STOP 



Early Journal Content on JSTOR, Free to Anyone in the World 

This article is one of nearly 500,000 scholarly works digitized and made freely available to everyone in 
the world by JSTOR. 

Known as the Early Journal Content, this set of works include research articles, news, letters, and other 
writings published in more than 200 of the oldest leading academic journals. The works date from the 
mid-seventeenth to the early twentieth centuries. 

We encourage people to read and share the Early Journal Content openly and to tell others that this 
resource exists. People may post this content online or redistribute in any way for non-commercial 
purposes. 

Read more about Early Journal Content at http://about.jstor.org/participate-jstor/individuals/early- 
journal-content . 



JSTOR is a digital library of academic journals, books, and primary source objects. JSTOR helps people 
discover, use, and build upon a wide range of content through a powerful research and teaching 
platform, and preserves this content for future generations. JSTOR is part of ITHAKA, a not-for-profit 
organization that also includes Ithaka S+R and Portico. For more information about JSTOR, please 
contact support@jstor.org. 



L 171 ] 



VII. Experimental Researches in Electricity. — Twenty-third Series. By Michael 
Faraday, Esq., D.C.L., F.R.S., FullerianProf. Chem. Royal Institution, Foreign 
Associate of the Acad. Sciences, Paris, Ord. Boruss. Pour le Merite, Eq., Memb. 
Royal and Imp. Acadd. of Sciences, Petersburgh, Florence, Copenhagen, Berlin, 
Gbttingen, Modena, Stockholm, Munich, Bruxelles, Vienna, Bologna, &;c. 8§c. 

Received January 1, — Read March 7 and 14, 1850. 

§ 29. On the polar or other condition of diamagnetic bodies. 

2640. Jb OUR years ago I suggested that all the phenomena presented by diamag- 
netic bodies, when subjected to the forces in the magnetic field, might be accounted 
for by assuming that they then possessed a polarity the same in kind as, but the re- 
verse in direction of, that acquired by iron, nickel and ordinary magnetic bodies 
under the same circumstances (2429. 2430.). This view was received so favourably by 
Plucker, Reich and others, and above all by W. Weber*, that I had great hopes it 
would be confirmed ; and though certain experiments of my own (2497.) did not in- 
crease that hope, still my desire and expectation were in that direction. 

2641. Whether bismuth, copper, phosphorus, &c, when in the magnetic field, are 
polar or not, is however an exceedingly important question ; and very essential and 
great differences, in the mode of action of these bodies under the one view or the 
other, must be conceived to exist. I found that in every endeavour to proceed by 
induction of experiment from that which is known in this department of science to 
the unknown, so much uncertainty, hesitation and discomfort arose from the unsettled 
state of my mind on this point, that I determined, if possible, to arrive at some experi- 
mental proof either one way or the other. This was the more needful, because of the 
conclusion in the affirmative to which Weber had come in his very philosophical 
paper ; and so important do I think it for the progress of science, that, in those im- 
perfectly developed regions of knowledge, which form its boundaries, our conclusions 
and deductions should not go far beyond, or at all events not aside from the results 
of experiment (except as suppositions), that I do not hesitate to lay my present 
labours, though they arrive at a negative result, before the Royal Society. 

2642. It appeared to me that many of the results which had been supposed to in- 
dicate a polar condition, were only consequences of the law that diamagnetic bodies 
tend to go from stronger to weaker places of action (2418.) ; others again appeared to 
have their origin in induced currents (26. 2338.) ; and further consideration seemed 

* Poqqendorff's Annalen, January 7, 1848, or Taylor's Scientific Memoirs, v. p. 477. 

z2 



172 DR. FARADAY'S EXPERIMENTAL RESEARCHES IN ELECTRICITY. (SERIES XXIII.) 

to indicate that the differences between these modes of action and that of a real pola- 
rity, whether magnetic or diamagnetic, might serve as a foundation on which to base 
a mode of investigation, and also to construct an apparatus that might give useful 
conclusions and results in respect of this inquiry. For, if the polarity exists it must 
be in the particles and for the time permanent, and therefore distinguishable from 
the momentary polarity of the mass due to induced temporary currents ; and it 
must also be distinguishable from ordinary magnetic polarity by its contrary direc- 
tion. 

2643. A straight wooden lever, 2 feet in length, was fixed by an axis at one end, 
and by means of a crank and wheel made to vibrate in a horizontal plane, so that its 
free extremity passed to and fro through about 2 inches. Cylinders or cores of metal 
or other substances, 5 \ inches long and three-quarters of an inch diameter, were fixed 
in succession to the end of a brass rod 2 feet long, which itself was attached at the 
other end to the moving extremity of the lever, so that the cylinders could be moved 
to and fro in the direction of their length through the space of 2 inches. A large 
cylinder electro-magnet was also prepared (2191.), the iron core of which was 21 inches 
long and 1*7 inch in diameter; but one end of this core was made smaller for the 
length of 1 inch, being in that part only 1 inch in diameter. 

2644. On to this reduced part was fixed a hollow helix consisting of 516 feet of fine 
covered copper wire : it was 3 inches long, 2 inches external diameter, and 1 inch 
internal diameter : when in its place, 1 inch of the central space was occupied by the 
reduced end of the electro-magnet core which carried it ; and the magnet and helix 
were both placed concentric with the metal cylinder above mentioned, and at such a 
distance that the latter, in its motion, would move within the helix in the direction of 
its axis, approaching to and receding from the electro-magnet in rapid or slow suc- 
cession. The least and greatest distances of the moving cylinder from the magnet 
during the journey were one-eighth of an inch and 2 - 2 inches. The object of course 
was to observe any influence upon the experimental helix of fine wire which the metal 
cylinders might exert, either whilst moving to or from the magnet, or at different 
distances from it*. 

2645. The extremities of the experimental helix wire were connected with a very 
delicate galvanometer, placed 18 or 20 feet from the machine, so as to be unaffected 
directly by the electro-magnet; but a commutator was interposed between them. 
This commutator was moved by the wooden lever (2643.), and as the electric currents 
which would arrive at it from the experimental helix, in a complete cycle of motion 
or to and fro action of the metal cylinder (2643.), would consist of two contrary por- 
tions, so the office of this commutator was, sometimes to take up these portions in 
succession and send them on in one consistent current to the galvanometer, and at 

* It is very probable that if the metals were made into cylinders shorter, but of larger diameter than those 
described above, and used with a corresponding wider helix, better results than those I have obtained would 
be acquired. 



SUPPOSED POLARITY OF DIAMAGNETIC BODIES. 173 

other times to oppose them and to neutralize their result ; and therefore it was made 
adjustible, so as to change at any period of the time or part of the motion. 

2646. With such an arrangement as this, it is known that, however powerful the 
magnet, and however delicate the other parts of the apparatus, no effect will be pro- 
duced at the galvanometer as long as the magnet does not change in force, or in its 
action upon neighbouring bodies, or in its distance from, or relation to, the experi- 
mental helix ; but the introduction of a piece of iron into the helix, or anything else 
that can influence or be influenced by the magnet, can, or ought to, show a corre- 
sponding influence upon the helix and galvanometer. My apparatus I should ima- 
gine, indeed, to be almost the same in principle and practice as that of M. Weber 
(2640.), except that it gives me contrary results. 

2647. But to obtain correct conclusions, it is most essential that extreme precaution 
should be taken in relation to many points which at first may seem unimportant. All 
parts of the apparatus should have perfect steadiness, and be fixed almost with the 
care due to an astronomical instrument; for any motion of any portion of it is, from 
the construction, sure to synchronize with the motion of the commutator ; and por- 
tions of effect, inconceivably small, are then gathered up and made manifest as a 
whole at the galvanometer ; and thus, without care, errors might be taken for real 
and correct results. Therefore, in my arrangements, the machine (2643, &c), the 
magnet and helix, and the galvanometer stood upon separate tables, and these again 
upon a stone floor laid upon the earth ; and the table carrying the machine was care- 
fully strutted to neighbouring stone-work. 

2648. Again, the apparatus should itself be perfectly firm and without shake in its 
motion, and yet easy and free. No iron should be employed in any of the moving 
parts. I have springs to receive and convert a portion of the momentum of the whole 
at the end of the to and fro journey ; but it is essential that these should be of ham- 
mered brass or copper. 

2649. It is absolutely necessary that the cylinder or core in its motion should not 
in the least degree disturb or shake the experimental helix and the magnet. Such a 
shake may easily take place and yet (without much experience) not be perceived. It 
is important to have the cores of such bodies as bismuth, phosphorus, copper, &c, as 
large as may be, but I have not found it safe to have less than one-eighth of an inch 
of space between them and the interior of the experimental helix. In order to float, 
as it were, the core in the air, it is convenient to suspend it in the bight or turn of a 
fine copper wire passing once round it, the ends of which rise up, and are made fast 
to two fixed points at equal heights but wide apart, so that the wire has a V form. 
This suspension keeps the core parallel to itself in every part of its motion. 

2650. The magnet, when excited, is urged by an electric current from five pairs of 
Grove's plates, and is then very powerful. When the battery is not connected with 
it, it still remains a magnet of feeble power, and when thus employed may be re- 
ferred to as in the residual state. If employed in the residual state, its power may for 



174 DR. FARADAY'S EXPERIMENTAL RESEARCHES IN ELECTRICITY. (SERIES XXIII.) 

the time be considered constant, and the experimental helix may at any moment be 
connected with the galvanometer without any current appearing there. But if the 
magnet be employed in the excited state, certain important precautions are necessary; 
for upon connecting the magnet with the battery and then connecting the experi- 
mental helix with the galvanometer, a current will appear at the latter, which will, in 
certain cases, continue for a minute or more, and which has the appearance of being 
derived at once from that of the battery. It is not so produced, however, but is due 
to the time occupied by the iron core in attaining its maximum magnetic condition 
(2170. 2332.), during the whole of which it continues to act upon the experimental 
helix, producing a current in it. This time varies with several circumstances, and in 
the same electro-magnet varies especially with the period during which the magnet has 
been out of use. When first employed, after two or three days' rest, it will amount to 
eighty or ninety seconds, or more. On breaking battery contact and immediately re- 
newing it, the effect will be repeated, but oceupy only twenty or thirty seconds. On 
a third intermission and renewal of the current, it will appear for a still shorter period; 
and when the magnet has been used at short intervals for some time, it seems capa- 
ble of receiving its maximum power almost at once. In every experiment it is neces- 
sary to wait until the effect is shown by the galvanometer to be over ; otherwise the 
last remains of such an effect might be mistaken for a result of polarity, or some pecu- 
liar action of the bismuth or other body under investigation. 

2651. The galvanometer employed was made by Ruhmkorff and was very sensible. 
The needles were strengthened in their action and rendered so nearly equal, that a 
single vibration to the right or to the left occupied from sixteen to twenty seconds. 
When experimenting with such bodies as bismuth or phosphorus, the place of the 
needle was observed through a lens. The perfect communication in all parts of the 
circuit was continually ascertained by a feeble thermo-electric pair, warmed by the 
fingers. This was done also for every position of the commutator, where the film of 
oxide formed on any part by two or three days' rest was quite sufficient to intercept a 
feeble current. 

2652. In order to bring the phenomena afforded by magnetic and diamagnetic 
bodies into direct relation, I have not so much noted the currents produced in the 
experimental helix, as the effects obtained at the galvanometer. It is to be under- 
stood, that the standard of deviation, as to direction, has always been that produced 
by an iron wire moving in the same direction at the experimental helix, and with the 
same condition of the commutator and connecting wires, as the piece of bismuth or 
other body whose effects were to be observed and compared. 



2653. A thin glass tube, of the given size (2643.), 5 J by f inches, was filled with a 
saturated solution of protosulphate of iron, and employed as the experimental core: 
the velocity given to the machine at this and all average times of experiment was 



SUPPOSED POLARITY OF DIAMAGNETIC BODIES. 175 

such as to cause five or six approaches and withdrawals of the core in one second ; 
yet the solution produced no sensible indication at the galvanometer. A piece of 
magnetic glass tube (2354.), and a core of foolscap paper, magnetic between the 
poles of the electro-magnet, were equally inefficient. A tube filled with small crystals 
of protosulphate of iron caused the needle to move about 2°, and cores formed out of 
single large crystals, or symmetric groups of crystals of sulphate of iron, produced the 
same effect. Red oxide of iron (colcothar) produced the least possible effect. Iron 
scales and metallic iron (the latter as a thin wire) produced large effects. 

2654= Whenever the needle moved, it was consistent in its direction with the effect 
of a magnetic body ; but in many cases, with known magnetic bodies, the motion 
was little or none. This proves that such an arrangement is by no means so good a 
test of magnetic polarity as the use of a simple or an astatic needle. This deficiency 
of power in that respect does not interfere with its ability to search into the nature 
of the phenomena that appear in the experiments of Wkber, Reich and others. 

2655. Other metals than iron were now employed and with perfect success. If 
they were magnetic, as nickel and cobalt, the deflection was in the same direction as 
for iron. When the metals were diamagnetic, the deflection was in the contrary direc- 
tion ; and for some of the metals, as copper, silver and gold, it amounted to 60° or 70°, 
which was permanently sustained as long as the machine continued to work. But 
the deflection was not the greatest for the most diamagnetic substances, as bismuth 
or antimony, or phosphorus ; on the contrary, I have not been able to assure myself, 
up to this time, that these three bodies can produce any effect. Thus far the effect 
has been proportionate to the conducting power of the substance for electricity. Gold, 
silver and copper have produced large deflections, lead and tin less. Platina very little. 
Bismuth and antimony none. 

2656. Hence there was every reason to believe that the effects were produced by 
the currents induced in the mass of the moving metals, and not by any polarity of 
their particles. I proceeded therefore to test this idea by different conditions of the 
cores and the apparatus. 

2657- In the first place, if produced by induced currents, the great proportion of 
these would exist in the part of the core near to the dominant magnet, and but little 
in the more distant parts ; whereas in a substance like iron, the polarity which the 
whole assumes makes length a more important element. I therefore shortened the 
core of copper from h\ inches (2643.) to 2 inches, and found the effect not sensibly 
diminished ; even when 1 inch long it was little less than before. On the contrary, 
when a fine iron wire, 5| inches in length, was used as core, its effects were strong; 
when the length was reduced to 2 inches, they were greatly diminished ; and again, 
with a length of 1 inch, still further greatly reduced. It is not difficult to construct 
a core of copper, with a fine iron wire in its axis, so that when above a certain 
length it should produce the effects of iron, and beneath that length the effects 
of copper. 



176 DR. FARADAY'S EXPERIMENTAL RESEARCHES IN ELECTRICITY. (SERIES XXIII.) 

2658. In the next place, if the effect were produced by induced currents in the 
mass (2642.), division of the mass would stop these currents and so alter the effect ; 
whereas if produced by a true diamagnetic polarity, division of the mass would not 
affect the polarity seriously, or in its essential nature. (2430.). Some copper filings 
were therefore digested for a few days in dilute sulphuric acid to remove any adhe- 
ring iron, then well-washed and dried, and afterwards warmed and stirred in the air, 
until it was seen by the orange colour that a very thin film of oxide had formed upon 
them : they were finally introduced into a glass tube (2653.) and employed as a core. 
It produced no effect whatever, but was now as inactive as bismuth. 

2659. The copper may however be divided so as either to interfere with the 
assumed currents or not, at pleasure. Fine copper wire was cut up into lengths of 
b\ inches, and as many of these associated together as would form a compact cylin- 
der three-quarters of an inch in diameter (2643.) ; it produced no effect at the galva- 
nometer. Another copper core was prepared by associating together many discs of 
thin copper plate, three-quarters of an inch in diameter, and this affected the gal- 
vanometer, holding its needle 25° or 30° from zero. 

2660. I made a solid helix cylinder, three-quarters of an inch in diameter and 
2 inches long, of covered copper wire, one-sixteenth of an inch thick, and employed 
this as the experimental core. When the two ends of its wire were unconnected, 
there was no effect upon the experimental helix, and consequently none at the gal- 
vanometer ; but when the ends were soldered together, the needle was well affected. 
In the first condition, the currents, which tended to be formed in the mass of moving 
metal, could not exist because the metal circuit was interrupted ; in the second they 
could, because the circuit was not interrupted ; and such division as remained did not 
interfere to prevent the currents. 

2661. The same results were obtained with other metals. A core cylinder of gold, 
made of half-sovereigns, was very powerful in its effect on the galvanometer. A cy- 
linder of silver, made of sixpenny pieces, was very effectual ; but a cylinder made of 
precipitated silver, pressed into a glass tube as closely as possible, gave no indications 
of action whatever. The same results were obtained with disc cylinders of tin and 
lead, the effects being proportionate to the condition of tin and lead as bad con- 
ductors (2655.). 

2662. When iron was divided, the effects were exactly the reverse in kind. It was 
necessary to use a much coarser galvanometer and apparatus for the purpose ; but 
that being done, the employment of a solid iron core, and of another of the same size 
or weight formed of lengths of fine iron wire (2659.), showed that the division had 
occasioned no inferiority in the latter. The excellent experimental researches of 
Dove* on the electricity of induction, will show that this ought to be the case. 

2663. Hence the result of division in the diamagnetic metals is altogether of a 
nature to confirm the conclusion, that the effects produced by them are due to in- 

* Taylor's Scientific Memoirs, v. p. 129. I do not see a date to the paper. 



SUPPOSED POLARITY OF DIAMAGNETIC BODIES. 177 

duced currents moving through their masses, and not to any polarity correspondent 
in its general nature (though opposed in its direction) to that of iron. 

2664. In the third place (2656.), another and very important distinction in the ac- 
tions of a diamagnetic metal may be experimentally established according as they 
may be due either to a true polarity, or merely to the presence of temporary induced 
currents ; and as for the consideration of this point diamagnetic and magnetic pola- 
rity are the same, the point may best be considered, at present, in relation to iron. 

2665. If a core of any kind be advanced towards the dominant magnet and with- 
drawn from it by a motion of uniform velocity, then a complete journey or to and 

from action might be divided into four parts ; the to, the stop after it ; the from, and 
the stop succeeding that. If a core of iron make this journey, its end towards the 
dominant magnet becomes a pole, rising in force until at the nearest distance, and 
falling in force until at the greatest distance. Both this effect, and its progression 
inwards and outwards, cause currents to be induced in the surrounding helix, and 
these currents are in one direction as the core advances, and in the contrary direction 
as it recedes. In reality, however, the iron does not travel with a constant velocity ; 
for, because of the communication of motion from a revolving crank at the machine 
(2643.), it, in the to part of the journey, gradually rises from a state of rest to a maxi- 
mum velocity, which is half-way, and then as gradually sinks to rest again near the 
magnet : — and the from part of the journey undergoes the same variations. Now as 
the maximum effect upon the surrounding experimental helix depends upon the velo- 
city conjointly with the intensity of the magnetic force in the end of the core, it is 
evident that it will not occur with the maximum velocity, which is in the middle of 
the to or from motion ; nor at the stop nearest to the dominant magnet, where the 
core end has greatest magnetic force, but somewhere between the two. Nevertheless, 
during the whole of the advance, the core will cause a current in the experimental 

- helix in one direction, and during the whole of the recession it will cause a current 
in the other direction. 

2666. If diamagnetic bodies, under the influence of the dominant magnet, assume 
also a polar state, the difference between them and iron being only that the poles of 
like names or forces are changed in place (2429. 2430.), then the same kind of action 
as that described for iron would occur with them ; the only difference being, that the 
two currents produced would be in the reverse direction to those produced by iron. 

2667. If a commutator, therefore, were to be arranged to gather up these currents, 
either in the one case or the other, and send them on to the galvanometer in one 
consistent current, it should change at the moments of the two stops (2665.), and 
then would perform such duty perfectly. If, on the other hand, the commutator 
should change at the times of maximum velocity or maximum intensity, or at two 
other times equidistant either from the one stop or from the other, then the parts of 
the opposite currents intercepted between the changes would exactly neutralize each 
other, and no final current would be sent on to the galvanometer. 

mdcccl. 2 A 



178 DR. FARADAY'S EXPERIMENTAL RESEARCHES IN ELECTRICITY. (SERIES XXIII.) 

2668. Now the actions ©£ the iron is, by experiment, of this nature. If an iron wire 
be simply introduced or taken out of the experimental helix with different conditions 
of the commutator, the results are exactly those which have been stated. If the ma- 
chine be worked with an iron wire core, the commutator changing at the stops (2665.), 
then the current gathered up and sent on to the galvanometer is a maximum ; if the 
commutator change at the moments of maximum velocity, or at any other pair of 
moments equidistant from the one stop or the other, then the current at the com- 
mutator is a minimum^, or 0. 

2669. There are two or three precautions which are necessary to the production of 
a pure result of this kind. In the first place, the iron ought to be soft and not pre-, 
viously in a magnetic state. In the next, an effect of the following kind has to be 
guarded against. If the iron core be away from the dominant magnet at the begin- 
ning of an experiment, then, on working the machine, the galvanometer will be seen 
to move in one direction for a few moments, and afterwards, notwithstanding the con- 
tinued action of the machine, will return and gradually take up its place at 0°. If the 
iron core be at its shortest distance from the dominant magnet at the beginning of the 
experiment, then the galvanometer needle will move in the contrary direction to that 
which it took before, but will again settle at 0°. These effects are due to the circum- 
stance, that, when the iron is away from the dominant magnet, it is not in so strong 
a magnetic state, and when at the nearest to it is in a stronger state, than the mean 
or average state, which it acquires during the continuance of an experiment ; and 
that in rising or falling to this average state, it produces two currents in contrary 
directions, which are made manifest in the experiments described. These existing 
only for the first moments, do, in their effects at the galvanometer, then appear, pro-, 
during a vibration which gradually passes away. 

2670. One other precaution I ought to specify. Unless the commutator changes 
accurately at the given points of the journey, a little effect is gathered up at each 
change, and may give a permanent deflection of the needle in one direction or the 
other. The tongues of my commutator, being at right angles to the direction of mo- 
tion and somewhat flexible, dragged a little in the to and from parts of the journey : 
in doing this they approximated, though only in a small degree, to that which is the 
best condition of the commutator for gathering up (and not opposing) the currents; 
and a deflection to the right or left appeared (2677-)- Upon discovering the cause 
and stiffening the tongues so as to prevent their flexure, the effect disappeared, and 
the iron was perfectly inactive. 

2671. Such therefore are the results with an iron core, and such would be the 
effects with a copper or bismuth core if they acted by a diamagnetic polarity. Let 
us now consider what the consequences would be if a copper or bismuth core were to 
act by currents, induced for the time, in its moving mass, and of the nature of those 
suspected (2642.). If the copper cylinder moved with uniform velocity (2665.), then 
currents would exist in it, parallel to its circumference, during the whole time of its 



SUPPOSED POLARITY OF DIAMAGNETIC BODIES. 179 

motion ; and these would be at their maximum force just before and just after the to 
or inner stop, for then the copper would be in the most intense parts of the magnetic 
field. The rising - current of the copper core for the in portion of the journey would 
produce a current in one direction in the experimental helix, the stopping - of the copper 
and consequent falling of its current would produce in the experimental helix a current 
contrary to the former ; the first instant of motion outwards in the core would produce 
a maximum current in it contrary to its former current, and producing in the experi- 
mental helix its inductive result, being a current the same as the last there produced; 
and then, as the core retreated, its current would fall, and in so doing and by its final 
stop, would produce a fourth current in the experimental helix, in the same direction 
as the first. 

2672. The four currents produced in the experimental helix alternate by twos, Le. 
those produced by the falling of the first current in the core and the rising of the 
second and contrary current, are in one direction. They occur at the instant before 
and after the stop at the magnet, i. e. from the moment of maximum current (in the 
core) before, to the moment of maximum current after, the stop ; and if that stop is 
momentary, they exist only for that moment, and should during that brief time be 
gathered up by the commutator. Those produced in the experimental helix during 
the falling of the second current in the core and the rising of a third current (iden- 
tical with the first) in the return of the core to the magnet, are also the same in 
direction, and continue from the beginning of the retreat to the end of the advance 
(or from maximum to maximum) of the core currents, i. e. for almost the whole of the 
core journey ; and these, by its change at the maximum moments, the commutator 
should take up and send on to the galvanometer. 

2673. The motion however of the core is not uniform in velocity, and so, sudden 
in its change of direction, but, as before said (2665.), is at a maximum as respects 
velocity in the middle of its approach to and retreat from the dominant magnet ; and 
hence a very important advantage. For its stop may be said to commence immediately 
after the occurrence of the maximum velocity; and if the lines of magnetic force were 
equal in position and power there to what they are nearer to the magnet, the con- 
trary currents in the experimental helix would commence at those points of the jour- 
ney ; but, as the core is entering into a more intense part of the field, tshe current in 
it still rises though the velocity diminishes, and the consequence is, that the maxi- 
mum current in it neither occurs at the place of greatest velocity, nor of greatest 
force, but at a point between the two. This is true both as regards the approach and 
the recession of the core, the two maxima of the currents occurring at points equi- 
distant from the place of rest near the dominant magnet. 

2674. It is therefore at these two points that the commutator should change, 
if adjusted to produce the greatest effect at the galvanometer by the currents excited 
in the experimental helix, through the influence of, or in connection with, currents 
of induction produced in the core ; and experiment fully justifies this conclusion. 

2 a2 



180 DR. FARADAY'S EXPERIMENTAL RESEARCHES IN ELECTRICITY. (SERIES XXIII.) 

If the length of the journey from the stop out to the stop in, which is 2 inches (2643. 
2644.), be divided into 100 parts, and the dominant magnet be supposed to be on 
the right-hand, then such an expression as the following, 50]50, may represent the 
place where the commutator changes, which in this illustration would be midway in 
the to and from motion, or at the places of greatest velocity. 

2675. Upon trial of various adjustments of the commutator, I have found that from 
77|23 to 88|12, gave the best result with a copper core. On the whole, and after 
many experiments, I conclude that with the given strength of electro-magnet, 
distance of the experimental core when at the nearest from the magnet, length of the 
whole journey, and average velocity of the machine, 86|14 may represent the points 
where the induced currents in the core are at a maximum and where the commutator 
ought to change. 

2676. From what has been said before (2667.)) it will be seen that both in theory 
and experiment these are the points in which the effect of any polarity, magnetic or 
diamagnetic, would be absolutely nothing. Hence the power of submitting by this 
machine metals and other bodies to experiment, and of eliminating the effects of mag- 
netic polarity, of diamagnetic polarity, and of inductive action, the one from the others : 
for either by the commutator or by the direction of the polarity, they can be sepa- 
rated ; and further, they can also be combined in various ways for the purpose of 
elucidating their joint and separate action. 

2677- For let the arrows in the diagram represent z, <r «?. , ff 

the to and from journey, and the intersections of the 
lines a, b or c, d, &c. the periods in the journey when 
the commutator changes (in which case c, d will cor- 
respond to 50|50, and e, / to 86|14), then a, b will 
represent the condition of the commutator for the ^ % ^ 

maximum effect of iron or any other polar body. If the line a, b be gradually re- 
volved until parallel to c, d, it will in every position indicate points of commutator 
change, which will give the iron effect at the galvanometer by a deflection of the 
needle always in the same direction ; it is only when the ends a and b have passed 
the points c and d, either above or below, that the direction of the deflection will 
change for iron. But the line a, b indicates those points for the commutator with 
which no effect will be produced on the galvanometer by the induction of currents 
in the mass of the core. If the line be inclined in one direction, as i, k, then these 
currents will produce a deflection at the galvanometer on one side ; if it be inclined 
in the other direction, as /, m, then the deflection will be on the other side. There- 
fore the effects of these induced currents may be either combined with, or opposed 
to, the effects of a polarity, whether it be magnetic or diamagnetic. 

2678. All the metals before mentioned (2655.), namely, gold, silver, copper, tin, lead, 
platina, antimony and bismuth, were submitted to the power of the electro- magnet 
under the best adjustment (2675.) of the commutator. The effects were stronger 




"7n- 



SUPPOSED POLARITY OF DIAMAGNETIC BODIES. 181 

than before, being now at a maximum, but in the same order ; as regarded antimony 
and bismuth, they were very small, amounting to not more than half a degree, and may 
very probably have been due to a remainder of irregular action in some part of the 
apparatus. All the experiments with the divided cores (2658, &c.) were repeated with 
the same results as before. Phosphorus, sulphur and gutta percha did not, either 
in this or in the former state of the commutator, give any indication of effect at the 
galvanometer. 

2679. As an illustration of the manner in which this position of the commutator 
caused a separation of the effects of copper and iron, I had prepared a copper cylinder 
core 2 inches in length having an iron wire in its axis, and this being employed 
in the apparatus gave the pure effect of the copper with its induced currents. Yet 
this core, as a whole, was highly magnetic to an ordinary test-needle ; and when the 
two changes of the commutator were not equidistant from the one stop or the other 
(2670. 2677 ■)» the iron effect came out powerfully, overruling the former and pro- 
ducing very strong contrary deflections at the needle. The platinum core which I 
have used is an imperfect cylinder, 2 inches long and 0*62 of an inch thick : it points 
magnetically between the poles of a horseshoe electro-magnet (2381.), making a 
vibration in less than a second, but with the above condition of the commutator 
(2675.) gives 4° of deflection due to the induced currents, the magnetic effect being 
annulled or thrown out. 

2680. Some of the combined effects produced by oblique position of the commu- 
tator points were worked out in confirmation of the former conclusions (26770* 
When the commutator was so adjusted as to combine any polar power which the 
bismuth, as a diamagnetic body, might possess, with any conducting power which 
would permit the formation of currents by induction in its mass (2676.), still the 
effects were so minute and uncertain as to oblige me to say that, experimentally, it 
is without either polar or inductive action. 

2681. There is another distinction which may usefully be established between the 
effects of a true sustainable polarity, either magnetic or diamagnetic, and those of 
the transient induced currents dependent upon time. If we consider the resistance 
in the circuit, which includes the experimental helix and the galvanometer coil, as 
nothing, then a magnetic pole of constant strength passed a certain distance into 
the helix, would produce the same amount of current electricity in it, whether the 
pole were moved into its place by a quick or a slow motion. Or if the iron core be 
used (2668.) the same result is produced, provided, in any alternating action, the 
core is left long enough at the extremities of its journey to acquire, either in its 
quick or slow alternation, the same state. This I found to be the fact when no com- 
mutator nor dominant magnet was used ; a single insertion of a weak magnetic pole 
gave the same deflection, whether introduced quickly or slowly ; and when the resi- 
dual dominant magnet, an iron wire core, and the commutator in its position a, b 
(26770 were used, four journeys to and from produced the same effect at the galva- 
nometer when the velocities were as 1 : 5 or even as 1 : 10. 



182 DR. FARADAY'S EXPERIMENTAL RESEARCHES IN ELECTRICITY. (SERIES XXIII.) 

2682. When a copper, silver, or gold core is employed in place of the iron, the effect 
is very different. There is no reason to doubt, that, as regards the core itself, the 
same amount of electricity is thrown into the form of induced circulating currents 
within it, by a journey to or from, whether that journey is performed quickly or 
slowly: the above experiment (2681.) in fact confirms such a conclusion. But the 
effect which is produced upon the experimental helix is not proportionate to the 
whole amount of these currents, but to the maximum intensities to which they rise. 
When the core moves slowly, this intensity is small ; when it moves rapidly, it is 
great, and necessarily so, for the same current of electricity has to travel in the two 
differing periods of time occupied by the journeys. Hence the quickly moving core 
should produce a far higher effect on the experimental helix than the slowly moving 
core ; and this also I found to be the fact. 

2683. The short copper core was adjusted to the apparatus, and the machine 
worked with its average velocity until forty journeys to and from had been completed ; 
the galvanometer needle passed 39° west. Then the machine was worked with a 
greater rapidity, also for forty journeys, and the needle passed through 80° or more 
west ; finally, being worked at a slow rate for the same number of journeys, the needle 
went through only 21° west. The extreme velocities in this experiment were pro- 
bably as 1:6; the time in the longest case was considerably less than that of one 
vibration of the needle (2651.), so that I believe all the force in the slowest case was 
collected. The needle is very little influenced by the swing or momentum of its 
parts, because of the deadening effect of the copper plate beneath it, and, except to 
return to zero, moves very little after the motion of the apparatus ceases. A silver 
core produced the same results. 

2684. These effects of induced currents have a relation to the phenomena of revul- 
sion which I formerly described (2310. 2315. 2338.), being the same in their exciting 
cause and principles of action, and so the two sets of phenomena confirm and illus- 
trate each other. That the revulsive phenomena are produced by induced currents, 
bas been shown before (2327. 2329. 2336. 2339.) ; the only difference is, that with 
them the induced currents were produced by exalting the force of a magnet placed at 
a fixed distance from the affected metal ; whilst in the present phenomena, the force 
of the magnet does not change, but its distance from the piece of metal does. 

2685. So also the same circumstances which affect the phenomena here affect the 
revulsive phenomena. A plate of metal will, as a whole, be well-revulsed ; but if it 
be divided across the course of the induced currents it is not then affected (2529.). 
A ring helix of copper wire, if the extremities be unconnected, will not exhibit the 
phenomena, but if they be connected then it presents them (2660.). 

2686. On the whole, the revulsive phenomena are a far better test and indication 
of these currents than the present effects ; especially if advantage be taken of the 
division of the mass into plates, so as to be analogous, or rather superior, in their 
action to the disc cylinder cores (2659. 2661.). Platinum, palladium and lead in leaf 
or foil, if cut or folded into squares half an inch in the side, and then packed regularly 



SUPPOSED POLARITY OF DIAMAGNETIC BODIES. 183 

together, will show the phenomena of revulsion very well ; and that according to the 
direction of the leaves, and not of the external form. Gold, silver, tin and copper 
have the revulsive effects thus greatly exalted. Antimony, as I have already shown, 
exhibits the effect well (2514. 2519.). Both it and bismuth can be made to give 
evidence of the induced currents produced in them when they are used in thin plates, 
either single or associated, although, to avoid the influence of the diamagnetic force, 
a little attention is required to the moments of making and breaking contact between 
the voltaic battery and the electro-magnet. 

2687. Copper, when thus divided into plates, had its revulsive phenomena raised 
to a degree that I had not before observed. A piece of copper foil was annealed and 
tarnished by heat, and then folded up into a small square block, half an inch in the 
side and a quarter of an inch thick, containing seventy-two folds of the metal. This 
block was suspended by a silk film as before (2248.), and whilst at an angle of 30° or 
thereabouts with the equatorial line (2252.), the electro-magnet was excited ; it im- 
mediately advanced or turned until the angle was about 45° or 50°, and then stood 
still. Upon the interruption of the electric current at the magnet the revulsion came 
on very strongly, and the block turned back again, passed the equatorial line, and 
proceeded on until it formed an angle of 50° or 60° on the other side ; but instead of 
continuing to revolve in that direction as before (2315.), it then returned on its course, 
again passed the equatorial line, and almost reached the axial position before it stood 
still. In fact, as a mass, it vibrated to and fro about the equatorial line. 

2688. This however is a simple result of the principles of action formerly developed 
(2329. 2336.). The revulsion is due to the production of induced currents in the 
suspended mass during the falling of the magnetism of the electro-magnet; and the 
effect of the action is to bring the axis of these induced currents parallel to the axis 
of force in the magnetic field. Consequently, if the time of the fall of magnetic force, 
and therefore of the currents dependent thereon, be greater than the time occupied 
by the revulsion of the copper block as far as the equatorial line, any further motion 
of it by momentum will be counteracted by a contrary force; and if this force be 
strong enough the block will return. The conducting power of the copper and its 
division into laminae, tend to set up these currents very readily and with extra power ; 
and the very power which they possess tends to make the time of a vibration so short, 
that two or even three vibrations can occur before the force of the electro-magnet has 
ceased to fall any further. The effect of time, both in the rising and falling of power, 
has been referred to on many former occasions (2170. 2650. ), and is vexy beautifully 
seen here. 



2689. Returning to the subject of the assumed polarity of bismuth, I may and 
ought to refer to an experiment made by RBrcH, and described by Weber*, which, if 

• Tayiok's Scientific Memoirs, v. p. 480. 



184 DR. FARADAY'S EXPERIMENTAL RESEARCHES IN ELECTRICITY. (SERIES XXIII.) 

I understand the instruction aright, is as follows : a strong horseshoe magnet is laid 
upon a table in such a position that the line joining its two poles is perpendicular to 
the magnetic meridian and to be considered as prolonged on one side; in that line, 
and near the magnet, is to be placed a small powerful magnetic needle, suspended by 
cocoon silk, and on the other side of it, the pole of a bar magnet, in such a position 
and so near, as exactly to counteract the effect of the horseshoe magnet, and leave 
the needle to point exactly as if both magnets were away. Then a mass of bismuth 
being placed between the poles of the horseshoe magnet is said to react upon the 
small magnet needle, causing its deflection in a particular direction, and this is sup- 
posed to indicate the polarity of the bismuth under the circumstances, as it has no 
such action when the magnets are away. A piece of iron in place of the bismuth 
produces the contrary deflection of the needle. 

2690. I have repeated this experiment most anxiously and carefully, but have never 
obtained the slightest trace of action with the bismuth. I have obtained action with 
the iron ; but in those cases the action was far less than if the iron were applied 
outside between the horseshoe magnet and the needle, or to the needle alone, the 
magnets being entirely away. On using a garnet, or a weak magnetic substance of 
any kind, I cannot find that the arrangement is at all comparable for readiness of 
indication or delicacy, with the use of a common or an astatic needle, and therefore 
I do not understand how it could become a test of the polarity of bismuth when these 
fail to show it. Still I may have made some mistake ; but neither by close reference 
to the description, nor to the principles of polar action, can I discover where. 

2691. There is an experiment which Plucker described to me, and which at first 
seems to indicate strongly the polarity of bismuth. If a bar of bismuth (or phos- 
phorus) be suspended horizontally between the poles of the electro-magnet, it will 
go to the equatorial position with a certain force, passing, as I have said, from 
stronger to weaker places of action (2267. )• If a bar of iron of the same size be fixed 
in the equatorial position a little below the plane in which the diamagnetic bar is 
moving, the latter will proceed to the equatorial position with much greater force 
than before, and this is considered as due to the circumstance, that, on the side 
where the iron has N polarity, the diamagnetic body has S polarity, and that on the 
other side the S polarity of the iron and the N polarity of the bismuth also coincide. 

2692. It is however very evident that the lines of magnetic force have been altered 
sufficiently in their intensity of direction, by the presence of the iron, to account 
fully for the increased effect. For, consider the bar as just leaving the axial position 
and going to the equatorial position ; at the moment of starting its extremities are in 
places of stronger magnetic force than before, for it cannot be doubted for a moment 
that the iron bar determines more force from pole to pole of the electro-magnet than 
if it were away. On the other hand, when it has attained the equatorial position, 
the extremities are under a much weaker magnetic force than they were subject to 
in the same places before ; for the iron bar determines downwards upon itself much 



SUPPOSED POLARITY OF DIAMAGNETIC BODIES. 185 

of that force, which, when it is not there, exists in the plane occupied by the bismuth. 
Hence, in passing through 90°, the diamagnetic is urged by a much greater difference 
of intensity of force when the iron is present than when it is away ; and hence, pro- 
bably, the whole additional result. The effect is like many others which I have re- 
ferred to in magnecrystallic action (2487-2497.), and does not, I think, add anything 
to the experimental proof of diamagnetic polarity. 

2693. Finally, I am obliged to say that I can find no experimental evidence to 
support the hypothetical view of diamagnetic polarity (2640.), either in my own 
experiments, or in the repetition of those of Weber, Reich, or others I do not 
say that such a polarity does not exist ; and I should think it possible that Weber, 
by far more delicate apparatus than mine, had obtained a trace of it, were it not that 
then also he would have certainly met with the far more powerful effects produced 
by copper, gold, silver, and the better conducting diaraagnetics. If bismuth should be 
found to give any effect, it must be checked and distinguished by reference to the 
position of the commutator, division of the mass by pulverization, influence of time, 
&c. It appears to me also, that, as the magnetic polarity conferred by iron or nickel 
in very small quantity, and in unfavourable states, is far more readily indicated by 
its effect on an astatic needle, or by pointing between the poles of a strong horseshoe 
magnet, than by any such arrangement as mine or Weber's or Reich's, so diamag- 
netic polarity would be much more easily distinguished in the same way, and that 
no indication of that polarity has as yet reached to the force and value of those 
already given by Brugmann and myself. 

2694. So, at present, the actions represented or typified by iron, by copper and 
by bismuth, remain distinct ; and their relations are only in part made known to us. 
It cannot be doubted that a larger and simpler law of action than any we are yet 
acquainted with, will hereafter be discovered, which shall include all these actions at 
once ; and the beauty of Weber's suggestion in this respect was the chief induce- 
ment to me to endeavour to establish it. 

2695. Though from the considerations above expressed (2693.) I had little hopes 
of any useful results, yet I thought it right to submit certain magnecrystallic cores 
to the action of the apparatus. One core was a large group of symmetrically 
disposed crystals of bismuth (2457.) ; another a very large crystal of red ferroprus- 
siate of potassa; a third a crystal of calcareous spar; and a fourth and fifth large 
crystals of protosulphate of iron. These were formed into cylinders of which the 
first and fourth had the magnecrystallic axes (2479.) parallel to the axis of the 
cylinder, and the second, third and fifth, had the equatorial direction of force (2594. 
2595. 2596.) parallel to the axis of the cylinder. None of them gave any effect at 
the galvanometer, except the fourth and fifth, and these were alike in their results, 
and were dependent for them on their ordinary magnetic property. 

2696. Some of the expressions I have used may seem to imply, that, when employing 
the copper and other cores, I imagine that currents are first induced in them by the 

mdcccl. 2 B 



186 DR. FARADAY'S EXPERIMENTAL RESEARCHES IN ELECTRICITY. (SERIES XXIII.) 

dominant magnet, and that these induce the currents which are observed in the ex- 
perimental helix. Whether the cores act directly on the experimental helix or indi- 
rectly through their influence on the dominant magnet, is a very interesting question, 
and I have found it difficult to select expressions, though I wished to do so, which 
should not in some degree prejudge that question. It seems to me probable, that 
the cores act indirectly on the helix, and that their immediate action is altogether 
directed towards the dominant magnet, which, whether they consist of magnetic or 
diamagnetic metals, I'aises them into power either permanently or transiently, and 
has their power for that time directed towards it. Before the core moves to approach 
the magnet, the magnet and experimental helix are in close relation ; and the latter is 
situated in the intense field of magnetic force which belongs to the pole of the former. 
If the core be iron, as it approaches the magnet it causes a strong convergence and 
concentration of the lines of magnetic force upon itself; and these, as they so con- 
verge, passing through the helix and across its convolutions, are competent to produce 
the currents in it which are obtained (2653. 2668.). As the iron retreats these lines of 
force diverge, and again crossing the line of the wire in the helix in a contrary direc- 
tion to their former course, produce a contrary current. It does not seem necessary, 
in viewing the action of the iron core, to suppose any direct action of it on the helix, 
or any other action than this which it exerts upon the lines of force of the magnet. 
In such a case its action upon the helix would be indirect. 

2697- Then, by all parity of reasoning, when a copper core enters the helix its ac- 
tion upon it should be indirect also. For the currents which are produced in it are 
caused by the direct influence of the magnet, and must react equivalently upon it. 
This they do, and because of their direction and known action, they will cause the 
lines of force of the magnet to diverge. As the core diminishes in its velocity of mo- 
tion, or comes to rest, the currents in it will cease, and then the lines of force will 
converge; and this divergence and convergence, or passage in two directions across 
the wire of the experimental helix, is sufficient to produce the two currents which 
are obtained in the advance of the core towards the dominant magnet (2671. 2673.). 
A corresponding effect in the contrary direction is produced by the retreat of the 
core. 

2698. On the idea that the actions of the core were not of this kind, but more 
directly upon the helix, I interposed substances between the core and the helix during 
the times of the experiment. A thick copper cylinder 2*2 inches long, 07 of an inch 
external diameter, and 0*1 of an inch internal diameter, and consequently 0*3 of an 
inch thick in the sides, was placed in the experimental helix, and an iron wire core 
(2668.) used in the apparatus. Still, whatever the form of the experiment, the kind 
and amount of effect produced were the same as if the copper were away, and either 
glass or air in its place. When the dominant magnet was removed and the wire 
core made a magnet, the same results were produced. 

2699. Another copper lining, being a cylinder 2*5 inches long, 1 inch in external 



SUPPOSED POLARITY OF DIAMAGNETIC BODIES. 187 

diameter, and one-eighth of an inch in thickness, was placed in the experimental 
helix, and cores of silver and copper five-eighths of an inch in thickness, employed as 
before, with the best condition of the commutator (2675.) : the effects, with and 
without the copper, or with and without the glass, were absolutely the same (2698.). 

2700. There can be no doubt that the copper linings, when in place, were full of 
currents at the time of action, and that when away no such currents would exist in 
the air or glass replacing them. There is also full reason to admit, that the 
divergence and convergence of the magnetic lines of force supposed above (2697.) 
would satisfactorily account for such currents in them, supposing the indirect action 
of the cores were assumed. If that supposition be rejected, then it seems to me that 
the whole of the bodies present, the magnet, the helix, the core, the copper lining, or 
the air or glass which replaces it, must all be in a state of tension, each part acting 
on every other part, being in what I have occasionally elsewhere imagined as the 
electro-tonic state (1729.). 

2701. The advance of the copper makes the lines of magnetic force diverge, or, so 
to say, drives them before it (2697-)> No doubt there is reaction upon the advancing 
copper, and the production of currents in it in such a direction as makes them com- 
petent, if continued, to continue the divergence. But it does not seem logical to say, 
that the currents which the lines of force cause in the copper, are the cause of the 
divergence of the lines of force. It seems to me, rather, that the lines of force are, so 
to say, diverged, or bent outward by the advancing copper (or by a connected wire 
moving across lines of force in any other form of the experiments), and that the re- 
action of the lines of force upon the forces in the particles of the copper cause them 
to be resolved into a current, by which the resistance is discharged and removed, and 
the line of force returns to its place. I attach no other meaning to the words line 
of force than that which I have given on a former occasion (2149.). 



Royal Institution, 
14 Dec. J 849. 



2 b2 



188 DR. FARADAY'S EXPERIMENTAL RESEARCHES IN ELECTRICITY. (SERIES XXIII.) 



Plan of the Apparatus employed in Dr. Faraday's Researches, Series XXIII. p. 172. 




a, b, c frame board ; d, d, d wooden lever, of which e is the axis, / the crank-wheel, and g the 
great wheel with its handle h ; i the bar connecting the crank- wheel and lever ; k a cylinder or core 
of metal to be submitted to experiment ; I the rod connecting it with the lever ; m the helix of the 
electro-magnet ; n the iron core, and o the exciting battery ; p the experimental helix ; q the galvano- 
meter, 20 feet from the electro-magnet ; r the commutator ; w, w connecting wires ; s, s springs of 
brass or copper ; t a copper rod connecting the two arms of the lever, to give strength. The plan is 
to a scale of one-fifteenth : the part at the electro-magnet and experimental helix is in section ; the 
further description is in paragraphs 2643, 2644, 2645 and 2648 of the Experimental Researches.