95 



of which was kept constant within some thousandths of a degree. 

 Its temperature was 97°. 10. 



The tube and its contents is heated to 99° and put into the 

 tliermostat. When the temperature of the metal iiad become 97°. 10 

 the stirrer was put in motion. The temperature rose to 97°. 51 and 

 remained constant for some time. The experiment was repeated, 

 tlie molten metal being cooled 0°.7 lower than before. Aflei' slirring 

 the temperature rose again to 97°. 51. 



The beginning of solidification consequently occurs at 97°. 51, 

 while the range of melting covers 0°.3 C. 



9. As the changes of volume which play a role in these inves- 

 tigations are only small, and as the glass of the dilatometer was 

 exposed in our experiments to sudden and strong changes of temper- 

 ature (about 100°) we thought it important to prove that the glass 

 used did not show thermal hysteresis. For tliis purpose we filled 

 our dilatometer with rock oil and heated it for some hours in a 

 thermostat at 95°. 00. After having noted the place of the meniscus 

 (358.0) we dipped the instrument into petroleum which had been 

 cooled (by means of solid carbon dioxyde) to — 20° C. After half 

 an hour we put it again into the thermostat at 95°. 00. After two 

 hours the level of oil was again 358.0. Twenty-four hours later it 

 had not changed : consequently thermal hysteresld had not occurred. 



10. The dilatometer was now filled in the way described above 

 (comp. § 3) with molten sodium and some cc. of petroleum. After 

 this it was cooled very slowly in the thermostat to 15° C, so that 

 the metal might be (ransformed into the «-modification. 



The following results were obtained (Comp. Table I). 



TABLE I. 



Temperature 



50° 

 68.5 

 90.0 

 96.0 



At 97°. 3 the melting of the metal occurred which is accompanied 

 bv a verv mai'ked increase in volume. 



