1879.] 



Quantitative Spectroscopic Experiments. 



485 



pelled from the bottle, it almost immediately began to be sucked back, 

 and that nearly, if not quite, as rapidly as it had been expelled, until 

 not only the whole gas expelled was drawn in again, but generally a 

 small quantity of nitrogen in addition. In a few seconds, not only 

 had the sodium vapour ceased to have any sensible tension, but a 

 portion of the nitrogen (2 to 3 cub. centims.) had been absorbed. 

 There was no absorption of nitrogen before the sodium was put in, 

 and when the equilibrium had been attained, which was always in less 

 than a minute, there was no absorption afterwards. 



These phenomena seemed to indicate, (1), either that the tempera- 

 ture of the sodium vapour was higher than that of the iron, or that 

 the emissive power of sodium vapour for light of the wave-length of 

 D was greater than that of iron at the temperature of the furnace ; 

 (2), that the width of the bright D line, or band, increased with the 

 depth of the stratum of sodium when the density was not increased, 

 inasmuch as the pressure to which the sodium vapour was subject was 

 always the same, and the only, or the principal, change during the 

 expansion of the sodium vapour in the bottle was in the thickness of 

 the layer of vapour; (3), that iron either has the power of forming a 

 compound with sodium, which is very stable at the temperature of the 

 furnace, or else it has a power of transmitting sodium vapour analo- 

 gous to that which platinum has in regard to hydrogen. This last 

 supposition seems to be negatived by the fact that we found that out 

 of "220 grm. of sodium put into the bottle in a series of experiments 

 on one occasion, *198 grm. was afterwards extracted when the bottle 

 was cold by washing it out with hydrochloric acid. The effect was 

 not due to the nitrogen, since the sodium appeared to be taken up in 

 the same way when carbonic oxide was substituted for nitrogen. 



This behaviour of iron towards sodium explains a phenomenon 

 which had puzzled us when operating before with iron tubes, viz., 

 that it required a very large quantity of sodium, many times the 

 normal amount, to fill one of the tubes with sodium vapour. It also 

 seems to explain the fact observed by Mr. Lockyer (Proc. Roy. Soc, 

 vol. xxii, p. 371), that sodium heated in a long iron tube gives the 

 D absorption line " no thicker than when seen under similar con- 

 ditions in a test tube." 



We came, of course, to the conclusion that it was useless to pursue 

 the investigation of the quantitative measure of sodium vapour by 

 experiments in iron vessels ; and we had a stout platinum tube made, 

 half an inch in diameter and 23 inches long, connected by a union 

 joint with an iron tube with fittings similar to those of the top of the 

 iron bottles before described ; and also a bottle similar to the iron 

 bottles, but entirely of platinum. When used, the tube was packed 

 with magnesia inside a wider porcelain tube, and the bottle was 

 packed with lime in blacklead crucibles. 



