8 . Experiments on the Undulation ofEUcli icily. 



on the one HJe than the other, and this irregularity may be inacafcd alfo from other caufcs. 

 The irregular fupply will therefore be made at more dillant intervals by the fimple ma- 

 chine, ^t more frequent reiterations in the machine with the multiplying wheel, and per- 

 haps uniformly from a plate. 



V. The cffiia of this undulation may be fhewn from various fafts. i. A fmall wire of 

 many yards in length, communicating from a b.ill to the ground, will be rendered luminous 

 through its whole length, by fparks of pofitive eledrieity given to the ball ; but it will not be , 

 at all luminous when the fame quantity, of eledricity is given in a more continued Rream by 

 placing the ball in contaft with the prime condudor. 2. If a metallic rod be infulated 

 with a ball at one end, of a diameter fuflicicnt to prevent the fpontaneous flaflilng of elec- 

 tricity when it forms part of the prime condudor, and at the other end of the Item be 

 fixed another ball of a proper fize to draw the fpark ; when this laft ball is placed in con- 

 taa with the prime conduftor, the other ball will not throw out flaQies ; but if it be with- 

 drawn fo as to receive fparks, the external ball, though certainly no more elearified than 

 before, will throw out a Hadi every time the fpaik is emitted. 3. A brafs ball of four 

 inches diameter was conneaed by a metallic Hem to the end of the prime condudor of an 

 clearical machine in the pofitive itate, fufficiently vigorous to throw a flaOi into the air now 

 and then. The metallic ftcm, which was about fix inches long, was then changed for ano- 

 ther of the fame dimenfions of deal wood. In this ftate the ball threw out continual llallies 

 into the air. The experiment was frequently repeated, and the laft refult may naturally be 

 fiippofed to depend on the imperfedion or difcontinuity of the condudUng matter in the 

 wooden fteni. 4. A pointed wire was inferred in the pofitive condudor of Nairne's elec- 

 trical machine, with the point upwards. It was then covered with a clean Florence flafl:. 

 '1 he point of the wire occupied the centre of the bottle. "Whenever the pofitive fpark was 

 drawn from the condudor, the point of the wire exhibited the luminous negative fign. But 

 ■when the experiment was repeated on the negative condudor and the fpark drawn, the point 

 emitted the pofitive flafti fo as to fill the whole capacity of the bottle with ramified light. 

 In thefe experiments it may be prefumed, that the efcape at the point was occafioned by un- 

 dulation. 5. The lateral fpark in difcharging a jar may be urged as an inftance of the 

 fame kind. 



Thefe obfcrvations feem to give the advantage in favour of plate -machines as far as re- 

 lates to the efcape. I fhall therefore proceed to defcvibe that of Dr. Van Alarum*, and 

 afterwards ftate from his report, compared with my own experiments, what may be the 

 proportions of eledricity coUeded from each fquare foot, which pafles the cudiion in ma- 

 chines of both kinds. 



Plate III exhibits a perfpedive view of the machine, and Tlatc IV a fcdion, cxckifive of 

 the culhions. In the view it may be obl'erved that the eufliions are each feparately infu- 

 lated upon pillars of glafs, and are applied nearly in the dircdton of the horizontal diameter 

 of the plate, inftead of the vertical diameter as heretofore. The ball diametrically oppofite 

 to the handle is the prime condudor, and the femicircular piece with two cylindrical ends 

 fcrvcs, in the pofition of the drawing, to receive the eledricity from the plate. By the happy 

 contrivance of altering the pofition of tlris femicircular branch from vertical to nearly hori- 



* Abridged from the Seconite CmlinualioH des Expiriciicnfiuus put U Moyen de la Machine Ele<liique XtyU- 

 liennc. 



zontal, 



