j^6 D^fcriptkn of the Revolving Douiler, 



D, will conftitute another mafs. All.the experiments yet made concur to prove, that thefe 

 two maflcs will not poflefs the fame eledric ftate ; but that, with refpeft to each other, 

 their eledricities will be plus and minus. Thefe ftates would be fimple and without any 

 compenfation, if the mafles were remote from each other ; but as that is not the cafe, a 

 part of the redundant cle£lricity will take the form of a charge in the oppofed plates A 

 and B. From other experiments I find that the effeft of the compenfation on plates op- 

 pofed to each other, at the dillance of one-fortieth part of an inch, is fuch that they 

 require, to produce a given intenfity, at leaft one hundred times the quantity of eleftricity 

 that would have produced it in either, fingly and apart. The redundant electricities in the 

 mafles under confideration will therefore be unequally diftributed : the plate A will have 

 about ninety-nine parts, and the plate C one; and, for the fame reafon, the revolving 

 plate B will have ninety-nine parts of the oppofite eleflricity, and the ball D one. The 

 rotation, by deftroying the contacts, preferves this unequal diftribution, and carries B from 

 A to C, at the fame time that the tail K connefts the ball with the plate C. In 

 this Ctuation, the eleftricity in B ads upon that in C, and produces the contrary ftate, 

 by virtxie of the communication between C and the ball ; which laft muft therefore ac- 

 quire an eledlricity of the fame kind with that of the revolving plate. But the rotation 

 again deflroys the contadt, and reftores B to its firft fituation oppofite A. Here, if we 

 attend to the effe£l: of the whole revolution, -we fhall find that the eledtric ftates of the 

 refpedlive rnafles have been greatly increafed : for the ninety-nine parts in A and in B 

 remain, and the one part of electricity in C has been increafed fo as nearly to com- 

 penfate ninety-nine parts of the oppofite electricity in the revolving pla-te B, while the 

 communication produced an equal mutation in the electricity of the ball. A fecond 

 rotation will, of courfe, produce a proportional augmentation of thefe increafed quantities ; 

 and a continuance of turning will foon bring the intenfities to their maximum, which is 

 limited by an explofion between the plates. 



If one of the parts be connected with an eleCtrometer, more efpecially that of Bennet, 

 thefe effects will be very clearly feen. The fpark is ufuaily produced by a number of 

 turns between eleven and twenty ; and the eleCtrometer is fenfibly aCted upon by ftill 

 fewer. 



If the ball be conneCled with the lower part of Bennet's eleCtrometer, and the plate A 

 with the upper part, and any weak eleCtricity be communicated to the eleCtrometer, while 

 the pofition of the apparatus is fuch that the crofs-piece GH touches the two pins ; a very 

 few turns will render it perceptible. But here, as well as in the common doubler, the 

 eflTeCl is rendered uncertain by the condition, that the communicated eleCtricity muft be 

 tftrong enough to deftroy and predominate over any other eleCtricity the plates may 

 poflefs. I fcarcely need obferve, that if this difficulty fhould hereafter be removed, the 

 inftrument will have great advantages as a multiplier of eleCtricity in the facility of its ufe, 

 the very fpeedy manner of its operation, and the unequivocal nature of its refults. 



