J. AITKEN ON THE NUMBER OF DUST PARTICLES IN THE ATMOSPHERE. 11 



capacity of 450 c.c, while the volume of the pump is 150 c.c. In making an experiment, 

 one stroke of the pump is made, and after the drops inside the receiver have settled, the 

 stopcock F is opened and air allowed to enter, the air passing through the filter on its 

 way. It was found that with a length of 1 inch (2 "5 cm.) of wool in the filter, the 

 air was very far from being freed from all its dust, as a shower always fell after expansion. 

 But, if the rush of the air through the filter was checked, and it was made to pass slowly 

 through, then all the dust was taken out. More and more wool was now packed into the 

 tube, with the result, that when a length of 3 or 4 inches was attained the air might be 

 allowed to rush unchecked through the filter, and yet be perfectly freed from all dust. 

 Part of this better filtration is the result of the greater length of wool passed through by 

 the air, but very much is due to the checking of the rush of air by the greater length 

 of wool. When more wool was packed into the tube, no difference whatever was detected 

 in its filtering power. It was, however, found that with even 12 inches of wool the 

 filtration, though perfect under the conditions described, was imperfect if the air was 

 made to rush through it more violently. If three strokes of the pump were made instead 

 of one, before the stopcock F was opened, the rush of air was sufficiently violent to carry 

 some dust through even 12 inches of tightly packed cotton wool. 



These results point clearly to the necessity of testing the filter connected with any 

 apparatus, under exactly the conditions under which it will be used. When working with 

 certain arrangements of apparatus, such as that shown in the Plate, where an air-pump is 

 used, it does not matter how much wool we pack into the filter, the pressure employed 

 is so great that the resistance of the filter does not practically interfere with the working 

 of the apparatus, in which case it is advisable to have a great thickness of wool. But 

 when working with the gasometer arrangement, the pressure available for drawing the air 

 through the filter is very small, and a great length of wool would retard the filtering and 

 waste much time. The filter for the gasometer was therefore made of very large area, to 

 reduce the rate per unit of area at which the air passes through it ; and after enough of wool 

 had been packed in to give perfect filtration, when used under the pressure at which it was 

 to be worked, a large amount more was added to give a good factor of safety. In work- 

 ing the gasometer arrangement, it has been customary to test the action of the filter every 

 time the gasometer is being filled with filtered air. This is necessary not only to see 

 that the filter is acting perfectly, but also to make sure that all dusty air has been washed 

 out before we begin to fill with filtered air, as it is difficult to get rid of the last trace 

 of dusty air in the gasometer and its connecting pipes, and this can be done only by 

 washing with filtered air and testing. 



What I have here stated definitely with regard to the filtering powers of different 

 lengths of cotton wool was very far from being clear while the investigation was going 

 on. At that time, what has been stated seemed to be the general conclusion, but any 

 thing like confidence in it was destroyed by the many exceptions. For instance, the 12- 

 inch filter seemed to do its work perfectly, filtering the air from every dust particle, so 

 long as the rush of air through it was limited to that produced by one stroke of the pump. 



