DowLiNG — A Direct Reading Ultra-Micromeier. 187 



certain value of the capacity, sometimes more than one, for which the plate 

 current reaches a maximum. It is clearly necessary to adjust the apparatus- 

 so that it is not functioning near such maximal points. 



The first trial of the apparatus as a micrometric device was carried out 

 as follows; — Two circular plates of steel, 6'4 cm. in diameter, and ground 

 reasonably fiat, were mounted on the stand of a Leitz microscope. One was- 

 cemented to the platform, while the other was carried on the end of a 

 glass rod which projected from the microscope tube, in which it was firmly 

 supported. The two plates were adjusted to be exactly parallel. The fine 

 adjustment screw of the microscope w'as furnished with a divided head, each 

 division of which corresponded to a vertical motion of the tube through 

 2-0-5- mm. It was thus found possible to move the upper plate through this 

 distance with reasonable certainty, the screw adjustment being remarkably 

 good. 



Observations were made as follows : — The galvanometer being shunted, 

 the plates were screwed apart a certain number of divisions, and the sliding 

 resistance S adjusted until the galvanometer was at zero. The shunt was then 

 gradually reduced ; slight further adjustments of S being made, if necessary,, 

 while this was being done. Except in the most sensitive settings, the shunt 

 was reduced to unity, and observations then made of the deflection of the 

 galvanometer resulting from the rotation of the microscope screw through 

 one scale division. It was generally found that a deflection of at least two or 

 three hundred divisions was thus produced ; but, by selecting a suitable 

 initial separation of the plates, much larger deflections could be obtained. It 

 was easy to set the apparatus so that about six hundred galvanometer 

 divisions corresponded to one division of the screw; that is to say, a displace- 

 ment of -aoVo cm- produced a deflection of the galvanometer of six hundred 

 scale divisions. In other words, a displacement of t.too.oTo cm. was detect- 

 able unmistakably. On more than one occasion a much higher sensitivity 

 was reached, but its magnitude could not be stated with certainty. The 

 smallest movement that could be given with any certainty to the plates was 

 that represented by one division on the screw-head. The highest sensitivity 

 that could be utilized was, therefore, already reached when the galvanometer 

 deflection extended, as just mentioned, six hundred divisions from one end 

 of the scale to the other. By shunting the galvanometer, the effect of moving 

 the plate could be reduced, and an estimate formed of higher degrees of 

 sensitivity. However, I am not at present certain if the sensitivity thus 

 estimated is correctly given by multiplying the galvanometer reading by the 

 shunt ratio. 



In order to produce displacements of sufiiciently minute nature, another 



