252 Professor Sir James Dewar [Jan. 23, 



This arrangement thus charged only needs a little li(|nid air 

 sucked in every one and a-half hours. The liquid hydrogen vessel 

 will not need replenishing for at least four hours. The level of the 

 liquid hydrogen in the calorimeter does not fall 1 cm. in six hours 

 with constant use. The bulk of the materials added roughly com- 

 pensates for the volume of the liquid hydrogen evaporated. It is 

 important that this level should not materially change, since, after 

 striking the shoulder, bodies move more slowly, are deflected on to 

 the cold wall, and low results are obtained due to the longer cooling 

 of the materials in the vapour before being immersed in the liquid 

 hydrogen. 



The isolation of the calorimeter was such that less than 10 c.c. 

 of hydrogen gas evaporated from it per minute. The whole vacuum 

 vessel combination is supported between cork-lined spring jaws mounted 

 on a heavy metal base, on which the outer vacuum vessel rests. 



The cooling vessel H is connected by india-rubber tube to the 

 top of the calorimeter. It consists of an ordinary cylindrical slit 

 silvered vacuum vessel, 30 cm. long and 7 cm. wide, with a central 

 axial open tube K sealed in below. This tube passes through the 

 liquid in the vacuum vessel. It has the same diameter below as the 

 neck tube of the calorimeter. Near the top of the central tube a 

 side tube J, of about the same diameter, and some 3 cm. long, serves 

 for the introduction of the weighed pieces of material, which are all 

 cooled previously to the temperature of liquid air, and then fall on 

 to a thin metal pan P fitting loosely the tube E, where they remain 

 about fifteen minutes. P is supported by being hinged to two thin 

 ebonite rods, L and M, fixed to a brass fitting cemented on to the 

 top. The rod L is not fixed directly to the disc but to a metal ring 

 R. Prom the ring R two thin vertical steel wires are connected 

 freely to two points on the circumference of the pan below. This rod 

 and the attached ring can be given a vertical motion by a crank N 

 in the top fitting, thereby tipping the pan and releasing the piece of 

 material, which then falls freely down into the calorimeter. The 

 level of the pan is approximately one-third the vertical height of the 

 cooling vessel. Quartz was found to be safer than glass for the 

 construction of this vessel. A high vacuum was maintained by filling 

 with charcoal a cross-tube S, opening to the annular space. 



At the temperature of boiling nitrogen, the convection currents 

 in the central tube of such a vessel, when connected to the calori- 

 meter below, have no serious effect on the temperature in the tube 

 at a reasonable distance from the bottom, provided the central tube 

 be not wide. The difference of temperature in the tube and in the 

 surrounding liquid was found to be only 0'5° when the tube was 

 1 • 5 cm. wide. With a larger pattern vessel the width of the central 

 tube was increased to 2*2 cm., and even here the difference was 

 under 3° at the level of the pan. These temperatures were measured 

 bv a small helium thermometer, consistinsr of a -4 c.c. bulb to which 



