634 Proceedings of the Royal Society of Edinburgh. [Sess. 
so that the instrument was by no means in its most sensitive adjustment. 
The delicacy of the above method of detecting moisture is, however, so 
great, that a surface of conducting glass, even when warm, and after careful 
polishing with a cloth, may give a deflection of as much as 100 mms. in 
three seconds. 
A freshly made plate, tested directly it had cooled sufficiently, gave no 
deflection at first. 
After a few hours it was again examined, and a large deflection obtained. 
Old plates were found to give a much smaller action than new ones which 
had been exposed to air and dust for a few hours. 
In addition to providing a means of numerically expressing the changes 
in the hydroscopic property of the surface, the electrical test served to 
eliminate many of the glasses which for one reason or another had seemed 
sufficiently well constituted. It was soon found, for instance, that glass 
No. 50, although resisting acid comparatively well, and being satisfactory 
in other ways, would not improve with age, and was always greedy for 
moisture. A plate of it left exposed over night would generally give a 
deflection of 150 mms. in two seconds the next morning. 
The experiments of Kohlrausch, Mylius and Foerster strongly support 
the view that the surface of certain glasses may, as regards its resistance to 
water and acids, be improved by contact with those liquids. 
It may be pointed out, for instance, that lead glass is chosen for holding 
wine because long contact with dilute acid is found to increase the power 
of the glass to resist further attack, and that water boiled in a flask 
dissolves out more alkali during the first than during subsequent hours. A 
plate of glass No. 50 was therefore boiled for five minutes to see whether 
the surface would be improved thereby. 
It was placed afterwards in a warm muffle to dry, and unfortunately 
came out opaque with fine scratches. The previous electrical tests of the 
surface should have served as a warning. It remained, therefore, to revert 
to the original mixture and to add a trace of sodium antimoniate in order 
to remove colour. A plate newly made was broken in two. One half was 
boiled for five minutes, and both halves were then placed in the muffle 
for two hours. Their surfaces were unaffected by this treatment and the 
electrical test showed that there was no moisture whatever upon either piece. 
Both halves were then laid side by side, and exposed upon the bench to air 
and dust. After three hours the unboiled piece A gave a deflection of 20 mms. 
in 100 seconds, while the boiled portion B gave no action. The next morning 
A gave 105 mms. in 20 seconds (fig. 2), and from that time varied with the 
saturation of the atmosphere, while B produced no deflection whatever for 
