44 ATMOSPHERIC DUST. 



consisted of a mixture of daily catches taken during' pai't 

 of June and part of July in 1895. These were thoroughly 

 mixed before the analysis was made. Two samples which 

 were taken, one on the upper cloth and one on the lower, 

 on the nineteenth of August the same year, were sep- 

 arately examined, as was also some other material col- 

 lected in the same manner under some trees in a o-rove 

 about a qiiarter of a mile from the pole pr-eviouslj- 

 referred to. The dust taken in this way resembles per- 

 fectly that AA'hich was caught on the broom-corn. The 

 percentages of the several grades correspond almost to 

 within two perx-ent iu the two averages (Tab. XXVII). 



It will be noticed that the composition of the mixed 

 sample for June and July is very much like the average 

 for the dust taken on the broom-corn iu June, but it is 

 somewhat coarser than that taken on muslin in August. 

 The dust taken undei' the trees in the grove is also a little 

 coarser than the latter. 



Another devise for collecting dust from the atmosphere 

 consisted of a hollow cylinder, with apertures on the side 

 for receiving the wind, and with strips of muslin suspended 

 inside. These strips as well as the inner surface of the 

 cylinder wei'e washed once a week, and adhering particles 

 thus secured. Eight samples were taken by this method 

 during the months of July, August and September in 

 1895 (Tab. XXYIII)). The cylinder was suspended at 

 the same height and f]-om the same flag pole as the 

 broom-corn and the muslin previously mentioned. In 

 this series of samples, also, there was a con-espondence 

 between the wind velocities and the quantities of dust 

 caught, though not so well marked as in the other in- 



