Mr. J. B. Hannay's Examination of Vacua. 233 



was made with this method ; but a new difficulty arose. The 

 fusible metal expands very much after solidification, and often 

 bursts the tube containing it ; in fact it always bursts it if 

 allowed to cool far enough ; but, as its principal expansion- 

 point is 35°, or 60° below its freezing-point, it was hoped 

 that by keeping the temperature between 50° and 70° a good 

 stoppage would be obtained. I may mention that the expan- 

 sion of the metal at 35° is so great that, if allowed to solidify 

 in a common deep goblet, it invariably splits it ; but it may 

 be allowed to cool and solidify in a shallow stew-pan with 

 safety. After working for some weeks in vain attempts to 

 regulate the temperature of the stoppers, the difficulty was 

 found to be so great that (some point being momentarily 

 overlooked) the apparatus was invariably broken. When it 

 is remembered that working at such high temperatures the 

 glass often cracks, being extremely brittle and having little 

 tenacity, it will be understood that failures Avere often due 

 to this cause ; but when something had gone wrong, and 

 I could not get the metal out without cooling the apparatus, I 

 have often had to stand by and see two days' glass-blowing 

 slowly disintegrated without being able to save a tube. 



It became manifest that frozen stoppers would not do, at 

 least where many were required, and stopcocks were not 

 reliable when in communication with the vacuum; so a new 

 apparatus was constructed, where the stopcocks were under 

 pressure, and therefore not liable to leakage, and where leakage 

 could only take place to the unimportant side of the appa- 

 ratus. The pumping action was also abandoned, and the 

 apparatus shown at fig. VI. used. The tube A was attached 

 to a mercury-pump, and the cocks «, b, and c closed. When 

 evacuating, b was slightly opened, and dipped into a pot of 

 metal so that it filled up to the inside of the stopcock, 

 which was then closed. When a good vacuum had been 

 obtained, the tubes B and C were heated and allowed to be 

 sucked in till nearly closed. The same was done with the 

 tubes above and below the bulbs E ; and they were also heated 

 nearly to the softening-point. B was then sealed off, and the 

 metal poured through F into D till nearly full, and then 

 also sealed off from the Sprengel pump. Some metal was 

 left in F, and new burnt india-rubber applied to a. 



The bulbs E were then strongly heated and c opened, so 

 that the metal rose and swept the residual air into H, where it 

 was frozen by cold air. No leakage can take place by b or c, 

 as they are under more than an atmosphere of pressure, while 

 leakage by a is of little consequence, but seldom occurred. The 

 tube at the bottom of D was frozen by the same blast as was 



