362 



Prof. J. Dewar. 



point of nitrogen ; the other on a simple arrangement for keeping the 

 more volatile gases from getting into solution after separation by 

 partial exhaustion. By the latter mode of working something like 

 1/34 000th of the volume of the air liquefied appears as uncondensed 

 gas. The latter method is only a qualitative one for the recognition 

 and separation of a part of the hydrogen in air. In a former paper on 

 the " Liquefaction of Air and the Detection of Impurities,"* it was 

 shown that 100 c.c. of liquid air could dissolve 20 c.c. of hydrogen at 

 the same temperature. The crude gas separated from air by the 

 second method gave on analysis — hydrogen 32*5 per cent., nitrogen 

 8 per cent., helium, neon, &c, 60 per cent. After removing the 

 hydrogen and nitrogen the neon can be solidified by cooling in liquid 

 hydrogen and the more volatile portions separated. 



There exists in air a gaseous material that may be separated without 

 the liquefaction of the air. For this purpose air has to be sucked 

 through a spiral tube filled with glass wool immersed in liquid air. 

 After a considerable quantity of air has been passed, the spiral is 

 exhausted at the low temperature of the liquid air bath. The spiral 

 tube is now removed and allowed to heat up to the ordinary tempera- 

 ture, and the condensed gas taken out by the pump. After purifica- 

 tion by spectroscopic fractionation, the gas filled into vacuum tubes 

 gives the chief lines of xenon. The spectroscopic examination of the 

 material will be dealt with in a separate paper by Professor Liveing 

 and myself. A similar experiment made with liquid air kept under 

 exhaustion, the air current allowed to circulate being, to prevent lique- 

 faction, under a pressure less than the saturation pressure of the liquid, 

 resulted in crypton being deposited along with the xenon. 



A study of fifteen electric resistance thermometers as far as the 

 boiling point of hydrogen has been made, and the results reduced by 

 the Callendar and Dickson methods. The following table gives the 

 results for seven thermometers, viz., two of platinum, one of gold, 

 silver, copper, and iron, and one of platinum-rhodium alloy. It will 

 be noted that the lowest boiling point for hydrogen was given by the 

 gold thermometer. Next to it came one of the platinum thermo- 

 meters, and then silver, while copper and the iron differ from the gold 

 value by* 26 and 32 degrees respectively. The gold thermometer 

 would make the boiling point 23° - 5 instead of the 20° -5 given by the 

 gas thermometer. Then the reduction of temperature under exhaus- 

 tion amounts to only 1° instead of 4° as given by the gas thermometer. 

 The extraordinary reduction in resistance of some of the metals at the 

 boiling point of hydrogen is very remarkable. Thus copper has only 

 l/105th, gold l/30th, platinum l/35th to 1/1 7th, silver l/24th the 

 resistance at melting ice, whereas iron is only reduced to l/8th part of 

 the same initial resistance. The real law correlating electric resistance 

 * ' Chem. Soc. Proc.,' 1897. 



