524 Professor Lodge [May 28, 



on a glass surface when tbe end of a hot cylinder is held against it 

 is similarly completely explicable. 



But, it may be asked, if dust gets driven towards cold bodies 

 instead of away from them, how is it that any dark or dust-free 

 plane forms beneath them ? At very low temperatures I believe it 

 does not, but rather a bright or dusty layer forms instead. From a 

 rod a few degrees below the air a fine dark plane is visible, however, 

 edged by two bright ones. I do not know what Lord Eayleigh's view 

 of this descending plane is, but it may be due to the gravitative 

 settling of the dust through the air immediately beneath the cold 

 body: a thing which is shown to occur by the existence of a thin 

 dark half-coat formed underneath, but not above, such a rod, and by 

 the deposit of dust on its upper surface. The rod shelters the 

 air immediately beneath it from the shower, and so a layer of clear 

 air forms and streams downward continuously. If the downward 

 convection currents are too rapid the settling has not time to occur, 

 and no dark plane is visible. 



The original plan of experiment included a series of measurements 

 of the thickness of the dark-coat at different temperatures (both 

 excess and absolute), at different pressures, and in different kinds of 

 gas. This research Mr. Clark had indeed begun to arrange for, but 

 his untimely death last year cut short this jjart of the investigation. 

 Another very necessary thing is to see how it varies with size of 

 dust particles. Meanwhile I had devoted myself to developing a 

 branch of the subject which we had accidentally hit upon in the 

 course of testing one hypothesis which had at an early stage occurred 

 to us as perhaps a clue to the cause of the phenomenon of the dust- 

 free plane. We thought it possible that air in streaming over 

 the surface of the solid might get electrified, and that, from air so 

 electrified, dust might somehow be expelled. To test this hyjDothesis, 

 we purposely electrified the rod, positively and negatively, to see what 

 happened. A hundred volts or two produced a barely noticeable 

 effect ; positive electrification causing a slight widening, negative 

 electrification, a slight narrowing, of the dust- free coat. But as soon 

 as the potential rose to a few thousand volts, and brush discharge 

 began to be possible, a very violent and remarkable effect was 

 noticed : the dark coat widened enormously and tumultuously, and 

 the whole box was rapidly cleared of smoke. 



To specially observe this new phenomenon is very easy. Fill a 

 bell-jar with any kind of smoke : tobacco, camj^hor, turpentine, mag- 

 nesia, amnionic chloride, ammonic sulphite, brown paper, steam, 

 phosphoric oxide, lead fume, zinc fume, no matter what, and then 

 discharge electricity into it from a point connected with a Voss or 

 Wimshurst machine, or even with a common frictional machine, the 

 other pole of which is connected with the ground or with the base of 

 the bell-jar. 



In a second or two, aggregation of the smoke particles sets in, 

 they form in masses or flakes along the lines of force, and in another 



