Barometric Variability at Kimberley and ^Isetoheix. 109 
lished. For the purposes of this inquiry the Argentine constants would 
he of great value ; but the results of observations for that region in 
my possession end with Vol. XIV. (1901), carrying the record for Staten 
Island no farther than 1896, and for Cordoba only to 1898 ; so that only 
these two stations — and for a limited period at that — could be used. The 
Anales are, moreover, an expensive luxury. 
14. Hong Kong. The variability for the 14 years 1884, 1886-7, 
1902-12 has been computed for the sake of including one station in the 
Northern Hemisphere. 
Table 3 gives the monthly mean constants of barometric variability, 
derived from differences such as those of Table 1, for 13 stations in the 
Southern Hemisphere as well as for Hong Kong. The average of all the 
Southern Hemisphere stations appears on the bottom line. 
According to this Table the minimum variabihty for the Southern 
Hemisphere generally falls in February, and the maximum in August. 
There is besides a secondary minimum in July. 
The South African stations show the February minimum, as do also 
Wellington and Cordoba. In Australia, however, the minimum is retarded 
into March and April, stations to the north and east feeling it later than 
do those to the south and west. It is worth notice in this connection, too, 
that the two easterly South African stations — i.e. East London and Durban 
— show a secondary minimum in April. 
Cape Town has its maximum variability in August, whereas this is 
retarded into September at East London, and even into October at Durban. 
Kimberley conforms in a way to the rules of both East London and 
Durban in that the maximum is shown in October as well as August. 
At Port Darwin and Perth the maximum variability falls in June. At 
Adelaide it comes in September, at Alice Springs between September and 
October, and at Sydney in October. Alice Springs and Sydney have 
besides a fairly strongly marked June maximum, and there is also a 
tendency to this maximum at Adelaide. Thus on the whole the rule that 
the more easterly of the south-coast stations have the later maximum 
variability holds good in Australia as it does in South Africa. 
Wellington has the greater maximum variability in June and the lesser 
in August. Cordoba has one maximum, in July, and Staten Island prob- 
ably one in August. Further, at quite one-half of the stations the 
December variability is greater than that of November. 
Either the July minimum, or a tendency to it, is shown at most of the 
Southern Hemisphere stations. 
One deduction to be drawn from Table 3, and particularly from the 
summary on the bottom line, is that the "equinoctial gales," so far as 
barometric changes can indicate them, have no existence in fact. 
Table 4 shows the maxima of barometric variability in each month, 
