632 
Proceedings of the Royal Society of Edinburgh. [Sess. 
Messrs Johnson & Mathey made for me a slightly tapered silica tube 
two feet long, and having at its thick end a thin platinum sleeve welded on 
so as to extend eight inches up the tube. A cap of platinum closes the 
lower end, and the upper narrow one is left open to allow air to escape 
when the rod is lowered into the molten glass. By this means a light, 
handy stirrer is obtained at a moderate cost, and having even greater 
rigidity than a solid platinum bar of the same dimensions. An asbestos 
screen protects the hand from the heat of the furnace, and the mixture 
may be stirred for thirty minutes or more, continuously, without 
discomfort. 
Since stirring introduces a number of fresh bubbles, the crucible has to 
go into the vacuum pot another four times to remove them, and care must 
then be taken not to let the temperature of the mass become too high. At 
a full white heat the evaporation of the borax is greatly accelerated 
in vacuo , and striae are abundant in the finished glass in spite of the 
additional mixing which the vacuum undoubtedly also produces. After 
eight times the bubbles have all disappeared, so that the contents may 
be poured out after about two hours of full heat. The glass is then cast 
or pressed into sheets as required, and removed to the annealing muffle. 
It is necessary to have this at a dull red heat, and to ensure that its 
temperature shall fall very slowly. A prism of conducting glass made 
last April was annealed for only ten hours, and showed afterwards no 
trace of stress whatever. 
A plate of glass so prepared indicated, however, by a milky appearance 
of its surface after exposure to moist air, that it was not sufficiently stable. 
It was also seen that the clay crucible used for roughly melting up the 
mixture before transferring it to the platinum pot had given a bluish-green 
colour to the plate. This was traced to impurities in the clay. A drop 
of HC1 (1 in 5) rapidly attacked the surface, and left a white crust of 
crystals above a layer of fine cracks. 
The problem of improving the glass as regards this attack by acid was 
partially solved in a new series by adding more lead and less borax. The 
Powell glass was left out and the following mixture — 
32 parts of sodium silicate, 
7 parts of borax, 
*7 part of lead oxide (red lead), 
T part of sodium antimoniate, 
gave a material better in some respects than the original, but easily 
devitrifying on re-heating. 
