New Evgine for RuUikj. 63 



so as to projec-t below the inlet at least one inch. This device will 

 prevent the withdrawal of any coarse sediment from the bottom of 

 the container in case the clamp should not have been replaced in 

 .time. After syphoning off to a safe level the clamp is replaced, the 

 tube and float withdrawn and a further supply of untreated sedi- 

 ment and water poured into the container and the syphoning 

 conducted as before, afterwards water only is added and the 

 -syphoning continued so long as any considerable proportion of 

 abrasive remains in suspension. The water withdrawn through the 

 syphon carries with it most of the fine material suspended in the 

 container; the bulk of the heavier particles, maintaining their 

 initial advantage in descent during the stated interval, make their 

 way to the bottom of the vessel. By this method of treatment, as 

 will readily be perceived, a much larger proportion of fine material 

 can bo withdrawn in a given time, while the regulated discharge 

 ensures a more uniform grading. 



Respecting the fine material thus withdrawn, the vessels contain- 

 ing it may be left to settle until the water appears comparatively 

 clear, when it may be poured off and the sediment transferred to 

 an evaporating dish preparatory to drying off. Before drying, 

 the material should be examined under a microscope and the average 

 size of the grains determined, in case it should be neecssary to 

 repeat the separation process, with a slightly longer time interval, 

 in the event of too great a disparity between the size of the various 

 grains. 



Much the same mode of procedure as the foregoing applies to 

 each of the crude separations first made, with correspondingly 

 coarser or finer jets and longer or shorter periods for settlement. 

 It must be borne in mind in this connection, that a small quantity 

 of carefully separated abrasive is far more effective and uniform 

 in its action than many times the same quantity imperfectly 

 treated. Hence, though the process of separation may appear 

 tedious, it will prove to be a wise economy to carry it through. 

 Pour grades only of any particular abrasive thus treated are all 

 that are really necessary to complete the grinding of a precision 

 screw, providing the thread has already been carefully cut. With 

 a view to present to the reader a graphic representation of the uni- 

 formity in size of grain attainable by this mode of separation, a 

 series of photo-micrographs with a magnification of approximately 

 75 diameters has been prepared and are shown as Plate XII.. 

 Figs. 1 to 6. Respecting tliese photographs it is first necessary 



