THE SPECTRUM 



309 



A 



into which slides a metal tube, enlarging at the front end into 

 a long cone r, and at the back end bearing a plate with a slit 

 s. The cone r is made to extend nearly to the lens which 

 focusses the slit ; the latter being made thus, to slide out, 

 because it will need cleaning pretty frequently. To the door 

 at the back is attached, by a projecting piece, a metal ring, in 

 which can be supported capsules, b, for the burning sub- 

 stances. 



In using this combustion-lantern, as I venture to term it, 

 it is arranged in front of 

 the ordinary lantern, in the 

 optic axis, with the door 

 open as in the second 

 figure, so that the light 

 from the optical lantern 

 can be passed through the 

 slit s, and enable the latter 

 to be focussed on the 

 screen, with its spectrum 

 all in focus as described 

 in 175. The optical 

 lantern may then be turned 

 off or removed. The pow- 

 der used is then placed 

 in the capsule 6, in a little 

 heap, a piece of cotton wick 

 about an inch long is stuck 

 perpendicularly into it so 

 as to be half-buried in the 

 powder ; the wick is lighted, and the door shut ; Prof. Wein- 

 hold advises soaking the wick in lead chromate. As soon as 

 the flame reaches the powder it flames up, and gives excellent 

 line-spectra, though for a short time only. 



Prof. Weinhold recommends the following mixtures as 

 effective. For sodium lines : 3 parts sodium nitrate, 1 part 



FIG. 169 



