Possible Lunar Influence upon the Velocity of the Wind at Kimherley. 147 
winds of Kimberley during the day are moving normally down the gradients 
of pressure set up by the lunar tide in the air, thus augmenting their 
speeds, whereas during the night they are moving up the gradients and 
thus diminishing their speeds. A similar sort of explanation applies to 
the ''horizon" curve, and to a good portion of the "midnight" curve; but 
it gives no adequate reason for the minimum near moonrise common to all 
the curves, and which is probably of a more general character. Thus 
the curves shown in the diagram are largely due to the superimposition 
of the lunar air tide upon the diurnal variation of wind direction. 
The acceptance of this explanation would demand that the velocity 
deviation curves of no two places even on the same circle of latitude would 
be quite alike, as the lunar air tide (to which they are in great part subject) 
is. Even so the deviations of velocity cannot be regarded as mere surface 
phenomena, but rather as the outcome of conditions common to the whole 
depth of the atmosphere. 
In a later paper I hope to discuss the velocity deviations for the moon in 
apogee. My wife, as usual, has checked off the averages : without her help 
the investigation could not have been undertaken. 
In the diagram A is the mean curve for all the observations used in the 
previous paper ; P, M, N, H are the mean, " midnight," " noon," and 
*' horizon" perigee curves of columns 2, 4, 3 and 5 respectively. 
