the Radiation of Heated Gases. 469 



circulate in the apparatus, no sodium being admitted. A 

 distinct and fine sodium line was visible in the spectroscope, 

 which increased in brightness when a little air was mixed 

 with the gas, but which gradually faded to invisibility when 

 the phosphorus tube was heated so as to eliminate oxygen. 



The explanation of this result appears to be simple enough. 

 The trace of oxygen remaining in the coal-gas combines with 

 the hydrogen when it reaches the hot part of the tube, and 

 the " flame " so formed (which, however, is not visible as 

 such, except when a large quantity of is present) becomes 

 tinted by the salts of sodium, which in excessively minute 

 quantity are known to be driven off from the porcelain at a 

 red heat, just in the same way as the Bunsen flame outside is 

 tinted. This fine double D line, therefore, may not be the 

 result of heat alone, since it is developed as a consequence of 

 chemical reactions. 



II. A pellet of clean sodium was placed in the steel spoon, 

 and the gas — coal-gas — allowed to circulate, the phosphorus 

 being heated. The line seen in Experiment I. was watched, 

 and some time after it had quite disappeared the current was 

 stopped and the sodium pushed into the hot part of the tube. 

 Instantly the central bore of the porcelain was filled with 

 light, which in the spectroscope was found to be perfectly 

 continuous, but crossed by a very wide black line at D. 

 Gradually the continuous spectrum faded, and as it became 

 fainter the dark D line was seen to be bordered with a fringe 

 of light on each side ; and as the vapour became less dense, 

 owing to the distillation of the sodium into cooler parts of the 

 tube, the D line went through the changes represented in 

 fig. 3, in the order a, b, c, d, finally persisting as a rather 

 wide bright line in which a very fine dark line could 

 usually be made out *. But at any stage of the experiment, 

 the dark central line could easily be extinguished by allowing 

 a gentle current of gas to push back the cooler absorbing- 

 layer into the hotter regions. Now the question to be decided 

 was whether this broad bright, hazy D line was or was not 

 the result of chemical activity. 



III. In this experiment the phosphorus tube and the 

 drying-tube were cut out of the circuit, the gas-holder A 

 being connected directly with the heating- tube. With the 

 current of gas stopped, the D line appeared as in the last 

 experiment, but observations were somewhat impeded by 

 opaque clouds of oxide which hung about the cooler parts 



* The dispersion of the spectroscope employed in all the sodium 

 experiments being* insufficient to separate the two components of tha 

 D line, it is evident that when this line appeared widened the two 

 were really fused into one broad band. 



Phil. Mag. S, 5. Vol. 39. Nq. 240. May 1895. 2 I 



