ATTAINMENT OF VERY LOW TEMPERATURES. 



13 



It will be seen that as soon as the circulation of the air through 

 the apparatus was suspended the liquid air in it began to evaporate 

 rapidly, heat being conveyed to it by conduction from the outside. 

 However, long before the whole of the liquid air in the apparatus had 

 evaporated the temperature of the coils had become steady and the 

 quantity of liquid air thus evaporating per minute could be taken as a 

 measure of the quantity of heat entering the apparatus during the 

 process of liquefying air. This quantity corresponds to about 2.6 

 liters of gaseous air per minute, or to a loss of 200 grams of liquid air 

 per hour. 



During an actual experiment it was the duty of one assistant to 

 maintain a steady current of air, escaping at an almost constant pressure, 

 through the apparatus. The pressure gauges were read every minute ; 

 the gas meter readings were taken at the end of every half hour, the 

 liquid air was drawn off every five minutes, and the vacuum vessels 

 removed and weighed every half hour ; the temperature at which the 

 air entered and left the apparatus and the barometer readings were 

 taken at intervals during an experiment. The following are the re- 

 sults of some observations : 



December 14, 1902. Machine started at 10:20 A. m. 



Duration of experiment 



Weight of liquid air collected 



Loss by evaporation 



Loss due to heat absorption by liquefier 



Weight of air passing through gas meter 



Total weight of air passing through ap- 

 paratus 



Theoretical quantity of liquid air per 

 hundred of gaseous air 



Quantity collected per hundred of gase- 

 ous air 



Efficiency of liquefier 



Mean pressure 



Temperature interchange 



10:55 to 11:25 



460 gms. 



40 gms. 



TOO gms. 



6,620 gms. 



7,120 gms. 



8.4 



6.4 



76.2 

 167 atm. 

 0.4° C. 



11:25 to 11:55 



490 gms. 

 35 gms. 



100 gms. 

 6,525 gms. 

 7,050 gms. 



8.9 



69 



77-5 

 165 atm. 

 0.4° C. 



The results of other sets of observations led to practically the same 

 result. The errors of the various observations probably reduce the 



