438 On Attraction due to the Flow of Liquids. 



passage of vibration being more rapid in water than air, then 

 globules of mercury, bromine, and bisulphide of carbon, all of 

 which have a greater velocity for sound than water, should 

 either float or show some tendency to do so. A great number 

 of experiments, under varying conditions, failed to show that 

 they were in any way affected. 



As an air-bubble sinks as rapidly as a grain of sand would 

 do, I tried hydrogen in place of air, and mercury in place of 

 water in the tube A. When the hydrogen reaches the mercury 

 it rests a little, and then suddenly sinks with the same appa- 

 rent velocity as the air in water. 



If in these experiments the bubble is large it does not de- 

 scend; and this points to the close analogy with the experi- 

 ments of Guthrie (Proc. Roy. Soc. 1870, pp. 35-41), who 

 found that the vibration of a solid body caused the " approach " 

 of a piece of cardboard from a considerable distance. We 

 might say, conventionally, that the attraction was less between 

 the air and the sounding body than the cardboard and the 

 latter. Now in the tube A, C being a node, any particle near 

 C is relatively at rest to one nearer E, and the attraction be- 

 tween them is greater than either for the bubble. Hence the 

 particle at C tends to move towards E, or, what is the same 

 thing, the bubble moves to 0. When the bubble is large it is 

 still repelled, but it leaves no way for the approach of the 

 attracted water. A great number of small bubbles may be 

 successively sunk till a large one results at C, showing that 

 the conditions of the experiment only proscribe the motion of 

 the large one. By inverting fig. 5 it may be used to produce 

 the water-figures; but they can seldom be obtained free from 

 harmonics by such means. 



On the Attraction due to the Flow of Liquids from an Expanded 



Orifice. 



M. Lacoutre has shown in Les Mondes for May 10, 1866, 

 that when a liquid flows perpendicularly from a cylindrical 

 tube upon a flat disk, its motion is arrested till the disk is 

 brought to within half a diameter of the tube (but varying 

 with pressure), when the outflow^ is greater than if the disk be 

 entirely removed. I find that if the tube terminate in a similar 

 disk, attraction ensues at the position of greatest outflow. 



A glass filter funnel containing a light cone with a weight 

 dependent from its apex, or a tube terminating in a hemi- 

 spherical cup with a light ball, have more stability than the 

 disks, and answer equally well. In all cases the apparatus 

 may be immersed in liquid and the attraction continued. 

 The explanation usually accepted of the corresponding pneu- 



