McCliolland and Gill — Self -Ignition of Ether- Air ^fixtures. 113 



temperatures developed at the various positions indicated. In each series 

 the numbers arrived at are the means of several observations. The agree- 

 ment between successive observations tal^en under apparently identical 

 conditions was not very exact. Tlie tap may not have been always opened 

 in exactly the same manner. The strong rush of air down tlie wide tube 

 may also liave affected the delicate thermo-j unction. It was impossible to 

 obviate this possibility. i 



Table I. 



Eelation between the galvanometer dellection and the length of the 



Thus the temperature produced increased in tubes up to 6 feet long, but 

 had decreased for a length of 11 feet. It was not convenient to use tubes 

 between 6 and 11 feet long. 



2 4 6 8 10 II 



Feet. 



Length of Tube. 



Curve 1. — Showing relation between deflections and length of explosion-tube. 



