1/2 CHEMICAL COMPOSITION OF RAIN. 



the Cape Province which have their rainy seasons respectively in 

 winter and in summer. Subsequent consideration, however, led 

 to an endeavour to embrace the whole Union within the scope of 

 the investigation. The collection of rain water tyi:)ical of each of 

 the four Provinces of the Union was therefore aimed at. In the 

 Transvaal I decided, on Mr. Innes's suggestion, to have rain 

 water collected (T) at the Johannesburg Observatory, (2) at the 

 Potchefstroom Experimental Farm, and (3) at the Government 

 Nurseries at Ermelo. On Mr. Stewart's advice, the following six 

 collecting centres were resolved on for the Cape Province : Re- 

 treat, Mr. Stewart's residence (Cape Peninsula), Calvinia, 

 Douglas, Kokstad (Mr. H. D. Coyte), Grahamstown (gaol), 

 and Cradock (gaol). In the Province of Natal, Weenen (Ex- 

 perimental Farm) was selected as representing Upper Natal, 

 Cedara (Government School of Agriculture), as typical of Cen- 

 tral Natal, and Durban Observatory for the coast region. In the 

 Orange Free State the centres were Bloemfontein, jModderpoort, 

 and Lindley. 



The method of collecting the rain water for analysis pre- 

 sented problems from the very start, not due only to the circum- 

 stances above alluded to, but owing as well to the climatological 

 and other conditions of the country. In order to obtain a suf- 

 ficient supply of rain water for monthly analysis by means of the 

 eight-inch gauge, a monthly rainfall of about three inches would 

 be needed. Rut if the whole Union was to be represented by the 

 rain typical of each rainfall district, some of these arid areas 

 would not afford much more tlian three inches of rain in an 

 entire year. In any case, small rainfalls had to be considered, 

 and so the course was adopted of mixing as one sample the rain 

 collected during two or more months in succession. 



On account of the long storage involved, it appeared un- 

 desirable to adopt this course more often than necessity com- 

 pelled, but. on the other hand, it was impracticable to provide a 

 sufficient number of gauges of the dimensions necessary to over- 

 come the chemical difficulties occasioned by small rainfall. It 

 was therefore decided in extreme cases to confine tlie determina- 

 tions to nitrogen. 



Another difficulty was involved in the dust storms frequent 

 in certain parts of the Karroo, as well as in the more northerly 

 Provinces of the Union, which would cause the accumulation of 

 large quantities of dust in the gauges during the high winds that 

 frequently precede ram. This would require the exercise of con- 

 stant care to keep the gauge free from dust up to the very instant 

 that a shower commenced. Such dust is frequently calcareous, 

 and would therefore im])art to the rain an alkaline reaction, while 

 in the neigbourhood of great alkali tracts the dust might even 

 contain nmch soluble salts. Of course, it will be sufficiently 

 recognised that, quite apart from dust-bestrewn gauges, first 

 rains would, in the districts mentioned, necessarily encounter 

 considerable c|uantities of calcareous and alkali dust in their mere 

 passage through the air. 



