244 On Spectral Lines of Low Temperature. 



simpler one, at least in some cases, may be offered. Every pho- 

 tographer must have observed the singular power which glass, 

 exposed to an inhabited atmosphere, possesses of attracting not 

 only moisture, but also thin films of grease. If a photographic 

 glass plate be thoroughly cleaned and put aside for some days, it 

 can very seldom be used at the end of that time without further 

 treatment. Unless it be cleaned afresh with some solvent of 

 grease, such as ether or strong acid, or rubbed with tripoli, it will 

 probably show on the finished picture the streaks which betoken 

 an imperfectly cleaned plate. What is true of a glass plate is 

 probably true of the inner surface of a tube. Unless it is cleaned 

 with something more powerful than water, it is pretty certain to 

 retain a film of grease. I interpret these carbon-lines, therefore, 

 as a simple proof of slovenly preparation. Their presence or 

 absence may serve as an easy test, not of the skill, but of the care 

 with which the thermometer has been made. I find that it de- 

 pends more upon the reputation of the maker than on the price 

 charged for the instrument. One of Elliott's that showed only 

 mercury-lines is a common two-shilling instrument. In one 

 costly, but apparently very dirty, therm ometer— besides the four 

 carbon-lines, one or two other very faint green lines showed 

 themselves in the part of the spectrum near the solar line E ; 

 but they were too faint to be identified. In this thermometer, 

 after frequent exposure to the action of the coil, a slight grey 

 deposit was observable in the tube just over the point where the 

 mercury ordinarily stood. 



But if these carbon-lines show grease, why does the hydrogen 

 present in grease not show its characteristic lines, especially the 

 line F, which is the first seen of the hydrogen-lines ? I believe 

 the reason to be that hydrogen will not become luminous under 

 electric influence at a low temperature, while the vapour of (com- 

 bined) carbon will. This may be shown in a very simple way. 

 Place a Geissler's tube, exhausted from some hydrocarbon com- 

 pound (paraffin oil answers well), upright on the insulated plate 

 connected with one secondary pole. Let a wire in connexion with 

 the other pole be so arranged that it can be at will applied to or 

 withdrawn from the upper end of the Geissler's tube while the coil 

 is at work. First apply it, and let the full discharge pass. The line 

 F and the carbon-line I have named// will be seen close together 

 with a deep-black space between them, giving the effect (in a 

 one-prism instrument) of a groove with luminous edges. With- 

 draw the movable electrode; F will immediately disappear; but 

 the carbon line //will remain, little diminished in brightness. 

 Approach the movable electrode to the top of the tube until a 

 spark occasionally passes. Whenever the spark passes, the line 

 F will shoot across the spectrum ; but as long as there is no 



