McCi-ELLAND AND Gii.L — Self ■ Ignition o J Ether- Air Mixtures. 117 



All experiment was made to determine tlie effect of placing a pad of 

 cotton wool at the end of the tube. In the first experiment the junction 

 was 8 cm. from the end of the tube. The pad was about 3 cm. thick, The 

 pressure of air in the tube was 2 cm. The length of the tube was 3 feet, 

 and the diameter -5 inches. 



Without pad, . . . Deflection, 151. 

 With pad, ... „ 110. 



In the second experiment the pad was not so thick, and the junction was 

 placed quite near it. The other conditions were the same. 

 Without pad, . . . Deflection, 155. 

 With pad, ... „ 143. 



In both cases there was apparently a diminution of temperature, which 

 agrees with the result already described as to absence of ignition of an 

 ether-air mixture under the same conditions. 



'J'he part played by the connecting tube was also tested by this method- 

 The tube usually employed was 30 inches long. This was replaced by a 

 tube only 2 inches long. In experiments carried out under similar condi- 

 tions, except for the lengths of the connecting tubes, there was no appre- 

 ciable difference in the deflections due to the thermo-couple. This result 

 was to be expected from the conclusions already arrived at. 



C. — Actual temperature reached hy the air in the above experiments. 



The numbers given above, expressed in terms of the galvanometer 

 deflections, merely indicate the relative values of the temperature. To find 

 the actual temperature reached we employed the following method : — 



A, B (fig. 2) represents a potentiometer through which a current can be 

 sent from the cell E when the key K is closed. G is a galvanometer and J 

 the thermal -junction mounted inside the tube into which the air is allowed 

 to rush. M is an electro-magnet, from which the wires (« and h) are taken 

 to the point where the spring and lever arrangement opens the tap, allowing 

 the air to rush from the reservoir. It is arranged that the lever opens the 

 electro-magnet circuit, and consequently the galvanometer circuit closes at D. 

 The contact C is varied, so that the potentiometer electromotive force balances 

 the electromotive force of the thermal -junction. The potentiometer e.m.f. can 

 be varied, so that the slight sudden kick of the galvanometer is in either 

 direction when the tap is opened. Of course, the junction very soon begins 

 to cool, and then the galvanometer moves in the direction corresponding to 

 the e.m.f. of ,tfie potentiometer. The exact instant of closing the galvano- 

 meter circuit could be varied relatively to the time of opening the tap, and in 



SCIENT. PROC. H.D.S., VOL. XVI., NO. X. P 



