igoo-i.] Effkcts of Water on Foliage Leaves. 335 



with salt crystals, due to the residue of solution sprayed upon the inside; 

 and there had been placed inside the jar at the commencement of the 

 experiment a watch glass containing dry salt. Besides this, the salt 

 from the solution in the crystallizing dish had crept up round the margin 

 of the dish, and there had collected in considerable quantity, but only 

 on the half of the margin on the side towards the source of light. No 

 attempt is made to explain this last mentioned phenomenon. The 

 greater increase per cent, of one sample of iron in jar three as compared 

 with the same in jar one in experiments I. a. and II. a. is due to the 

 fact that the specimen in jar one had been fairly well coated with rust 

 from experiments I. and II. ; and so the iron having the smallest 

 amount of rust upon it at the commencement of the experiments I. a. 

 and II. a., made, as might be expected, the greatest gain in the suc- 

 ceeding experiments. 



From these experiments we might conclude : — 



(i) That sea-water causes the atmosphere to produce rust upon iron 

 to a greater extent than does fresh water. 



(2) That the presence of salt (NaCl) causes an accumulation of rust 

 upon iron when the humidity is high — about seventy to eighty — but 

 when the humidity is low — about thirty-five to forty-five — it prevents 

 rust formation. 



(3) That sea-water affecting the atmosphere in maritime localities 

 may also affect vegetation. 



The only investigations of importance relating to the question of 

 chlorides in the atmosphere were carried on at Rothampstead, England, 

 by Lawes and Gilbert (1883), and by Dr. Frankland (1881). Their 

 determinations were made by analyzing the water from rain-falls, and, 

 in the case of Lawes and Gilbert, the experiments extended over a 

 period of six years, 1877-83. The results obtained by these men did 

 not fall into any very regular series and several points were left by them 

 unexplained. Quoting from Law^es and Gilbert : — " In the account 

 given in the earlier results of this investigation it was pointed out that 

 the winter rain-fall was far richer in chlorine than the summer rain-fall ; 

 we are now able to take a step further, and show the general character 

 with respect to chlorine of each month of the year. The minimum 

 amount of chlorine occurs in the rain of July. In August and September 

 there is a distinct but not a very large increase in quantity. In October 



