191 -J] 



oti Heat Problems. 



585 



increased by silvering the inner wall to form a I'eliecting sur- 

 face. Experiments made by tilling the annular space with powdered 

 charcoal or silica showed that while with some flasks the loss of 

 heat could be reduced to about one-third, as a rale the results 

 obtained were not better than with the use of the silvered vessels. 

 This may be shown by the following experiment (Fig. G). Three 

 vacuum vessels, A, B, C, containing liquid air, are sealed to a common 



Fig. 6. 



exhaust. B is of the ordinary unsilvered variety ; A and C con- 

 tain powdered charcoal and silica respectively in the annular space. 

 The three tubes are fitted with glass delivery-tubes leading through 

 rubber corks to a glass tank I) filled with coloured water, con- 

 taining three uniform collecting-tubes, E, F, G. The evaporated 

 air from A, B, C is delivered simultaneously into these collecting- 

 tubes. In al)out five seconds F, connected with the plain tube B, 



