312 



OPTICAL PROJECTION 



the screen shows the radiation n, and the other half the 

 absorption A, from the same sodium flame. 



But the simplest, easiest, and best method is to produce 

 the vapour between the carbon poles themselves. To do this 

 with certainty, the positive carbon must be considerably 

 larger than usual, and this is another reason for employing a 

 hand-regulating lamp for spectrum experiments ; besides 

 which, the distances have to be adjusted by trial, which is 

 more conveniently done by hand. I doubt if better and more 

 certain reversals have ever been demonstrated than by my 

 long-time friend and correspondent in these matters, the Kev. 

 P. B. Sleeman, of Bristol ; and for sodium the arrangement 



adopted by him, as shown 

 at A, fig. 172, is probably 

 the best possible. The 

 figure is actual size, the 

 lower carbon being |-inch 

 diameter, with a groove 

 ^-inch wide cut in the end, 

 round a centre of ^-inch 

 diameter. In this groove 

 small pieces of sodium are 



placed (not too much). As soon as the upper carbon touches 

 the centre, the metal is volatilised. It will be seen at once 

 how and why, by this arrangement, as soon as the arc is 

 struck, the vapour is comparatively cool, and well outside 

 of the incandescent carbon. A very little adjustment of the 

 carbons brings out the reversal conspicuously. The ordinary 

 1 cup,' as in B of the same figure, will however answer very 

 well, and for more refractory metals like lithium is preferable, 

 the heat of the arc itself being needed to volatilise them. 



The guiding principle, in reversal experiments, is to get 

 a short arc. If we use carbons as in B, and separate the 

 carbons to give an arc of good length, we get the bright lines 

 and comparatively dull carbon poles ; then, as we lower the 



A B 



FIG. 172. Carbons for Reversals 



