180 
Transactions of the Royal Society of South Africa. 
scale than the corresponding perigee ones do, mean values in miles an hour 
being — 
Perigee. Apogee. Di^erence 
"Noon" .... 10-132 . 9 996 . -136 
Midnight" . . . 10-036 . 9-922 . -114 
"Horizon" . . . 9-882 . 9-844 . -038 
Mean .... 9-994 . ,9-899 . -095 
Further, while the velocity tends to rise when the moon is above the 
horizon at perigee, it tends to fall at apogee. And while the perigee curves 
end the lunar day at a higher level than they started at, the apogee curves 
run down. The relatively large range of variation shown by the apogee 
curve is probably due to this circumstance. 
In their general tendency the results so far obtained are pretty much 
w^hat might have been expected ; but the actual ranges of velocity are greater 
than I looked for, reasoning from the thermodynamical analogy of the semi- 
diurnal oscillation of barometric pressure. Their acceptance would demand 
for the purely dynamical problem of the air tide a lively circulation 
throughout the whole depth of atmosphere. On the negative side of the 
account it may be claimed that at the least if, at a place so favourably 
situated as Kimberley is for a research of this kind, the moon at its best 
cannot create a range of wind velocity of more than 0*7 mile an hour in the 
course of a whole day, then how can it possibly promote the storm and 
tempest attributed to it by the weather prophets ? 
Again I have to thank my wife for her assistance. 
