E.—_-GEOGRAPHY 105 
lesser (first) and greater (second) rainy season are recognised. Moreover, 
in Barotseland a fourth has to be added, named Munda—the floods—for 
there the regimen of the Zambezi and tributaries is of prime importance. 
There is thus a dry period of at least six months during which the 
temperature is first dropping to its minimum in July and then rapidly 
rising to its maximum in October or November. The rains then spread 
southward and eastward, the belt of maximum precipitation being in the 
north-west in November, and in the east around Lake Bangweolo in 
December. In January rain is more evenly distributed; in February 
the maximum is again in the east, from which it gradually withdraws 
northward again. ‘The total rainfall is over 50 in. near the Congo frontier, 
and decreases eastward to 35 in. on the Nyasaland border and southward 
to under 30 in. in the Zambezi valley. 
The annual rhythm of vegetation, of animal life, and the seasonal activi- 
ties of the population are matters upon which we have received much 
information, especially on the latter. These phenomena can now be 
closely related to the temperature and rainfall factors and the flooding of 
the rivers and variation of swamps and lakes. 
In the absence of real knowledge of Rhodesian soils we may legitimately 
have recourse to the Belgian pedological work in Katanga, where, 
however, the rainfall is heavier. Owing to the very great extent of 
surfaces of peneplane type, it is most probable that the Rhodesian soils 
as a whole are residual and old, deficient in soluble salts and more or less 
lateritic. Moreover, deforestation over large tracts has proceeded for 
long; and the removal of this natural protective cover results in the 
lowering of the water table during the dry season, and the loss of fertility, 
especially by the removal of humus, a contributory factor in this being 
the widespread annual grass fires. It may be regarded therefore as most 
likely that prevailing plateau soils are poor. ‘Their vegetation is savanna, 
or what Shantz has classified as dry woodland, in which the trees are mostly 
deciduous and where their stature and their density varies with available 
water. They are, of course, associated with grass which is renewed each 
rainy season, and which, like the trees, varies with the rainfall. 
Throughout the peneplanes are numerous shallow hollows known as 
dambos, filled by wash from the slopes, sandy and lateritic round their 
margins, clayey and marshy in their centres. Their soil is infertile and 
grass predominates in their vegetation. 
By far the most attractive soils are those of the alluvial areas. In the 
maps of the Katanga these are distinguished according to age—young, 
adult and old; and the District data from Northern Rhodesia would 
seem amply to justify this classification as one that is important in 
the human geography. But of course it is impossible to do more than 
guess the distribution of such types in any locality. The first class are 
annually inundated and renewed ; the second, which may occasionally 
be flooded, are typically dotted over with termite hills. The plant cover 
of both these types is herbaceous, and their edges would seem to form the 
sites of the great majority of native villages in the Protectorate, for such 
places are close to good soil, to water, and to trees, the three main 
desiderata of the Rhodesian cultivator. The old alluvium, on the other 
E2 
