^06 Prof. Oersted ofr an apparently 'paradoxical 
tween H and A, or, as in that in Fig. 3. formed by the copper 
JK, the zrnc KC, and the fluid. 
The contact of the copper and zinc above the water, or at the 
surface of the water, produces no eflect. In order to make 
myself sure that a collateral galvanic circuit was capable of pro- 
ducing such an effect on a homogeneous metallic circuit, I made 
the construction shewn at Fig 4., in which ABODE is the same 
homogeneous circuit as before, but J GF is an arch of copper, 
in contact with the zinc at J, and separated from it at F by a 
fold ^ ^ of paper- wetted in acidulated water. In alternately 
opening and closing this circuit, I found that the needle moved 
as in the preceding cases. This construction may therefore be 
considered as composed of a galvanic circuit JGFJ, and a con- 
ductor J CDEBAF, which transmits a current similar to that 
transmitted by JGF. 
This explanation is confirmed by an experiment made with 
the construction represented in Fig. 5. in which ZZ is a plate of 
zinc, CCCC a plate of copper, and A A a vessel of acidulated 
water. When a magnetic needle is placed at NS, it is deflect- 
ed according to the known laws ; but, if to this circuit there be 
added the conductor KKK, a part of the electricity passes in it, 
and acts more feebly on the needle, from being at a greater 
distance from it. The effect of the second conductor becomes 
more striking when the needle is placed at n s, and when (after 
having noted its deviation) the conductor KKK is added ; be- 
cause, in this case, the second conductor being above the needle, 
tends to give it a contrary deflexion to that given by the first 
conductor, which is below it. These experiments have the same 
result when ZZ is made of copper, and CCC of zinc. 
On - applying all this to the constructions in Figs 2. and 3.., we 
observe, that CDEBA is the same thing as the second conduc- 
tor in Fig. 5. and that the current in the part DE (Figs. S. and 
3.) should have the same direction as in the part CA (to which 
it is parallel), in the same manner that the currents are similar 
in CCC and KKK, in Fig. 5. This being granted, we can 
determine the direction of the current in all the other parts of 
CDEB A (Figs. 2. and 3»), and experiments with the needle will 
confirm the predictions of the theory. In some experiments, M. 
¥on Moll substituted a plate of zinc for the copper, with which he 
