CHKMICAL COMPOSITION OF RAIN. 1/5 



chlorine, to i.ooo c.c. of rain water, and to concentrate the water, in an 

 open basin, in a qniet room, to less than 250 c.c. After cooling, the water 

 is filtered through a filter previously washed with chlorine-free water, into 

 a marked 250 c.c. flask. The Flask is then filled up to the mark, and the 

 contents used for determination of chlorine in the usual way. 



Sulphur trio.vide. — For determining sulphates, it was proposed to col- 

 lect the rain water for six months in succession, so as to make up one 

 special sample of from four to seven litres of water. The rain used for 

 this exclusive purpose is collected through a glass funnel having a small 

 piece of asbestos cloth in the pipe no vulcanised caoutchouc being employed 

 for connections. The combined' quantity of 4 to 7 litres is concentrated 

 in a retort to a small bulk, and filtered through Swedish filter paper pre- 

 viously washed with hydrochloric acid. In the filtrate, heated to boiling, 

 sulphur trioxide is then gravimetrically determined by the addition of 

 boiling barium chloride solution. The concentration in a retort is esseri- 

 tial, because an open gas flame is a constant source of sulphuric acid. To 

 avoid this the use of an alcohol flame was suggested, which would permit 

 of a platinum dish being substituted for the glass retort, and thus avoid 

 the possibility of contaminating the Barium sulphate precipitate by silica 

 dissolved out from the glass. 



Ammonia. — The retort and condenser are freed from ammonia by 

 boiling some distilled water with a little magnesia in the former; after 

 which one litre of rain water is introduced and boiled with a small 

 quantity of magnesia until 250 c.c. have distilled over. In this distillate 

 duplicate determinations of ammonia are made by Nesslerising. 



Nitrogen as nitrites and nitrates. — The 750 c.c. of boiled rain water, 

 remaining in the retort after determination of the ammonia, are trans- 

 ferred to a wide-mouthed stoppered bottle supplied with six strips of 

 zinc foil converted into couple, and placed in an incubator at 21° to 

 24° for three days. The copper zinc couple used is prepared as follows: 

 Six strips of zinc foil, four inches long by i] inch wide, are bent at right 

 angles along their middle in order to obtain stiffness. The couple is then 

 further prepared in a series of five beakers, containing respectively (i) 

 a dilute solution of sodium hydrate {2) very dilute sulphuric acid, (3) a 

 three per cent, solution of copper sulphate (4) ordinary distilled water, 

 and (5) distilled water free from ammonia. Through these five beakers 

 the zinc foil is successively passed. It is rinsed both after the alkali and 

 after the acid, but after the copper has been deposited the strips are simply 

 drained, and carefully placed in the distilled water, as it is diflicult to 

 rinse without removing the copper. The couple shotdd be entirely sub- 

 merged when it is placed in the rain water. 



It wa.s not pos,sible to arrange for commencing the collection 

 of the rain water at all the designated localities simultaneou.slv^ ; 

 in fact, nearly a year had passed after the scheme of investiga- 

 tion had fir.^t been mooted before any practical beginning conld 

 be made at all. When the first results began to come in — those 

 froni rain collected at Cedara, Natal — they seemed to show that 

 the yearly amounts of nitrogen would be distinctly higher than in 

 England, and Dr. Miller suggested the possibility that, owing to 

 the higher temperature, the soil in Natal gives off more ammonia 

 At the satne time, the absence of ammonia from the Cedara water 

 in November, igio (and from the Grahamstovvn rain of October, 

 1912), suggested nitrification, and this in turn led to the sugges- 

 tion that such nitrification might be obviated by the addition of 

 a small amount of a lead salt to samples liable to contain dust 

 and nitrifving organisms. Another early modification of the 

 collecting arrangements arose from the presence of leaves in the 

 Durban samjile. Though, to judge by comparison between the 



