02 



Some preliminary determinations carried out by Griffiths proved 

 that tlie densities of the two modifications are different (at the same 

 temperature) and that this difference is of the order of 1 : 7000. 

 The modification formed by quenching cet. par. has the greater 

 specilic vohime. 



The investigations to be described below were carried out in order 

 to fix the limits of stability of both modifications, viz. to investigate 

 whether sodium is enantiotropic ') or monotropic. 



2. Although the change of volume which accompanies the trans- 

 formation of o-sodium into /^-sodium is small according to Griffiths' 

 measurements, yet the use of the dilatometer is suitable, if certain 

 precautions are taken, which enable us to carry out exact measure- 

 ments with this instrument. 



These precautions are : 



a. A large dilatometer must be used ; we employed an instrument 

 of about 380 cc. 



h. The bore of the capillary tube must be small (Bore of our 

 tube 1.2 mm.). 



c. The quantity of liquid put in, (rock oil) must be as small as 

 possible. 



(/. The temperature at which the readings are taken must be 

 constant witiiin some thousandths of a degree. 



3. Special care has to be taken in tilling the dilatometer. This 

 operation was carried out in the following manner : The metal 

 (sodium in rods from KAHLBAUM-Berlin, comp. § 6 and 8) was melted 

 under petroleum in a Jena-glass flask. The rock oil had been 

 prepared in the way to be described in ^ 4, while the metal had 

 been treated in a special manner (comp. § 5). The flask (Fig. 1) is put 

 into an oil bath RR, heated to J 30°. Pieces of sodium are added 

 until there are about 400 cc. of molten metal in 0.^) 



The dilatometer after being filled with rock oil, is placed in the 

 same bath. and G are connected by means of a glass tube PLH 

 the end of which (in G) is drawn out. 



The neck of the dilatometer is connected with a tube L by 

 means of rubber tubing K. 



1) In tlie paper mentioned above [These Proc. 23, 896 (1915)] the opinion was 

 expressed that there exists a transition point between 0^ and 90°. 



-) Generally there are formed spheres of metal which do not coalesce unless 

 the molten metal is cooled below the melting point and gently stirred at the 

 same time. 



