380 Transactions of the Boyal Society of South Africa. 
two rain-gauges, one at each extremity, and the mean rainfall at the two 
gauges is about the same, say 20 in., then in nine cases out of ten such the 
isohyet of 20 in. would be drawn with a semicircular sweep to follow 
the course of the hills, and in eight cases of the nine the isohyet might prove 
to be nearly correct. Such a plan has been followed to some extent here; 
but the facts as we find them show that in the lump inference is a poor 
substitute for knowledge, and tliat no such success in guessing can be 
expected in South Africa as is attained in countries where thunderstorms are 
fewer. Indeed, even in England, where the mean isoliyets are almost entirely 
of geographical inception, Mill has discovered that heavy rain, whether 
it fall in thunderstorms or during cyclonic disturbances, falls impartially on 
high or on low ground without any regard to the configuration of the country. 
" It seems," says Mill, " that the causes determining the fall of rain on 
these occasions are to be sought in the air alone, and that at a level so far 
above the surface that inequalities of ground, even amounting to several 
thousand feet, are without direct influence upon them.'"'' We may 
understand from this the great want of concordance between the isohyets 
and the contour of South Africa. If we would discover the reason why our 
isoliyets bend this way or that, we must first discover the reason why 
thunderstorms form and travel as they do. 
The same sort of considerations apply to the monthly maps. If we study 
them carefully, in order, month by month, we begin to see that the monthly 
mean rainfall over any area taken at random is not dependent so much 
upon the impinging of moisture-laden winds from the ocean upon lofty 
slopes, as upon the more general progressive movements of a rainfall belt to 
and from the equator. 
Consider, e. g., the isohyet of 8 in. In January this bends inland near 
latitude 17°, follows the course of the Zambesi to near longitude 25°, turns 
northwards to latitude 10°, crosses and recrosses this parallel, and finally 
runs down the meridian of 40° to the sea. Over all the land within this 
line the mean January rainfall exceeds 8 in. In February we find the area 
enclosed by the same isohyet somewhat further south, its southern boundary 
crossing the Zambesi and reaching the coast below Beira. In March it has 
retreated northwards again, its northern boundary being near latitude 8°, 
its southern boundary about 14°. In April it has nearly disappeared from 
the map and does not reappear until December. 
In January the 4-in. isohyet meanders from west to east between the 
7th and 30th meridians above and below the parallel of 21° ; it then turns 
towards the south-west as far as latitude 27°, from wdience it wanders 
south-east to the sea. To the left of this line the rainfall is greater than 
4 in. ; to the right it is less, and it gradually tails off to nearly nothing 
* H. E. Mill, " Map-studies of Rainfall," ' Quarterly Journ. R. Met. Soc./ April, 
1908. 
