WEATHER AND CHANGE IN WEIGHT OP BEE COLONY 21 



at the time that flights cease. The daily gains generally end abruptly, 

 a fact suggesting that the bees do not desert the field at this time for 

 lack of nectar. 



NOCTURNAL LOSS 



The nocturnal loss is calculated from the time that gain ceases 

 until the bees leave for the field the following morning (fig. 2, D-E). 

 Nocturnal loss is largely the result of the evaporation of water which 

 is given off in the process of ripening honey. In this regard, noctur- 

 nal loss bears an intimate relation to the net gain. 



The coefficient of correlation between noctural loss immediately 

 following the day's gain and the net gain is .4112 ±.0909. The 

 correlation of the diurnal gain with the loss of the second night 

 following is much less, being .1890±.1099. In the second instance 

 the coefficient of correlation is not twice its probable error, and so 

 can not be considered as especially significant. This would indicate 

 that the greater part of the necessary evaporation is accomplished 

 during the first night and that little is left over until the following 

 night. Evaporation naturally begins almost as soon as nectar is 

 gathered, and therefore occurs during the day of gathering and pro- 

 bably during the following day, but in neither case is it possible to 

 determine the amount of evaporation by day from a record of colony 

 weights. 



In 1923, colony 1, from May 18 to May 30, lost during the nights 

 25.29 per cent of the total amount gained. Colony 2 in the same 

 period lost 24.69 per cent of the total gain. For 11 days in May, 

 1922, colony AB lost during the nights 17.85 per cent of the total gain. 

 This lower percentage of loss in 1922 may have been occasioned by 

 the character of the nectar collected, since considerable locust honey 

 was available. It is also well known that the water content of nectar 

 varies from year to year with different varieties of plants and with 

 climatic factors. A calculation of Dufour's (11) data shows a loss 

 of 26.16 per cent for 14 consecutive nights in May and June, when 

 the minimum net gain was 970 grams, and the coefficient of correla- 

 tion between net gain and nocturnal loss for this period is found to 

 be .6663 ±.1002. Maujean (23) found a nocturnal loss of 22.53 per 

 cent of the net gain during the honey flow of 1905, and a loss of 27.40 

 per cent during that of 1904. A calculation of Maujean's data shows 

 a coefficient of correlation of .7868 ±.0411 between the net gain and 

 the following night's loss, a correlation of .5319 ±.0795 between the 

 daily net gain and the loss of the second night following the day con- 

 cerned, and of .2813±.1035 between net gain and the loss of the 

 third night following. These results would indicate that in this 

 particular instance evaporation was practically complete by the end 

 of the third night; a much slower evaporation than seems to have 

 occured in the present investigation. Hommell (18) states that ac- 

 cording to Sylviac seven days is the minimum observed time required 

 to change nectar completely to the consistency of honey; he thinks, 

 however, that seven days is entirely too long for this minimum period, 

 and points out that de Layens long ago discovered that in warm and 

 dry weather nectar may be of such consistency as to permit almost 

 immediate capping by the b^cs. Hommell points out that Huillon 

 likewise gives a much shorter period for the completion of evapora- 

 tion, and states that beekeepers can remove honey during the morn- 



