164 Profs. Liveing and Dewar on the Spectral Lines 



character and was always diffuse at the edges ; so that it was 

 not possible in this way to substantiate the minute difference 

 of wave-length indicated by theory, though the flash-line 

 certainly seemed a little the more refrangible of the two. ■ 



We then tried taking the explosion in a tube bent round so 

 as to be returned npon itself, the two parts of the tube being 

 parallel to each other, and the glass ends side by side (fig. 2). 

 The axis of the collimator (T) was made to coincide with that 

 of one limb of the tube, so that the flash in that limb was 

 seen directly; and by means of two reflecting prisms [r, r) the 

 light from the other limb was thrown into the slit, and the 

 two images were seen simultaneously one above the other. 

 As the gas was ignited from one end of the tube, the flash 

 was seen receding in one limb, approaching in the other, so 

 that the displacement of the two lines would be. doubled. 

 Still we were unable to substantiate any relative displacement 

 of the lines on account of their breadth and diffuse character. 

 By washing out the tube the breadth of the lines was con- 

 siderably reduced, but they remained diffuse at the edges, 

 and baffled any observation sufficiently accurate to establish a 

 displacement. Certainly there appeared to be a very slight 

 displacement, but it was not so definite that one could be sure 

 of it. 



These observations, however, led us to some other interesting 

 results. In the first place, one of the two images of the 

 lithium line almost always was reversed — that is, showed a 

 dark line down the middle. This was the line given by the 

 flash approaching the slit. The receding flash in the other 

 limb of the tube gave as broad a bright line, but it had no 

 dark line in its middle. This observation was made a great 

 many times; and the fact of the reversal established inde- 

 pendently in the case of some other metallic lines by means of 

 photographs. These reversals show that in the wave of 

 explosion the temperature of the gases does not reach its 

 maximum all at once ; but the front of the wave is cooler than 

 the part which follows and absorbs some of its radiation, while 

 the rear of the wave does not produce the same effect. One 

 would suppose that there must be cooler lithium-vapour in the 

 rear of the wave as well as in its front ; but it is possible that 

 the absorption produced by it extends over the whole width of 

 the line, and not only over a narrow strip in the middle. For 

 we observed that when a little lithium carbonate was freshly 

 put into the tube, the red line was so much expanded as to fill 

 the whole field of view — that is to say, it was some ten or 

 twelve times as wide as the distance between the two yellow 

 lines of sodium ; but by washing out the tube with water 



