22 BULLETIN 1339, U. S. DEPARTMENT OF AGRICULTURE 



ing following a heavy honey flow without much danger that the honey 

 will be too thin, since the greater part of the evaporation takes place 

 during the first night. 



The strength of the colony must likewise be an important factor 

 in the rate of evaporation. While it is true that a weak colony 

 collects less nectar than a strong one, it does not follow that the 

 efficiency of the two colonies is proportional to their strength. 



In studying the correlations in Table 2 it is seen that the external 

 factors have but little influence on nocturnal loss. If there existed 

 a strong internal individual influence in either one of these colonies, 

 one would not expect such a high correlation between the nocturnal 

 losses of colonies 1 and 2 (.8938, see p. 13). If external factors are 

 important so far as nocturnal loss is concerned they must be other 

 than temperature, temperature variation, relative humidity, or 

 variation of relative humidity. Again the need for strong colonies 

 is apparent. Weather conditions at night undoubtedly influence 

 weak colonies to a greater extent than they do strong ones. Phillips 

 and Demuth (80) in their wintering experiments have shown the 

 detrimental effect of low temperatures upon weakened colonies, 

 while strong colonies reacted normally to the same temperatures, 

 if we take the survival of strong colonies and the death of weak ones 

 as a criterion for differences in normal behavior. 



THE FALL PERIOD 



The data upon which the calculations for this period are based 

 were secured from September 4 to October 5, 1922, during the 

 period of the fall honey flow. No criterion for minimum gain was 

 used in selecting the day to be correlated, and all days irrespective 

 of gain or loss are included in the calculations. For this reason the 

 figures can not be compared directly with those of the spring period, 

 when only days having a minimum of 980 grams were selected. In 

 studying the figures representing days of the fall period, it will be 

 seen that the graphic presentations of hive weights are more signifi- 

 cant than graphs showing the difference in weight from hour to 

 hour. As in the analysis of the hourly weight changes of the spring 

 period, weather factors, as they affect changes of weight in the fall, 

 will be postponed for later discussion. 



MORNING LOSS 



The relation of morning loss to net gain appears to be the reverse 

 of what it is during the spring period. The correlation between 

 morning loss and net gain is .5769 ± .0825, so that not only is the 

 correlation between these two factors numerically greater than it is 

 in the spring, but it is positive rather than negative. The morning 

 losses generally were heavy, being from 390 to 1,690 grams. The 

 morning loss is primarily brought about by the rapid exit of bees, as 

 undoubtedly the returning bees with their loads of nectar more than 

 make up for the loss due to the consumption of stores for maintenance 

 of the colony, which constitutes a steady loss at all times. The 

 duration of these losses varied from four to nine hours, and often 

 equally as long a time would be required for the colony to regain 

 its original weight of the early morning, this in turn leaving only 



