1879.] 



Quantitative Spectroscopic Experiments. 



483 



the lines and the thickness, density, and temperature of the A r apours 

 conld be determined. TVe had snch a determination in view during 

 onr experiments on the reversal of the lines of metallic vapours heated 

 in iron tabes, and have since made a good many experiments expressly 

 for the purpose. We have not yet overcome all the difficulties attend- 

 ing these experiments ; nevertheless, some results, definite and in- 

 teresting, have been reached, which we now communicate to the 

 Society. 



The first experiments were made with wrought iron bottles, of the 

 form shown in the accompanying figure. To prevent the diffusion of 

 carbonic oxide from the fire we found it essential to have the wide 



part, a, and part of the tube, b, coated with borax. The bottle was 

 then packed in sand in two crucibles, one of them inverted, and heated 

 in a small furnace fed with gas retort carbon. The bottle was previ- 

 ously treated with fused caustic potash at a red heat, and washed with 

 acid. Any oxide remaining in the bottle was reduced by a current of 

 hydrogen introduced by the side tube e, and forced to pass to the 

 bottom by allowing it no exit except by a narrow tube passed through 

 the opening c and reaching nearly to the bottom of the bottle, as 



