62 Transactions of the Royal Society of South Africa. 
Considering individual months, it happens that of the 88 months having a 
fall less than the mean, 38 occur with a small gradient and 50 with a large 
one ; whereas of the 56 months having a rainfall greater than the mean, 34 
occur with a small gradient and 22 with a large one. That is to say, 
deficient rainfall tends towards occurrence with months of large, while 
abundant rainfall tends towards occurrence with months of small gradient. 
(2) Kimberley. — At Kimberley, on the whole, we have the opposite. The 
rainfall when the gradient is large exceeds that when the gradient is small 
by about 15 per cent , the effect being most marked in the winter. Eight 
months agree, three do not, and one (October) is about neutral. By a 
coincidence the number of months showing excess or defect is the same as 
at Cape Town though in the opposite direction like the quantities, i.e. of the 
88 mouths having a fall less than the mean, 38 occur with a large gradient 
and 50 with a small one : whereas of the 56 months having a fall greater 
than the mean, 34 occur with a large gradient and 22 with a small one. 
These facts may be summed up in — 
Mule 1. — An excess of rain inclines to occur either at Cape Town or 
at Kimberley according as the synchronal gradient is small or large 
respectively. 
The question now arises whether this tendency is something more than 
synchronal : e. g. whether it persists at all into the following months ? 
Obviously, by the above Rule, it can only persist at the best to a partial 
extent ; for if the months of small and large gradient were to alternate 
continually, any persistence would necessarily be masked by the current 
conditions. But the length of a spell of gradient, large or small, may be 
anything from one to six months, the average duration being about three. 
It seemed worth while, therefore, to compare the gradient of each month 
with the rainfall of the following month. This gives Table III. 
The unexpected result shown by Table III is a considerable reversal of 
the conditions shown by Table II : more rain going to Cape Town and less 
to Kimberley when the gradient of the previous month was large. With 
this qualification, that the effect is a winter one at Cape Town, and a 
summer one at Kimberley. In the summer half at Cape Town and the winter 
half at Kimberley the normal condition persists into the following month. 
Hence we have — 
Rule 2. — An excess of rain inclines to occur either at Cape Town in the 
winter or at Kimberley in the summer according as the gradient of the 
previous month is large or small respectively. 
Rule 3. — At Cape Town in the summer and at Kimberley in the winter 
the conditions under Rule 1 incline to continue unaltered into the following 
month. 
These results, based on a comparatively short period of observation, can, 
of course, only be regarded as provisional. As they stand they can only 
