58 ATMOSPHERIC DUST. 



quantity in the atmosphere, even in calms, cannot escape 

 being earned a hundred miles or more in a strong- mnd. 

 ]SIedium dust must be capable of being transported still 

 farther and fine, and very fine dust evidently settle with 

 great slowness even in perfect calm, unless present in such 

 quantity that flocculation will take place. This probably 

 seldom occurs excej)t near places of active wind erosion. 

 If we now take a review of all the analyses of atmos- 

 pheric dust here presented, that artificially collected as 

 weU as the storm dust, vre notice that the maxima are 

 scattered over three gi-ades. In sixteen samples the 

 maximum occurs in the coarse dust, in two it is right be- 

 tween this and the medium dust, in thii-ty-eight of the 

 samples it occure in the medium dust, and in one it is in 

 the fine dust. This one sample was collected fiom dried 

 foHage exposed to the winds for several months, during 

 which time a large proportion of the coarser particles 

 had, no doubt, been dislodged. In all the cases where the 

 maxima consist of coarse dust (except j)erhaps nos. 152, 

 153). special conditions of collecting account for the 

 greater quantity of coarse materials. The diversity in 

 composition of the atmosiiheric dust is hence more 

 apparent thcin real. In two of the samples ninety j)er 

 cent is disti'ibuted among five different grades : in seven- 

 teen samples, among foiir: in thirty-six. among thi'ee 

 grades, and in one sample it is divided between two. Tlie 

 average position of the precise maximum (as we may 

 designate that length of diameter, which, if taken as a 

 limit for separation, would divide the bulk of the dust 

 into two equal parts ) appears to be a little below but 

 not far removed from the limit between the coarse and 



