip tree. In 1922 the bees worked on, 

 locust actively for several days, while scarcely a locust blossom was 

 seen in 1923. A rainy May, especially rain for the first few weeks of 

 May, spells crop failure for the beekeeper of this region. In this con- 

 nection, attention should be called to the statement of Kenoyer (19) 

 that a rainy May scarcely fails to precede a good honey season in 

 the State of Iowa. This emphasizes the statement made earlier in 

 this bulletin that with the present limited knowledge of the various 

 honey plants a general application should not be made of data secured 

 in a single locality with a certain species of plant. Kenoyer's work 

 was based on data obtained in the clover region. 



The coefficients of correlation between the various factors were 

 calculated from the following usual formula for biometric calculations 

 (7): 



—\ 1 



— xy I 



Probable errors were calculated from the formula — 



-r^ .67449 



PE r = — ^-(i_ r2)==Xi(1 _ r2) 



The values of %i were taken from Pearson (29), and the values of 

 1-r 2 from Miner's Tables (25) 2 . 



2 All mathematical calculations are based on accepted biometric methods. The writer here records 

 his thanks to Dr. Sewall Wright, of the Bureau of Animal Industry, for his unfailing assistance and 

 advice in this phase of the work. 



