688 RAINFALL IN THE TRANSVAAL. 



After removal of irrelevant observations, 1,557 records 

 remain,, with a distribution over the year as follows : — - 



Jan. Feb. Mar. April. May. Tune. 



319 281 151 83 17 2 



July. Aug. Sept. (!)ct. Nov. Dec. Total. 



— 3 14 119 239 329 1,557 



The curves of Fig. i represent this frequencv distribution, 

 compared with the monthly rainfall. The former increases more 

 rajjidly than the latter in the early ])art of the rainy season, 

 reaching the maximum a month earlier ; a more rapid decrease in 

 March is also evident. A greater ]irol)ability of heavy down- 

 pours during thunderstorms suggests ])eriodicity in the different 

 ty|)es of pressure distribution over the countr\- as an ex])lanati()n 

 of these peculiarities. 



The most remarkable rainfall yet recorded in the 'J'ransvaal 

 occurred at W'^olhuter Kop on February i8th. 191 5, when 4.19 

 inches fell in 30 minutes. Although exceptional, this storm is 

 not without precedent outside the Transvaal. At Curtea-de- 

 Arges, Roumania, on July 7th. 1889, 8.05 inches were recorded in 

 20 minutes ; whilst the United States have on record a fall of 

 8.80 inches in one hour at Palmetto, Nevada, during August, 

 1890, and another of about 11.50 inches in about 80 minutes at 

 Campo, California, during August. 1891. There is some uncer- 

 tainty about this last fall, the raingaugc having been washed 

 away. 



These intense isolated rains, however, seem to be confined to 

 Certain mountainous regions, and their significance is restricted. 

 For the comparison of intensities experienced in dififerent coun- 

 tries, or dififerent areas of the same country, a number of falls 

 .scattered over an extended area furnish a better standard. es])e- 

 cially if they disclose a more or less constant relation between 

 time of duration and the amount of ]:)recipitation. In that case, 

 by taking only the heaviest falls recorded during various intervals 

 of time, some idea may be obtained of the probable maximum 

 amount of rainfall to be expected in that area during a stated 

 period. 



When seeking such a relation in the Transvaal, the decrease 

 in rainfall from east to west of the Province, concomitant with 

 increase of distance from the source of moisture su])ply and 

 alterations in the physical configuration of the country, suggests 

 a variation of intensities with geographical ]>osition. As this 

 evidently exists it is convenient to consider the data under three 

 divisions as shown in the map (Fig. 2), re]:)roduced by kind 

 permission of Mr. Tudor (1. Trevor, A.R..S.M., F.G.S. The 

 eastern slopes of the Main and Zoutpansberg plateaux and part 

 of the north-eastern High \^eld fall under the Fastern Divi- 

 sion ; the South-Western districts under the South-W^estern Divi- 

 sion, and the remainder of the country under the Central. For 

 the unhealthy parts of the Loiy \'eld onlv meagre information is 



