238 METALS IN THE ATMOSPHERE. 



on June 1, 10, and 11, 1870, amounted to from 23 to 7-5 mg. per liter; 

 at Paris it has run from 23 to 172, the maximum being 421 mg. This 

 resi(hie is hhickish gray, excei)t that collected in the fields, which is 

 entirely white, and it invariably contains the same relative propor- 

 tions of mineral substances and organic matter. 



The size of snowflakes and the leisurely manner in Avhich these 

 little spongy masses fall through the air makes them even better 

 fitted than rain drops to seize in their passage all the dust particles 

 and solid Imdies floating in the atmosphere. Moreover, solids are col- 

 lected in the waters fused into the flake. Conse(|uently, when Mon- 

 sieur Tissandier collected Avith all proper precaution the first snows 

 that fell on the towers of Notre Dame in the winter of 1875, he found 

 in each liter of the snow water a body of corpuscles varying in weight 

 from 5() to 118 mg. A liter of melted snow collected under the same 

 conditions in the country contains from 48 to 104 mg. Further- 

 more, as one might supi)ose, the corpuscles are less numerous after a 

 prolonged snowstorm, so much so that in Paris, after a heavy fall, 

 only l(j to 24 mg were found. The residu(^ obtained by the evapora- 

 tion of melted snow is ordinarily an impalpal)le grayish powder, con- 

 taining, at Paris, about 57 per cent of cinders, and in the country 

 about (')1 per cent. 



Ilail, because of its small size and its great density, does not collect 

 the dust particles so easily. Nevertheless, it has been collected and 

 examined in the same way. 



Iran. — The examination of the cinders in the dust collected in these 

 vaiMovis ways enables us to recognize in our atmos})here the i)resence 

 of a numl)er of nu^tals, the most important of which is iron. 



When a strong magnet was held near some of the atnu)spheric sedi- 

 ment thus obtained, a portion of the c()rj)uscles adhered to it and were 

 brushed off for microscopic or chemical examinations. It was then 

 discovere<l that these bits are nuule up essentially of iron. The same 

 results were obtained with sediments collected in several very differ- 

 ent localities. The examinations even went so far as to estimate ap- 

 proximately the (|uantity of iron contained in this niagnetic residue 

 -by the intensity of the coloration of sulphocyanide of potassium in the 

 dissolution of a known (juantity of dust. These ferruginous pai'licles 

 Aven^ found to be either [)ure iron or tliat metal associated in certain 

 ])r()p()rtions with other t'lcMuents, such as nickel and phosphoi'us. M. 

 N()rdenski<)l(l, at Stockholm in 1871, examining the surface of the 

 greatest fall of snow within the memory of man, foun<l small quanti- 

 ties of nu'tallic iron. l)Ut feai'ing this might have come from neigh- 

 boi'ing roofs, he had his brother examine the snow in a desolate plain 

 surrounded by the forests of Finland. The black powder seciu-ed 

 there was of the sauu^ character as that of Stockholm. The particles 

 of iron drawn out bv a magnet, when ti'iturated in an a^ate mortar, 



