Page:Scientific Memoirs, Vol. 1 (1837).djvu/629

I soon perceived that from this none but false results could be obtained. By considerably widening the circuit of the spiral, it advanced nearer and nearer to the upper bow of the magnet; so that by removing the armature, not only the sudden disappearance of the magnetism in it, but also the sudden removal which took place at the same time from that upper part of the magnet (the bay of the horseshoe) acts on the spirals, and indeed unequally with unequal diameters of the spiral; the electromotive power becomes thus greater in larger spirals than it would otherwise be. On this account I took two strong rectilinear magnetic systems, each of which consisted of ten single magnet bars; I laid them with their opposite poles against one another so that they lay in a straight direction, and brought the iron cylinder which had served me in the above-mentioned experiments as an armature to the horseshoe magnet, between their poles, while the spirals covered the cylinder; upon this I let the magnet be suddenly drawn by two assistants from each other in opposite directions.

I wound at first ten convolutions of wire No. 2 round the iron cylinder

the diameter of the convolutions ═ 0·73 inch;

upon this I wound ten convolutions of wire No. 2 round a wooden disc,

the diameter of the convolutions ═ 6·57 inches;

the wooden disc was perforated in the centre, into which the iron cylinder was inserted. The observation gave

	Angle of Deviation.				Mean.
	1	2	3	4
Narrow Conv.	24·6	27·1	26·4	26·5	26·15
Wider Convolutions.	22·8	22·7	22·0	22·5	22·50
	23·4	23·5	21·6	23·2	22·92
Narrow Conv.	24·8	27·7	26·3	26·6	26·35

I observed the deviation of the wider spirals between the narrower, in order that the fault which might have originated by diminishing the magnetic power of the magnet systems might be estimated: we therefore have

for the narrower spirals the angle of deviation $\alpha =26\cdot 25$
for the wider spirals the angle of deviation $\alpha '=22\cdot 71$

The length of the wire of the multiplier and of the conducting wires (reduced to the diameter of the first) was as in the former experiments, i. e. they amounted together to $673\cdot 25$, or $L+l=673\cdot 25,\,\lambda$ however is for the narrower convolutions $=28$, and for the wider $\lambda '=203$.