428 Prof. J. J. Thomson and Mr. H. F Newall. [Dec. 17, 



The separate rings always remain connected with each other by 

 threads of liquid ; this shows that there is no surface tension ; for if 

 there were, these threads would break up into drops. The only 

 stage at which the motion is discontinuous is when the drops are 

 changed into rings. The change from the ring shape to stage 2 

 (fig. 1) occurs quite gradually, and there is nothing analogous to the 

 separation which takes place when a cylinder of liquid splits up into 

 drops. We have made a great many experiments on the subject, and 

 have arrived at the conclusion that two conditions are necessary for 

 the splitting up of the rings. 



1. The liquid forming the ring must be of density different from 

 that of the liquid into which it falls. 



2. There must be motion in the liquid into which the drop falls. 

 We suppose that the splitting up takes place in the following way. 



In consequence of the motion in the column of liquid the ring gets a 

 little uneven, more of the liquid collecting in one part of the ring 

 than another. Now, if there is plenty of vortex motion in the ring, 

 this irregularity will not be permanent, as the anchor ring with 

 uniform cross section is the stable form for the motion, so that unless 

 the disturbing force is too great, the ring will oscillate about the 

 anchor-ring shape, and the irregularity will not increase. If, however, 

 the vortex motion in the ring be small, it may not be able to balance 

 the disturbing forces and the irregularities will increase. The dis- 

 turbance is due to the resistance of the liquid in the column, the 

 places where the liquid in the ring has collected falling faster than the 

 remaining portions of the ring ; the ring consequently takes in time 

 some such appearance as that shown in fig. 8 (7 and 8) . The thicker 

 portions will behave now as the drop did when it fell into the column, 

 and they will develop rings in just the same way. 



