190 On a Multiple Induction Machine. 



siphon glass tube, the water entirely ceased to pass. I am therefore 

 led to believe that the intervention of capillarity is essential to the 

 production of the effect. 



The tin cup is not essential to this experiment, and was merely 

 used to afford a separate lodgment for the water brought over from 

 the opposite side, and at the same time make the best possible contact 

 with the water it contained. 



Fig. 7 shows this experiment in a still more striking form. In 

 this case the positive glass was dispensed with, and the upper end of 

 the cotton coil was passed into a glass bulb through a nozzle, the 

 aperture in which was just sufficient for the cotton to fill without 

 appreciable friction in moving. A knot was made at the end of the 

 cotton to prevent its dropping back through the aperture, and water 

 was then poured into the bulb until nearly full, and a cork, with the 

 conducting wire through it, was inserted. The bulb was held over 

 the metallic cup with the nozzle dipping into the water. When the 

 current was turned on the cotton climbed up vertically, and continued 

 to pass until it filled the bulb. Even when the nozzle was lifted a 

 quarter of an inch clear of the water the cotton continued to travel. 

 The smallest particle of salt added to the water, or a minute quantity 

 of anything that increased its conducting power, destroyed the effect 

 and caused gas to appear at the electrodes. 



The only experiment that I know of that presents any analogy to 

 the effect thus obtained is that which appears to depend upon what 

 is called electrical endosmose, in which case a porous diaphragm has 

 the power of transmitting water from a cell containing a positive 

 electrode into an adjoining one containing a negative electrode. If 

 we may assume the capillarity of the cotton to represent the porosity 

 of the diaphragm, it is reasonable to suppose it capable of transporting 

 the water. Then as to the cotton travelling in the reverse direction, 

 that may possibly be due to the reaction attending the transmission 

 of the fluid. At all events it would appear that capillarity is con- 

 trollable by electricity, and the question arises, what is the relation- 

 ship between the one and the other ? 



The difference in the action of the current in this experiment and 

 in that which I made with the hydro-electric machine consists in 

 this: That in the early experiment one thin cofctbn thread was 

 moved with a high velocity, while in the latter experiment a mass of 

 cotton equal to more than 100 similar threads was moved with a 

 low velocity. This difference is probably due to the current being 

 continuous in the one case and pulsatory in the other, also to its 

 being of higher potential and smaller volume in the hydro- electric 

 than in the multiple machine. When I attempted to use a single 

 fine thread with my present machine the water upon it immediately 

 boiled, and the thread was destroyed by sparks. 



