METALS IN THE ATMOSPHERE. 243 



his <)l)S(M-vations on the ni('t(M)rito of Oi-ffiu'iK some of these particles 

 coiiie fi-oin the exj)I()siou of meteorites or from their simple disunion 

 wlien frial)le and prone to dlsintei^ration. Assuming that their mean 

 diameter is a constant of 0.01 mm., which, in fact, is a rather too lib- 

 eral estimate, it would take 2,."')00 of then) to cover a sfpiare millimeter 

 and 250.000 to equal in l)nlk a cubic millimeter. Therefore it is easily 

 appreciated that the deposit of iron on the earth's surface from this 

 source even in a considei-abU' interxal of time will not l)e fjreat. 



But, admittinii' that some of the iron dnst is of extraterrestrial 

 oriijin. it is equally true that a large part of it is swejit from the 

 earth's surface l)y the wind or carried uj) in the smoke of the foun- 

 dries. In the neio-hborhood of these works may l)e collected globules 

 of magnetic iron oxide, which, rising as sparks, took the globular 

 foru) in cooling. It is easily shown that bits of iron at high temper- 

 atures become spherical, and that a mass of it combining at red heat 

 with oxygen will divide into microscopic globular fragments, ^^'hen 

 \'ery hue iron filings are made incandescent l)y jjassing tlii"ough a 

 liydrogen fianx', they burn brilliantly. 'J'issandier has discovei'ed, 

 by collecting th(* products of this combustion on a porcelain jilate 

 and examining them with a microscope, that they ;iiv in the foi'm of 

 sph(M"ical globules, vase-necked spheres, irregular surfaced, or fibrous 

 fragments inc()m})letely fused. Powder obtained by striking a piece 

 of iron with a flint is made uj) of the same kind of globules. An ii'on 

 wire burning in oxygen will form globules of magnetic oxide visible 

 to the naked eye and at the same time others much smaller, which 

 uiay be collected in water at the bottom of the flask. These are visi- 

 ble only under great magnification, since their diameter rarely exceeds 

 0.01 of a millimeter. 



Idle coml)ustioii of coal in factories furnishes the air an abundance 

 of iron oxide from the dec()mj)osition of ferruginous [pyrites contained 

 in the coal. But all these pai'ticles obtained in the various ways men- 

 tioiunl, whether from the ciMubustion <d" iron, coal, oi* other substances, 

 are easily distinguishable from those of cosmic origin by the fact that, 

 they never contain nickel in any form. 



TIIKmI land and salt water, whence gusts of wind snatch u]) parti- 

 cles the more ininiite of which measui'e scarcely 0.001 mm., often con- 

 tains magnetic dusts in comi)arative abundance. A magnet passed 

 over their surface attracts tiny grains of magnetic iron oxide. This 

 is entirely independent of the particles due to the contimial destruc- 

 tion of enormous (piantities of iron in manufacturing. The dust 

 from pulverized magnetic iron ore and other feriaiginous minerals, 

 or the dust formed by oxidation in o})en air or from fresh or salt 

 water, iievei- occurs in the mammilated, fibi'ous, or spherical forms. 

 There are amorphic grayish powders which do not resemble those 

 of planetary origin and are very different from those j^roduced at 



