256 ANNUAL. KEPORT SMITHSONIAN INSTITUTION, 19 3 5 



always big compared with that of the gas. Starting at low tempera- 

 tures, however, the situation changes absolutely. Firstly, one gets a 

 much bigger amount of gas into the container at a given external 

 pressure, according to the gas laws, and secondly, the specific heats 

 of all solid bodies drop with temperature, disappearing on approach- 

 ing absolute zero. So for instance, at a temperature of about 12°, 

 1 cubic centimeter of helium gas compressed to 100 atmospheres 

 has the same heat capacity as 1 kilogram of copper. This means 

 we can neglect the heat capacity of the walls altogether, and we 

 have the advantage of working with mathematical walls. Thus, the 

 efficiency of the procedure, as I described it to you, becomes very 

 high, and working with suitable dimensions and a good isolation, 

 it is easy to keep the low temperatures, too. For instance, under 

 the conditions realized in this apparatus, about 60 percent of the 

 volume originally filled with the compressed helium remains filled 

 with the liquid phase, and in the apparatus we generally use we 

 can raise this efficiency still much higher. 



In this way one could liquefy any quantity of helium. But as 

 the specific heats at low temperatures are so very small, only tiny 

 amounts of liquid helium are really necessary for cooling down the 

 apparatus and making measurements for a number of hours, if the 

 apparatus is designed in a suitable way. For example, in this ap- 

 paratus we have liquefied about 50 cubic centimeters, which is suffi- 

 cient to cool down the whole system and to work for about 5 hours. 



In figure 4 you see a rough plan of the apparatus. Outside is the 

 vessel D with liquid hydrogen; inside, the space S that is evacu- 

 ated. Within this you see the container C that is first filled with 

 compressed helium, and after the expansion with liquid helium. 

 Attached to this container is a gas thermometer G, the readings of 

 which you saw before. 



To cool the apparatus down further, one could reduce the pressure 

 over the liquid helium; the temperature must then fall till it cor- 

 responds with the vapor pressure of the helium. For purely technical 

 reasons we do not do this, but we have a second vessel E that we can 

 fill with liquid helium by letting helium gas through the tube T. 

 Then it condenses in the tube T where it is in contact with C, and 

 drops down into the vessel. By pumping through T now, we reduce 

 the vapor pressure, and therefore the temperature falls. This vessel 

 and the surrounding Dewar vessel are of this peculiar shape because 

 we will afterwards apply magnetic fields to some substance situated 

 in the vessel, and, of course, one can generate strong magnetic fields 

 over a small distance more easily than over a big one. Now, we will 

 pump off the helium in this vessel, in order to reduce the temperature 

 here, and the temperature will fall below 2° within a short time. 



