120 
Proceedings of the Royal Society of Edinburgh. [Sess. 
the two electrodes of the lamp, thus forming a continuous electrical con- 
nection through the flexible conductor L, the heating coil I, the sleeve K, 
the clip N (figs. 1 and 3), the cap F (figs. 1 and 3), or the bayonet-holder P 
(figs. 4, 6, and 7), and thence through the leading-in wire E and the 
mercury of the lamp to the other electrode. When the current is switched 
on, the heating coil of platinum wire I becomes incandescent; the first 
bubble of mercury formed in the vessel A rises to the constriction C, where 
it is caught, thus breaking the continuous column of mercury in the lamp, 
and starting the arc. 
The lamp, as we have described it, may be made of Jena glass, uviol 
glass, or quartz, according to the purpose for which it is required. For 
example, when the lamp is to he made use of in work on the visible spectrum, 
it may very conveniently be made of Jena-glass combustion tubing. Our 
first experimental lamps were made of soda glass, because of the great ease 
with which it can be worked. The method of starting the lamps proved to 
be entirely satisfactory, the arc being struck about 15 seconds after the 
current was switched on.* Of course, if the lamp was allowed to burn, it 
soon collapsed, owing to the fact that soda glass cannot withstand the 
temperature of the incandescent mercury vapour. The lamps, however, 
served the purpose of testing the method of starting. Once the method 
was fully developed, our attention was turned to the making of lamps from 
Jena glass. Here we met with a considerable amount of difficulty, because 
the existing methods of sealing electrical conductors, such as platinum 
wires, through Jena glass were found to be very unsatisfactory. In the 
course of our experiments an entirely new method of sealing electrical 
conductors through glass was developed by one of us, and it has been used 
by us since with the most satisfactory results. This method is described in 
a paper at present in the hands of the Institution of Electrical Engineers.^ 
Briefly stated, it consists of fusing the glass round the electrical conductor 
and then cooling the seal thus formed, when at about a dull red heat, 
by successive immersions in a suitable cooling medium, such as sperm or 
other oil. 
The Jena-glass lamps which we constructed for spectroscopic work were 
designed for a pressure of 100 volts; they were made from Jena-glass 
combustion tubing of 4 mms. internal and 6 mms. external diameter, the 
* The time that elapses between the switching on of the current and the striking of the 
arc depends, for a given heating element and a given current strength, on two factors, 
namely, the pressure in the small vessel due to the head of mercury above the constricted 
portion, and the quantity of mercury contained in the small vessel. The lamp, as shown in 
fig. 8, was designed to cut down to a minimum the time of starting. 
t “ A New Method of Sealing Electrical Conductors through Glass,” G. B. Burnside. 
