repeated several times:  contact was broken at r, then broken at G, next made at r, and lastly renewed at G; thus any current from N to P due to breaking of contact was avoided, but any additional force to the current from P to N due to making contact could be observed.  In this way it was found, that a much greater decomposing effect (causing the evolution of iodine against P) could be obtained by a few completions of contact than by the current which could pass in a much longer time if the contact was continued.  This I attribute to the act of induction in the wire ABD at the moment of contact rendering that wire a worse conductor, or rather retarding the passage of the electricity through it for the instant, and so throwing a greater quantity of the electricity which the electromotor could produce, through the cross wire passage NP.  The instant the induction ceased, ABD resumed its full power of carrying a constant current of electricity, and could have it highly increased, as we know by the former experiments (1060.) by the opposite inductive action brought into activity at the moment contact at Z or C was broken.

1103.  A galvanometer was then introduced at x, and the deflection of the needle noted whilst contact was continued at G and E:  the needle was then blocked as before in one direction (1087.), so that it should not return when the current ceased, but remain in the position in which the current could retain it.  Contact at G or E was broken, producing of course no visible effect; it was then renewed, and the needle was instantly deflected, passing from the blocking pins to a position still further from its natural place than that which the constant current could give, and thus showing, by the temporary excess of current in this cross communication, the temporary retardation in the circuit ABD.

1104.  On adjusting a platina wire at x (1081.) so that it should not be ignited by the current passing through it whilst contact at G and E was continued, and yet become red-hot by a current somewhat more powerful, I was readily able to produce its ignition upon making contact, and again upon breaking contact.  Thus the momentary retardation in ABD on making contact was again shown by this result, as well also as the opposite result upon breaking contact.  The two ignitions of the wire at x were of course produced by electric currents moving in opposite directions.

1105.  Using the helix only, I could not obtain distinct deflections at x, due to the extra effect on making contact, for the reasons already mentioned (1088.).  By using a very fine platina wire there (1083.), I did succeed in obtaining the igniting effect for making contact in the same manner, though by no means to the same degree, as with the electro-magnet (1104).