330 



NA TURE 



February 6, 1896 



immersing the side arms attached to the flasks in liquid air boil- 

 ing under a low pressure, and subsequently hermetically sealing 

 off the side arms containing the solid, the residual air left in the 

 flasks may be preserved for analysis ; it is then ascertained that 

 the residual air still contains oxygen and nitrogen in the usual 

 proportion. In the earlier experiments, the argon solidified 

 before the nitrogen, but chemical nitrogen and air nitrogen with 



Fig. I. — A, air or oxygen inlet. B, carbon dioxide valve. D, regenerator 

 coils. F, air or oxygen expansion valve, g, vacuum vessel with liquid 

 oxygen. H, carbon dioxide and air outlet. O, air coil. • carbon dioxide 

 coil. 



its I "I per cent, of argon behaved in substantially the same way 

 on liquefaction. 



Olszewski has examined liquid nitric oxide, and describes it as 

 colourless, but on strongly cooling several carefully purified 



NO. 137 1, VOL. 53] 



samples of this gas, which had been prepared by different 

 methods, the author obtained in each case a nearly white solid 

 melting to a blue liquid. The colour is more marked at the 

 melting than at the boiling point, and the liquid is not magnetic. 

 Solid nitric oxide is not phosphorescent, nor does it show any 

 chemical action in liquid oxygen, provided the tube containing it 

 is completely immersed ; but if the tube full of liquid oxygen be 

 lifted into the air, a violent explosion almost instantly occurs. 



In a good vacuum vessel, specific gravities may be taken in 

 liquid oxygen as easily as in water. Some twenty substances 

 were weighed in liquid oxygen, and the apparent density of the 

 liquid calculated ; the results were then corrected, using Fizeau's 

 values for the variation of the coefficients of expansion of the 

 solids employed, and the real density of liquid oxygen was thus 

 calculated as I •1375 under a pressure of 766-5 m.m. The 

 variation of density is about ± o*cx3i2 for 20 m.m. pressure. 

 Wroblewski found the density of liquid oxygen at the boiling point 

 to be I -168, whilst Olszewski found 1-124. Fizeau's parabolic 

 law for the variation of the coefficient of expansion holds down 

 to - 183° ; the solid which contracted least during cooling was 

 a compressed cylinder of silver iodide, that which contracted 

 most a block of compressed iodine. 



Similarly the density of liquid air was found to be 0-910 

 and that of nitrogen at its boiling point 0-850. No great 

 accuracy attaches to the density of liquid air as thus determined, 

 for liquid air kept in a silvered vacuum vessel rises 1° in boiling 

 point every ten minutes during the first hour ; the density of 



r^ 



\^- 



Fig. 3. 



Fig. 4. 



liquid air, however, does not reach that of pure oxygen even 

 after thirty hours' storage. 



A small ignited jet of hydrogen burns continuously below the 

 surface of liquid oxygen, all the water produced being carried 

 away as snow. There is a considerable amount of ozone formed, 

 which concentrates as the liquid oxygen evaporates. In the 

 same way graphite, or diamond, when properly ignited, burns 

 continuously on the surface of liquid oxygen, producing solid 

 carbonic acid and generating ozone. If liquid oxygen is absorbed 

 in wood charcoal, or cotton wool, and a part of the body heated 

 to redness, combustion can start with explosive violence. 



The experiments of Joule and Thomson and Regnault on the 

 temperature of gas jets issuing under low pressures are well 

 known ; the following observations refer to the pressure required 

 to produce a lowering of temperature sufficient to yield liquid 

 in the gas jet. The apparatus used in the study of highly com- 

 pressed gas jets is sketched in Fig. 2 ; c is a vacuum tube 

 holding a coil of pipe about 5 mm. in diameter along with 

 carbon dioxide or liquid air for cooling the gas before expansion, 

 and A is a small hole in the silver or copper tube about i mm. in 

 diameter, which takes the place of a .stopcock. When carbon 

 dioxide gas, at a pressure of 30 or40atmos., is expanded through 

 such an aperture, liquid can be seen where the jet impinges on 

 the wall of the vacuum tube along with a considerable amount of 

 solid. If oxygen gas escapes from the small hole at the pressure 

 of 100 atmos., having been cooled previously to - 79° in the 

 vessel c, a liquid jet is just visible. It is interesting to note in 

 passing that'Pictet could get no liquid oxygen just below 270 

 atmos., owing to his stopcock being massive and outside the re- 



