92 Rood — Experiments on High Electrical Resistance. 



cores of the units about to be described are inaccessible to 

 moisture, and that in ordinary weather there is no surface con- 

 duction that interferes with their accuracy. Less is known 

 about the invariability of the units during long periods of 

 time; this matter is being investigated by Mr. H. C. Parker, 

 and the results thus far obtained are promising. If it should 

 turn out that none of the varieties I have devised has this 

 desirable quality, it would merely necessitate on each occasion 

 where accuracy was required, a galvanometer measurement, 

 and a building up process from it, involving two or three 

 hours of extra labor, as changes in the units do not occur 

 which would interfere with the ease of a building up process 

 once established in all its details. In every case thus far exam- 

 ined, the resistance of the unit immediately after making it 

 has increased with some regularity, week by week, but at a 

 diminishing rate, and now at the present time, it is not my 

 custom to employ units till they are at least three months old. 

 Electrometer. — The construction of this apparatus is shown 

 in the sketch, fig. 1, where all the shaded portions are metallic, 

 and the unshaded parts made of the best quality of ebonite. 

 B is a brass binding post, to which may be attached the wire 

 conveying the charge, also a Leyden jar or a small mica con- 

 denser. It terminates below in an aluminum leaf, bluntly 

 pointed, and provided with a long hinge of gold leaf for flexi- 

 bility. The latter is at- 

 tached to a brass wire W, 

 that slides into a well-tit- 

 ting tube, from which it 

 can readily be removed for 

 repairs. P is a springy 

 brass plate faced with plati- 

 num. When the apparatus 

 has received a sufficient 

 charge the aluminum leaf 

 is attracted to P, and 

 usually remains attached ; 

 Y is a narrow strip of thin 

 sheet brass fastened to an 

 axis, not shown in the dia- 

 gram, which is terminated 

 by the milled head H ; it 

 can be made to slightly 

 scrape the edge of the 

 plate P, and communicate 

 to it a trifling vibratory movement, which instantly detaches 

 the gold-aluminum leaf. Mere contact from behind, even 

 when made with some violence, is not sufficient for this pur- 



