NO. 17.] 



WIND. 



The Table and the Diagrams on PI. I (1 cm. = 1 m. p. s.) give 



Considering the possibility that the diurnal period of the velocity of the 

 wind might be different with different states of the sky, I have computed the 

 mean values of the velocity for the different hours in each month separately 

 for the days with clear sky (amount of cloud 0) and for the days with the 

 sky overcast (amount of cloud 10). The computation was made for the 3 

 years of the drift of the Fram, taking the average for each month, e. g. 

 October mean for 1893, 94 and 95. As September presented no clear days, 

 July only 4, and June and August only one, I have taken for these months 

 the days which give a mean amount of cloud of about 5. 



The following Table shows for each month the sum of the clear and 

 of the overcast days, and the respective means of the diurnal velocity of 

 the wind. 



