78 Repoht 8. a. a. Advancemknt of Science. 



and only in the south-west of the colon}' is there any large area subject 

 to a winter rainfall. A narrow strip along the south coast from Cape- 

 town to East London receives its rainfall fairly evenly distributed 

 throughout the twelve months. 



Kini-ojt'. — Very little reliable infurmatiun is as yet available re- 

 garding the proportion of run-uff to rainfall, but it is one of the 

 (juestions now engaging the attention of the Irrigation Department. 

 "Jliere is no doubt that the percentage has a wide i-ange throughout 

 the countr}', attaining a maximum on the mountains which lie within 

 zones of high rainfall, and a minimum on the plains within zones ()f 

 low rainfall. The run-off from any area will vary for difierent times 

 and seasons, being generally highest in the rainy season and lowest in 

 the dr}^ It will be influenced by many factors, e.g. the amount and 

 the intensity of rainfall, the temperature, condition of the ground 

 innnediately preceding the rain, itc, itc. ; of any single factor probably 

 the amoiitit of rain will have the greatest influence, and the amount 

 combined with intensity will certainly always show a predominating 

 influence. It is extremely diflicult to obtain reliable run-off }-esults 

 for even small catchment areas, owing chiefly to the ditticulty of 

 obtaining an accurate mean rainfall on the area, and on a large catch- 

 ment the very local nature of the rainfall renders it more difficult still. 

 Three gauges placed a ver}^ few miles apart at the points of an equila- 

 teral triangle e\en on the level Karroo will seldom give the same 

 record for any day, so that the flifficulty of obtaining reliable results 

 for extensive areas is apparent. As illustrating the variation of run- 

 off, it may not be uninteresting to give a few exaniples applying to 

 Table Mountain. 



The results are shown graphically on the diagrams exhibited. 

 No 1 shows the monthly variations in percentage of run-off throughout 

 the year for total monthl}'^ rainfalls varying from 1 to 10 ins. It will 

 be seen that the run-off varies from to almost 100 per cent, according 

 to the season and to the total amount of the rainfall, e.g. a monthl}'^ 

 rainfall of I ins. in January gives about 16 per cent, of runofJ!, 

 whereas in May the same rainfall would give 80 per cent., a figure not 

 ap})roa(;hed in .Jainiary by any ordinary rainfall. 



Diagram '1 shows for one mc^nth, viz., April, the variations in the 

 i-un-ofl" with variations in the total rainfall for the montli. It is com- 

 piled from diagram 1, and each month of the year )night be similarly 

 treated. Of course it must be understood that the curves represent 

 only mean results, i.e. the -1 ins. of rain which in January gives 

 only 16 per cent, run-off is 4 ins. distributed with averac/e January 

 intensity, and if the same amount fell in one heavy and contiiuious 

 rain no doubt it would, even in January, give a considerably highei* 

 run-off. 



As alreadv remarked, the Table ^lountain s rainfall is not I'cpre- 

 sentative of the rainfall in South Africa generally, although it may be 

 so of the mountain districts in the south-west. A diagram (not re- 

 produced) was exhibited showing the i-ainfall at Gannapan «m the 

 north-west Kai-roo, the mean of fourteen years' observations. 



