123 



glazed in front of about 5mm depth. The pins however, 

 which give equal déviations, hâve been replaced by what, 

 for the sake of brevity, I will call deviators. They are 

 pentagonal pièces of wood having one side horizontal, 

 two sides perpendicular and the two upper ones incHned 

 under a fixed angle (45°) to the horizon. 



For the sake of compactness that part of some of the 

 deviators at the right hand side of the vov^js, which is of 

 no practical importance, has been eut ofF. The deviators 

 of the same row hâve their bases on a horizontal line. 

 The consécutive rows contain 3, 4, 5 . . . . of thèse 

 deviators. Their breadth, measured from the middle of 

 the channels between them, has been taken proportional 

 to the distance of their tops from the left hand side of 

 the frame, which is vertical. Likewise, though this is not 

 essential, the breadth of channels between the consécutive 

 deviators has been taken proportional to the distance of 

 the middle of the channel to the same side of the frame. 

 The tops of the deviators of any row hâve been placed just 

 below the middle of the channels of the preceding row. 

 With an infinité or at least a very great number of 

 rows, we might, without changing the final resuit, hâve 

 placed them in any way excentrically in regard to thèse 

 channels. 



At the top of the machine, just above the top of the 

 middlemost deviator of the Ist row, has been placed a 

 funnel. If we fiU this funnel with fine sand, the grains of 

 the sand will fall on this deviator and will be deviated 

 one half to the right the other to the left. Arriving at 

 the second row, both parts will again be devided in equal 

 parts. The left hand part, however, will not now be so 

 much deviated as the right hand part because of the 

 différent breadth of the deviators. As is évident from the 

 construction of the apparatus the déviations will always 

 be proportional to the distance from the left hand side 



