484 



Profs. Liveing and De war. 



[Dec. 11, 



shown by the dotted lines. The hydrogen was then replaced by 

 nitrogen, the narrow tube removed, c closed by a glass plate through 

 which the spectrum was viewed, and the side tube e connected by a 

 thick rubber tube with a glass tube for collecting the gas driven out 

 by the metallic vapour, or with a manometer formed of an inverted 

 burette dipping into a vessel of water. The sodium, weighed in dry 

 carbonic acid gas, was introduced in little iron cups about the size 

 of gun caps through the oblique side tube d, and the orifice in- 

 stantly closed with a rubber stopper, for the purpose of measuring 

 the gas expelled as the sodium evaporated. This orifice was fitted 

 with a screw cap, for use when the pressure of the gas in the bottle 

 was to be varied. 



Attempts to measure the temperature of our furnace were not very 

 successful. Observations of temperature depending on the electro- 

 motive force of a thermo-electric junction of platinum and palladium, 

 which had been carefully calibrated in boiling mercury, boiling cad- 

 mium, and boiling zinc, failed at the temperature at which our experi- 

 ments were made from the fusion of the palladium. This indicates, 

 from Yiolle's experiments, that the temperature of the furnace was 

 about 1,500° C. 



The bottles, however carefully cleaned, were never quite free from 

 sodium, so that on observing, at first, the spectrum emitted by the hot 

 bottom of the bottle, the D absorption was seen as two fine lines. 



On introducing a fragment of sodium ("015 or -020 grm.) the 

 following phenomena were observed with the spectroscope. First D 

 appeared as a narrow bright band, considerably brighter than the general 

 spectrum of the bottle, with the fine dark pair of lines quite unex- 

 panded in its midst ; this bright band rapidly expanded, becoming 

 fainter and more diffuse as it expanded, was visible to a width equal 

 to fourteen or sixteen times the distance of the D lines from each other, 

 and then died away entirely, leaving the fine dark lines much as they 

 were at first. The addition of several pieces of sodium in succession 

 produced a repetition of the same phenomena, the D absorption lines 

 being but very little wider after the introduction of sodium enough to 

 fill the bottle with its vapour. „ 



Larger pieces of sodium produced an increased width of expansion 

 of the bright band ; but from the very diffuse character of the edges 

 of the band and from its being seen against a background only a little 

 less bright, it was not possible to get any exact measures of the widths 

 to which the bands expanded. 



On observing the gas expelled from the bottle, it was found that 

 the expansion of the bright band was synchronous with the expulsion 

 of gas, and ceased when gas ceased to come out ; that is to say, it was 

 synchronous with the expansion of the sodium vapour within the 

 bottle. It was found, however, that when the gas ceased to be ex- 



