DEVELOPMENT OF THE COASTAL BELT OF SOUTH AFRICA. 21 3 



Table i gives the list of places shown on the map together 

 with their elevation above sea-level and their average annual 

 rainfall. From this list it will be seen that the 40 inches rainfall 

 usually commences above i,ooo feet above, sea level, and increases 

 with the elevation. The few records we have in the coastal 

 ranges, alxjve the 3,000 feet level, show that the rainfall usually 

 exceeds 50 inches. Take, for instance, Wemniers Hoek, Berg 

 River Hoek, Table Mountain, Hogsback, Evelyn Valley, Baziya, 

 Tabankulu Forest Station, and Pietermaritzburg. In looking at 

 the sketch map, there is apparently a gap in the heavy rainfall 

 about Port Elizabeth. But this may be due to lack of records 

 rather than lack of rainfall. In other words, there are no rain 

 gauges kept up in these mountains. This is borne out by the 

 rainfall figures in the Annual Report of the City Engineer for 

 Port Elizabeth for the year 191 2, which was about an average 

 year in , the district for rainfall. The figures are as follows: — 



Sand River Area, Gauge at Reservoir 27.34 inches, average 

 of mountain gauges 42.29 inches. Bulk River Area, Gauge at 

 Reservoir, 38.82 inches, average of mountain gauges 45.30 

 inches. But for these mountain gauges, we should have no in- 

 formation as to there being 40 inches of rainfall in this di'strict. 



The distance around the rain-belt line of the Union from 

 Ceres to the Portuguese border is, roughly, 1,000 miles. The 

 author has made a careful study to ascertain the width of this 

 belt of intense rainfall, being personally acquainted with a good 

 deal of the coastal belt, and has gone through a large number 

 of reports and statistics, all of which would take up too much 

 space to include here. The conclusion he has come to is that the 

 width varies from 10 to 30 or 40 miles, and even more in parts 

 of Pondoland and the Transkei. Where the mountains rise 

 abruptly, the width is narrower, but the rainfall more intense. 

 Taking all the facts into consideration, it seems highly probable 

 that there is a belt of country from Ceres to the Portuguese 

 border, at least 20 miles in width, with an average rainfall of 

 40 inches per annum. Nevertheless, the author proposes in this 

 paper, to be on the safe side, to base his calculations on a belt 

 only 15 miles wide, which gives a total catchment area for this 

 belt of 15,000 square miles. 



We have first to consider what is the run-oii" from this area. 

 For this we possess very little precise information, so the author 

 proposes to take the records of the Perie catchment area of the 

 Buffalo River for the past eight years as the basis for the general 

 run-off of this rain-belt. It may be suggested that the Perie 

 catchment area is abnormal, and that it is not fair to take it 

 as a sample of the rain-belt, but the author knows of no evidence 

 to support this. For instance, if we take the western comer of 

 the rain-belt, we find that Table Mountain gives a run-off three 

 times as great as that of the Perie catchment area, and similar 

 results were obtained from measurements of the run-off from 

 the Berg River and Wemmers Hoek catchment areas. While 



