194 MESSRS. A. M. WORTHINGTON AND R. S. COLE 



wards discarded for spheres of marble and serpentine, into which a deep hole was 

 drilled, and into this a soft iron plug inserted. After each splash the air had to be 

 re-admitted, the electro-magnet removed, and the vessel opened and the sphere fished 

 out by means of a long, clean, bar-magnet. 



The exhaustion was always pushed to within 2 or 3 millims. of a perfect vacuum, 

 the vapour only of the liquid being left. The first observations were made in broad 

 daylight, with a highly polished nickel sphere (size 1), dusted and undusted, and also 

 with a rough marble sphere, 2 5 '4 millims. in diameter. The observations were 

 alternated by others, in which all conditions were the same except that the air was 

 not removed. In no case could any difference be observed with the naked eye. 

 Thus the polished sphere took down no air when clean, made a large bubble, or some- 

 times a " rough " column, when dusted, and the rough sphere always went in " rough " 

 and made a high column. The depth of fall was 19 '5 centims. 



When spark-illumination was used with the small rough sphere, the figures [6], 

 8, and 10, and the upper part of [7] of the rough-sphere splash of Series X., Plate 2, 

 were obtained repeatedly. 



Even with the Fleuss pump, which works very quickly, it was a work of some 

 minutes to exhaust this large bulb to within 1 or 2 millims. of a vacuum. During 

 this time the electro-magnet had to be kept running, and became hot, and the time of 

 de-magnetisation thus depended, more than was desirable, on the previous history of 

 the magnetisation. 



In order not to waste time and photographic plates, we exchanged this large vessel 

 for a tall " gas jar " of smaller volume, through which the image was, indeed, a good 

 deal distorted, but this does not much diminish the value of the record. We used 

 both water and Alexandra oil as the liquids, and give, on Sheets 7 and 8, a few of the 

 photographs thus obtained. 



Inspection of these photographs shows only two points of difference of importance 

 between the splash in vacuo and the splash in air. 



The first of these is the significant point that it is not so easy to secure a quite 

 smooth splash in vacuo for the reason, as we may confidently surmise, that the liquid, 

 being supersaturated, is liable to burst into ebullition at the surface of the entering 

 solid, where probably, as already explained in the note on p. 196, the velocities set up 

 correspond to a true negative pressure or tension, under which the liquid will readily 

 rupture and break away from the solid surface if any cavity is formed. Indeed, it 

 should be mentioned that until the vessel had been many times exhausted of air, 

 bubbles were very liable to form spontaneously in the liquid and rise from the bottom 

 (boiling by bumping). 



Fig. 1 of Series XXL, Plate 2, shows, for purposes of comparison, a smooth splash 

 in air (height of fall, 14 centims.), while fig. 2, and also [3] and [4] (not here 

 reproduced), show the near approach attained in vacuo, and the development of 

 bubbles. Fig. [5] (not here given) shows a splash in vacuo, in which at an earlier 



