28 BULLETIN" 1339, IT. S. DEPARTMENT OF AGRICULTURE 



Humidity is, of course, influenced, by temperature. In order, there- 

 fore, to arrive at a more correct value of the effect of temperature 

 upon nectar secretion, it is necessary to eliminate the effect of humid- 

 ity by reducing this variable to a constant. This is conveniently 

 done by the employment of the formula for partial correlation — 



-yJ(l—r 2 NB )(l—r 2 T H) 



Substituting in this formula the above coefficients of correlation, 

 and solving, we find that — 



hVnt = .ooOo 



It seems, therefore, according to Bonnier's data, that when humidity 

 is constant nectar secretion varies directly with temperature. 



The interpretation of similar data compiled by other investigators 

 would be highly desirable at this place, so that results could be com- 

 pared with those of Bonnier. The writer's calculation of Harrault's 

 (15) data gives the following results: 



r sa =. 2988 ±.1228 

 r TG =. 5258 ±.0639 

 r HT = . 2821 ±.1241 



In this case the net gain of a colony of bees was used, rather than 

 the volume of nectar, and G therefore equals net gain. By substi- 

 tuting these values in Pearson's formula and solving we find that — 



H r T6 =A82Q 



The coefficient of correlation between net gain and temperature after 

 correction is, therefore, somewhat lower than the direct correlation, 

 but it is still positive. 



In Table 2 the following coefficients of correlation are found: 



r Ba =~. 3806 ±.0935 

 r Te = .7529 ±.0473 

 r TH = -.1551 ±.1067 



In this case, by reducing the effect of humidity to a constant, we 

 have — 



H r T Q=. 75 94 



No attempt has been made in the present investigation to ascer- 

 tain the optimum temperature for the maximum change in weight of 

 a colony of bees. Kenoyer (20) found under experimental conditions 

 that the optimum temperature for nectar secretion for most of the 

 Leguminosae which he tested was 15° C. (59° F.). He also called 

 attention to the fact that the phenomenon of the accumulation of 

 stored sugar from starch at low temperature in the twigs of woody 

 plants is well known, and, further, presented data which show that 

 the same is applicable to floral tissues. 1 He likewise ascertained that, 



