DoWLiNG — Sensitive Valve Method for Measurement of Capacity. 183 



fellow, the cup was filled with molten fusible metal (90° C. m.p.), which 

 rapidly cooled and cemented the metal plate to the glass support. When cool 

 the plates were separated by turning the micrometer screw the desired 

 amount. This condenser allowed a much higher sensitivity to be attained. 

 Fig. 6 shows the first set of observations thus obtained. During this set the 

 apparatus was shielded in the metal box previously mentioned, and the con- 

 ditions were fairly steady. The points lie almost on a straight line, but the 

 gradual increase in sensitiveness with decreasing separation of the plates is 

 just noticeable, the points really lying on a parabolic curve. (The observations 

 were taken with the galvanometer-shunt at 10, but the ordinates in fig. 6 

 have been plotted as if the galvanometer was unshunted.) The average 

 sensitivity throughout the range is about 3200 scale divisions for 

 ToVo ^™' ' i-®- 1 division deflection corresponds to a displacement of 

 about 3 X 10"' cms. 



Several sets of observations were made with this apparatus with the plates 

 at different distances apart ; and it seemed probable that, under sufficiently 

 steady conditions, it would be possible to detect a displacement of at least 

 '5 X 10'' cms. Further work with the apparatus is still being carried on, and 



1 will not, therefore, give further particulars in this paper. 



A sensitive pressure-gauge may easily be constructed involving the same 

 principle. A shallow recess was turned in an ebonite disc, and a flat metal 

 disc (6'4 cm. diameter) was cemented therein, so that its surface lay about 



2 mm. below the ebonite rim. A similar disc, about '6 mm. in thickness, was 

 cemented to a sheet of india-rubber (such as is used for foot-bellows). The 

 rubber was then cemented to the ebonite rim while slightly stretched, the 

 metal plates being thus supported about l^S mm. apart. Suitable connexions 

 were, of course, made by fine wires from each of these plates, and a tube 

 through the ebonite gave connexion with the air-space inside. When the 

 pressure rises within, the plates are driven apart, and the capacity of the 

 system diminishes. The apparatus was roughly tested, and it was found that 

 a pressure excess of 1 cm. of water resulted in a deflection of the galvano- 

 meter (shunt = 10) of about 300 divisions — or, roughly, 3 galvanometer 

 divisions (unshunted) corresponded to a pressure of 1 dyne per cm.^ It is 

 •obvious that with improved construction, larger discs both of metal and 

 rubber, thinner rubber, and closer proximity of the plates, it should be 

 possible to measure very minute pressure differences indeed. 



I have said enough to indicate the possibilities of the method, and hope 

 to publish further particulars when I have succeeded in attaining satisfactory 



SOIENT. PEOC. R.D.S., VOL. XVI, NO. XVII. 2 E 



