8y8 



voiiiids it even all over. Of course liquid lijilrogeii can he iutrodu- 

 ced into the cryostai in the ordinary way, to immerse the experi- 

 mental tube in a bath of liquid hydrogen. 



We had a good deal of difficulty in })rocuring a rod of solid 

 oxygen of about 5 or 6 cm. length that was completehj homogeneous. 

 As in freezing the volnme of the oxygen diminishes by about ^j^ 

 there is a great tendency to form hollows. A homogeneous rod can 

 onh' be obtained by allowing the liquid in the cylindrical mould lo 

 gradnall}' freeze from t/ie bottom upwards. As soon as solid matter 

 settles in the neck of the tube, hollows must arise in the experimen- 

 tal mass, as the entrance of liquid oxygen is cut off ^). The diffi- 

 cnlty of procuring an homogeneous cylinder was increased by the 

 fact that in our experiments the freezing had to take place inside 

 a silvered vacuum glass, and could therefore not be followed by 

 the eye. 



In order to come a step nearer to the solntion of this problem, 

 which still presents difficulties, we made nse of the possibility of 

 regulating at pleasnre the heal conducting power of a double-walled 

 vacuum glass. The experimental tube in which the oxygen was frozen 

 was made double-walled (not silvered) ; as much hydrogen was put 

 in between the walls that the (small) pressure had exactly such a value, 

 that according to preliminary expei'iments in a transparent bath 

 the freezing took place under the most favourable circumstances. 

 It appeared to be favourable to this that in the upper part of the 

 tube latent heat was developed by the condensation of the gaseous 

 oxygen which was admitted during the freezing. When a once for- 

 med stick partially melted, it was difficult to get it back into a 

 dense condition. The double wall of the experimental tube is there- 

 fore also of use to eliminate temporary irregularities in the temperature 

 of the surrounding bath (vapour or liquid). It was found that the 

 temperature-range of the ti-ansparent modification did not extend 

 more than 5 or 6 degrees below the melting point. 



§ 3. Residt.s. The figures in the following table are averages. To 

 be able to express the results in absolute measure, the ftelds are 

 measured in absolute measure, and the dimensions of the tubes 

 carefully determined. The .tubes were of the sj^m metrical type (see 

 Comm. N". 139r? § 4), the lower part was evacuated, so that no 



-) We sliJill return later to the freezing of oxygen and the change from the 

 transparent to the opaque mass. When the acces.'; of gaseous oxygen is cut ofif, 

 a long shaped hollow usually forms which as it increases assumes the shape of 

 a worm fantastically coiled up in a confined space. 



