458 PROCEEDINGS OF THE AMERICAN ACADEMY 



may be of suflicient magnitude to destroy the value of the observation. 

 In a continuous record, we could not compare the results when this 

 effect was acting with those when it was not acting, and deduce any 

 valuable laws. It was thought that by connecting the positive pole 

 of the battery to one set of quadrants, and the negative pole to the 

 other set, and then connected first to the ground and then to the col- 

 lector, (the deflection of the needle being noted in each case,) the effect 

 of a local charge in the around could be eliminated. This was tried 

 with very good results, but it would not be accurate if rapid changes 

 were taking place at the time of observation, nor could it be used to 

 obtain a continuous record. 



In order, therefore, to overcome the various difficulties of the col- 

 lectors described, we made use of the following principle. 



Maxwell* says: "Now let us suppose a firm insulated wire carried 

 from the electrode of the electrometer to the place where the potential 

 is to be measured. Let the sphere be first completely discharged. 

 This may be done by putting it into the inside of a vessel of the same 

 metal which nearly surrounds it, and making it touch the vessel. Now 

 let the sphere thus discharged be carried to the end of the wire and 

 made to touch it. Since the sphere is not electrified, it will be at the 

 potential of the air at the place. If the electrode wire is at the same 

 potential, it will not be affected by the contact; hut if the electrode is 

 at a different potential, it will by contact with the sphere be made 

 nearer to that of the air than it was before. By a succession of such 

 operations, the sphere being alternately discharged and made to touch 

 the electrode, the potential of the electrode will continually approach 

 that of the air at the given point." 



We applied this prniciple to experiment 2 by making G a metal 

 cup , but the height the ball was raised was so small that we aban- 

 doned that method, and constructed the following collector, which 

 seems to be free from all the greater objections. It consists of a 

 framework supporting a brass cup. A non-conducting string with 

 a brass sphere at one end passes through a system of pulleys. A 

 brass plate is attached to the framework just above the cup. It is 

 insulated from the framework by a thick layer of paraffine. The cup 

 is connected to the ground, and the plate to the electrometer. By 

 means of the string the sphere can be made to touch the inside of the 

 cup and the plate successively, and thus reduce the plate to the 

 potential of the air. 



* Elect, and Mag., sect. 221. 



