METEOROLOGY 97 



Germans before the war. Two papers by Range (191 5) and Reenen 

 (1925) are valuable. In spite of the aridity due to the absence of 

 considerable rainfall, the climate is somewhat damp by reason of 

 the copious night fogs and dews caused by the cold Benguela 

 current. The provision for aviation appears to be undeveloped; 

 upper air research has been confined to fifteen months' intensive 

 observations with pilot balloons at Walvis Bay (Cox 1934, p. 230), 

 and similar observations are now in progress at Marienthal: all 

 these w^re carried out by the Union meteorological department. 

 Since the air service in this territory is run by Union Airways, this 

 seems to be a case where centralization of meteorological services 

 is desirable. 



In Southern Rhodesia the official meteorological service is part of 

 the Irrigation Division of the Agricultural Department. The col- 

 lection of climatic information dates from 1897. There are two 

 first order observatories at Salisbury and Bulawayo, where inten- 

 sive studies have been made at high altitudes with pilot balloons^ 

 and the results have been considered v/ith the South African data 

 by Cox (1934). More than forty other stations, well distributed 

 over the country, observe temperature and pressure in addition 

 to rainfall, and there are about 550 stations for rainfall only. The 

 records are made entirely by Europeans, a system of regular 

 inspection is in force, and the results are published in monthly 

 bulletins and annual reports (the latter since 1900). A system of 

 daily forecasting and the preparation of weather maps has been 

 in operation since 1923 (Robertson 1927; Sellick 1934), and more 

 recently a twice-daily service including upper wind observations 

 for aviation has been established. The radio exchange of synoptic 

 reports is in operation with Northern Rhodesia, Mozambique, 

 Madagascar, and the Union. 



A special feature of the service is the preparation of seasonal 

 forecasts according to Sir Gilbert Walker's methods. The pre- 

 diction of excess or defect in rainfall was published in ten years 

 between 1 922 and 1 934, and in eight of these the anticipation proved 

 correct. Sir Gilbert Walker (1933) points out that meteorological 

 conditions are more persistent in tropical than in temperate coun- 

 tries, and concludes that prediction is, accordingly, more likely to 

 be successful in the former. Since in southern Africa the upper 



