182 MESSRS. A. M. WORTHIXC4TON AND II. S. COI.K 



proverbially difficult to prove a negative, the conclusion was finally forced upon us 

 that this electrification, whether positive or negative, had no certain or direct 

 influence. We also tried the effect of holding a charged ebonite rod near the splash, 

 but with negative results. 



One test we wei-e able to make which appears to us to be crucial. If the flame acts 

 by diselectrifying the sphere, then the same result should follow from letting the 

 sphere touch in its fall an earth-connected brush near the water, but such contact had 

 no observable effect whatever, and we therefore came to the conclusion that the action 

 of the flame was not an electrical discharging action at all. 



In this connection also may be mentioned two other facts : ( 1 ) That the flame had 

 no observable effect on a roughened sphere ; and (2) that we could not detect any 

 accumulation of electricity when we let the sphere fall time after time through a tube 

 connected to an electroscope, to which tube the sphere could give up its whole charge 

 by touching a wire brush in the interior. 



Photographs of the Transition from " Smooth " to " Rough." 



Having found that the splash passed by gradual transition from "smooth" to 

 "rough" as the height of fall of a polished sphere was increased, we decided to 

 obtain a photographic record of the process. Series XIII. , Sheet 3, of which a 

 selection is here reproduced on Plate 2, shows a perfectly smooth splash produced by 

 a highly polished sphere of serpentine, just over 1 inch (2 '57 centims.) in diameter, 

 falling into water from a height of between 14 and 15 centims. 



The earlier figures of this series show how extremely thin is the enveloping sheath 

 in its early stages. In fig. 1 it is almost best seen by its reflected image in the 

 undisturbed surface. A nearly horizontal row of minute drops may be seen in 

 figs. [2] and 3. As we shall see later, the place of origin of any one of these is to be 

 found very approximately by drawing a tangent to the sphere through the drop 

 in question in the plane containing the axis, for unless the velocity of the sphere is 

 very materially altered after the moment of separation, the drop will remain on the 

 tangent along which it was projected. The " lug " at either side in figs. [2] and 3 is 

 probably not really the continuous jet it seems to be, but merely the fore-shortened 

 edge of a ring of separate droplets, as is more apparent in figs. [4] and 5. Where the 

 horizontal equator of the sphere is passed, the film will thicken by convergence, and 

 the number of segmentations will diminish accordingly, the drops becoming larger 

 and fewer. It will be noticed that in each of the figs. 5 and 6, a vertical tangent 

 to the sphere marks the limit of the larger drops in the air above. 



Fig. 8 shows by means of bubbles that the displacement of the liquid corresponds 

 approximately to stream-line flow in a perfect fluid. 



Series XIV., Sheet 4, of which fig. 1 is given on p. 183, shows the effect of increasing 



