ON IMPACT WITH A LIQUID SURFACE. 



I'..:; 



making the ribs more noticeable, and we therefore conclude that withiu these limits 

 the viscosity does not play an important part in determining the main course of the 

 phenomenon.] 



Experiments in vacuo. 



1 1 remained to examine what part was played by the air in the whole transaction. 

 Tli is could only be settled by removing the air. We accordingly made provision for 

 obtaining, by instantaneous illumination, observations of splashes in vacuo. The 

 method was simple enough, since, happily, very exact timing was not necessary. For 

 eye observations a large, strong "bolt-head" was employed, and a 1-inch thick 

 slab of india-rubber closed it air-tight (see fig. 4). This slab was pierced at one side 

 by a glass tube leading to a Fleuss pump, and centrally by a short thick conical piece 

 of soft iron, which served as the prolongation of the core of a straight electro-magnet 

 which could be Imd on the top. A pad of folded, fine woven, copper wire-gauze* 

 prevented the bottom of the vessel being broken by the impact. 



Fig. 4. 



The nickelled and polished steel spheres were at first employed, but these retained 

 so much of their magnetism that the timing was very uncertain, and they were after- 



* We have found a pad of this material very convenient and efficient in all our experiments. 

 VOL. CXCIV. A. 2 C 



