438 



Messrs. G. D. Liveing and J. Dewar. [Nov. 20, 



TO AIRPUMP. 



FROM GASHOLDER 



G for regulating the intake, and the tube h led from the distant end of 

 B to the air-pump. The air-pump was a large one worked by a gas- 

 engine capable of keeping the pressure down to a few millimetres, 

 even with a considerable leakage. Observations were made of the 

 discharge in B at various low pressures, sometimes with, and at other 

 times without, a Ley den jar in circuit. The sparks in A were gener- 

 ally taken with a jar, and there was ample proof, if proof were 

 needed, of the dust derived from the electrodes, since it formed a 

 visible deposit in the tube d, in the first bulb of B, and even on the 

 end E. The air or other gas passed into A was filtered through 

 cotton- wool to remove all dust before admission to the apparatus. 



Various metals were used as electrodes in A, magnesium, iron, 

 manganese, cadmium, fused calcium chloride, metallic sodium in a 

 little glass cup on a platinum wire, and fragments of the Dhurmsala 

 meteorite ; but in no case could the rays of any of the substances 

 employed be seen in the discharge through S, either when a Leyden 

 jar was in circuit or not. 



Incidentally, we found that magnesium electrodes were not so good 

 as some of the other metals for these experiments, because the appa- 

 ratus was never wholly free from traces of air, and lines or bright 

 edges of bands of nitrogen fall very near the most characteristic lines 

 of magnesium, and with small dispersion might easily be mistaken 

 for them. 



Air, hydrogen, carbon dioxide, and oxygen, were successively used 

 as the gases passing through the apparatus, and at various pressures 

 from 2 mm. up to 20, and, in some cases, up to 40 mm., but with the 

 same result; no rays, due to the electrodes in A, could be detected in 

 B. Even when one of the electrodes in A was sodium, and the 

 sodium rays, orange, yellow, citron, green, and blue, were brilliant in 

 the spectrum of A, not even the D lines could be detected in B. We 

 should have expected that some traces of sodium in the state of vapour 

 would have been carried by the stream of hydrogen into B ; but it 

 seems that it was not so ; nor could the apparent absence of rays 



