416 W. M. DAVIS — OBSERVATIONS IN SOUTH AFRICA 



rangement of the seasonal rainfall, so long as the zones have their present 

 positions; and herein lies one of the most peculiar as well as one of the 

 most difficult elements of the Dwyka problem. The Pleistocene ice- 

 sheets of northeastern America and northwestern Europe are sufficiently 

 explained by a moderate intensification of the present winters. Eecent 

 studies of Alpine glaciation show that it can be accounted for by 

 a lowering of the snowline such as would result from a reduction of the 

 mean annual temperature by a few degrees centigrade. The Dwyka 

 glaciation is peculiar in that an intensification of its winter climate would 

 leave its area free from ice. It is the summers of the Dwyka area that 

 must be made wintry before an ice-sheet can be provided; and then a 

 still colder winter, dry and clear, would, intervene between the snowy 

 summers. This may perhaps be made plainer by considering in some 

 detail various possible changes of factors which affect climate, such as 

 land area, land form, ocean currents, and seasonal migration of the wind 

 belts, in order to determine whether any other change than general re- 

 frigeration would effect the desired result. It should be remembered that 

 under the last of the headings here named it is not admissible to shift 

 the various belts of the terrestrial wind system arbitrarily, for they are 

 known to be intimately dependent on the path of the equator and on the 

 position of the poles. It is no more reasonable to postulate an arbitrary 

 change in the position of subtropical belts by which the winter rains are 

 furnished to the southern coastal border of Cape Colony than arbitrarily 

 to tilt the axis of the earth into a new geographical position. The ques- 

 tion is, then, to learn whether any change in the lands, currents, or winds, 

 independent of changes in the intensity of insolation or in the composi- 

 tion of the atmosphere, would produce snowfall in the Dwyka area. 



Changes of land area and form. — If South Africa had been somewhat 

 larger in Permian time, the trade winds would have been all the better 

 established ; for the greater the area of an equatorial continent the higher 

 the temperature of its equatorial belt and the stronger the indraft of the 

 trade winds. On the other hand, the larger the land area of a trade wind 

 region, the drier it will be, provided its altitude is low. There would 

 probably, however, be an increased migration of the heat equator on an 

 enlarged continent ; but this could only bring warm summer rains toward 

 the Dwyka area. It is quite possible that, if equatorial Africa were sub- 

 merged while South Africa were enlarged, extensive monsoons would be 

 established and the equatorial rains might be carried 20 or 30 degrees 

 south of the equator, for they are carried some such distance north of the 

 equator in India; but, in order that snow should be gathered under the 

 south-shifting equatorial cloud belt with the present distribution of mean 



