Unopiolar X 4

No^! 1 '] ELECTROMAGNETIC INDUCTION.
323
ON ELECTROMAGNETIC INDUCTION.
BY S. J. BARNETT.
1. In a recent article 1 on electromagnetic induction and relative

motion published in this journal, and in a preliminary paper2 published
in the Physikalische Zeitschrift, I referred incidentally and briefly, and
of necessity adversely, to some work on unipolar induction 3 by Mr. E. H.
Kennard, who has now published three articles4 in reply, the last in the
May number of this journal. Although, in my opinion, Mr. Kennard's
fallacies are quite apparent to anyone familiar with this part of elec-
trical theory, it seemed advisable to make a reply5 to the first of these
articles on its appearance; and a final reply is now necessary.
2. In each of the three articles Mr. Kennard either states or implies
that my chief criticism of his work is on account of his treatment of
the dielectric on the moving line hypothesis. My chief criticism of Mr.
Kennard's work has always been that even if his moving line theory were
absolutely correct when applied to the ordinary case of unipolar induc-
tion in which the 'complete field-producing agent rotates, his sweeping
conclusion from his experiments would be entirely without justification.
For in his experiments the iron core of the electromagnet rotated alone
while the magnetizing coil remained fixed like the condenser; yet his
calculation of the effect to be expected on the moving line hypothesis
assumes that on this hypothesis all the lines in his experiments would
move with the iron. But it is quite impossible to say what fraction of
the lines, on this hypothesis, would adhere to the iron and move, and
what fraction would adhere to the coil and remain at rest. For the
same reason, viz., that a part of the field-producing system was at rest,
instead of all being in motion, the experiments are without value in
connection with the matter of relativity.
3. In treating the problem of unipolar induction in a fixed dielectric
1
P H Y S I C A L R E V I E W , 35, 1912, p. 323-
2 P h y s . Zeit., 13, 1912, p . 803.
3
Phil. Mag., 23, 1912, p . 937.
4
P h y s . Zeit., 13, 1912, p . 1155, a n d 14, 1913, p . 250; PHYSICAL R E V I E W (2), 1, 1913,
p.355-
5
P h y s . Zeit., 14, 1913, p . 251.
[SECOND
324 S. J, BARNETT. LSERIES.
Poincdre, 1 Abraham, and myself have each assumed that throughout

the field (points within the rotating magnet not being included) the
electric density is proportional to the divergence of the total electric
intensity. Mr. Kennard says that this was the assumption made by
Hertz, 2 but that it is not always allowable on the theory of Lorentz.
These statements are both true. But it is also true and well known that
in fields such as that under discussion, in which the intensities are steady
and the ether and matter at rest, Lorentz}s theory and Hertz's theory
and Maxwell's theory are completely identical so far as the total intensity,
displacement, and density are concerned; and that in this case the
assumption referred to is not only allowable but necessary. Mr. Ken-
nard says that this assumption " may easily be shown to be incompatible
with the moving line theory," and attempts to prove his assertion as
follows: He writes the equation
div / = div (E + e) (i)
in which E, e, and / are the field, motional, and total intensities, and
applies the equation to a point in a dielectric of free ether at which the
electric density, and therefore div / , according to Poincare, Abraham,
and myself, are zero. This gives
div e = — div E. (2)
Mr. Kennard then assumes, without justification, that div E = o. Then
he calculates div e and finds that it does not vanish—a particular case of
a well-known relation to which he has repeatedly called -attention.
Hence he concludes that (1) and (2) cannot be correct, and therefore
that I and the others are wrong. The trouble, however, is only with the
false assumption div E = o.
4. To come to the next point, which I did not realize that my treat-
ment could have left in obscurity, I have certainly assumed that moving
lines of induction act on the ether precisely as on a material dielectric?
1
Poincare assigns to the dielectrics, magnet, and other conductors velocities which may
vanish in particular cases. His treatment of the subject on Lorentz's theory is not correct
for the case in which the velocities of the dielectrics differ from zero; and his conclusion that
so far as experiment can show "the theory of Lorentz leads to the same results as that of
Hertz " in the general case is erroneous. When the dielectrics move the two theories give
quite different results; but on each theory the result is independent of the hypothesis adopted
with reference to the seat of the electromotive force.
2
Mr. Kennard's statement that Hertz "rejected the motional intensity" cannot be
accepted.
3
This is also an implicit assumption in the work of Poincare and Abraham. For in the
case under discussion the total intensity / differs from the field intensity E (whenever there is
a difference) only by the motional intensity e. As another discussion of unipolar induction
manifestly based on the same assumption should be mentioned that of S. Valentiner (Phys.
Zeit., 6, 1905, p. 10), which, as far as it goes, is essentially identical with my own.
No"* 1 '] ELECTROMAGNETIC INDUCTION, 325
Mr. Kennard says that this is not consistent with current electrical
theory as he knows it. It is nevertheless old and sound doctrine, uni-
versally accepted since the days of Maxwell.1 It is involved in the funda-
mental relation between the electromotive force around a closed circuit
in the ether at rest and the rate of change of magnetic flux through the
circuit—a change produced, according to Maxwell, by the motion of the
lines of induction across the circuit, since the lines of induction are closed
curves—coupled with the relation between intensity and displacement.
Among many obvious examples, one whose mere mention should be
sufficient is the case in which the energy of a magnet set into motion is
transformed into electromagnetic energy and transferred to an adjacent
electric circuit in accordance with Poynting's theorem. As to Mr.
Kennard's ideas on the diametrical opposition of ether to matter, etc.,
no remarks seem necessary.
5. The calculation of the charge on the condenser in my experiments
to be expected on the moving line hypothesis when the field-producing
agent rotates and the condenser remains fixed (Case II.) Mr. Kennard
proceeds to make by assuming that the charge would be the same as
when the condenser rotates and the field-producing agent remains at
rest (Case I.). His reason for this appears to be the false assumption
that the moving lines would have no effect on the ether. Had Mr.
Kennard taken proper account of the ether, he would have obtained the
charge zero, exactly as if the lines had been assumed to remain fixed.
Because his calculation gave a charge independent of the magnitude of
the dielectric constant, he stated in a previous paper, referring to my
criticism that he had not taken proper account of the dielectric, that the
trouble could not be with his treatment of the dielectric. But the
neglected ether is an important part of the dielectric.
6. In his concluding paragraph Mr. Kennard says that he has elsewhere
called attention to the fact that my conclusion as to relative motion rests
in part on an inference.
Now it follows immediately and necessarily from the experiments of
Faraday, Lorentz, Rayleigh and others on the motion of conductors in
magnetic fields, together with the experiments of Blondlot, H. A. Wilson,
and myself on the motion of insulators in magnetic fields, that if the con-
denser, whatever the magnitude of the constant of its dielectric, rotates
while the agent producing the magnetic field remains fixed (Case I.),
the condenser receives a charge equal to the product of the capacity
as it would be with ether alone as dielectric by the rate at which the
1
See, for example, Heaviside's Electromagnetic Theory, I., § 48; Lodge's Modern Views of
Electricity, p. viii and §§ 114, 115; and S. Valentiner, loc. cit.
{"SECOND
326 S. J. BARNETT, L SERIES.
short-circuiting wire cuts across magnetic flux; and that if the material
part of the dielectric is air, as in my own recent experiments, it is of no
consequence whether the air moves with the conductors or not.
In the first of my original papers I referred only to the fact of the con-
denser's becoming charged; and in the second I referred, in addition,
only to the "experiments of Faraday and others'' as establishing the fact.
Probably I should have gone into greater detail. This I have done in the,
Physikalische Zeitschrift, 14, 1913, p. 251; and in Science, January 17
and February 21, 1913.
The brief statement of fact, without reference to authority, in my first
paper drew from Mr. Kennard the criticism to which he refers. He said
that I had failed to give experimental proof of my statement. References
to the experimental work having been given later, however, Mr. Kennard
now objects to the experiments of Blondlot, Wilson, and myself on insula-
tors on account of sliding contacts, stationary connecting wires, absence
of a conducting screen, etc. These objections are entirely inconsequential
and irrelevant and I shall not consider them further. No one will object
to the repetition of any or all of these experiments, either modified or
unmodified, by anyone who is sufficiently interested in them; but it is
quite certain what the results will be.
7. In conclusion it seems desirable to consider briefly what happens
on the theories of Hertz, Einstein, and Lorentz in each of the two principal
experiments involved in the discussion of relative motion. As above, the
case in which the condenser rotates will be referred to as Case I.; that in
which the condenser remains at rest, as Case II.
On the theory of Hertz, the condenser is uncharged in Case I. and also
in Case II.
On the theory of Einstein, the condenser is charged as indicated in § 6
in Case L, and charged in the same way in Case II.
On the theory of Lorentz, the condenser is charged as indicated in § 6
in Case I., and is uncharged in Case II.
Lorentz's theory is thus the only one which is consistent with both sets
of experiments.
T H E OHIO STATE UNIVERSITY,
June 7, 1913.

Unopiolar X 4

Uploaded by

Copyright:

Available Formats

Unopiolar X 4

Uploaded by

Document Information

Original Title

Copyright

Available Formats

Share this document

Share or Embed Document

Sharing Options

Did you find this document useful?

Is this content inappropriate?

Copyright:

Available Formats

Unopiolar X 4

Uploaded by

Copyright:

Available Formats

No^! 1 '] ELECTROMAGNETIC INDUCTION.

1. In a recent article 1 on electromagnetic induction and relative

Poincdre, 1 Abraham, and myself have each assumed that throughout

You might also like