Dismiss Notice
Join Physics Forums Today!
The friendliest, high quality science and math community on the planet! Everyone who loves science is here!

Formulae of the first edition of the Maxwell's Treatise

  1. Jan 22, 2005 #1
    Dear Friends

    I'm looking for the formulae of the first edition of the Maxwell's Treatise, of the 1873 year, with 20 equations in 20 unknowns (they are explicitly listed in the paper). It seems that the equations taught today in universities as ‘Maxwell's equations' are actually Heaviside's equations.

    Has anybody this 20 formulae?
  2. jcsd
  3. Jan 22, 2005 #2


    User Avatar
    Science Advisor
    Homework Helper

    Did u check the closest University library??I doubt u can find the original book of Maxwell (scanned,of course) anywhere on the internet...

    Who said about Heaviside??

    Bring proof...

  4. Jan 22, 2005 #3


    User Avatar

    Staff: Mentor

    I seem to remember reading that Heaviside was a major promoter of the use of our current vector notation in E&M. Before that, most people wrote out the individual equations for the x, y and z components. Certainly that would have been true in Maxwell's day. Modern vector notation wasn't even invented until around 1900, I think.

    So... the 20 equations Raparicio is referring to, must be related to the individual components of what we now call Maxwell's Equations.

    Gauss's Law for E and for B each give one equation, because the divergence is a scalar quantity. The two equations with the curl of E and of B, each give three equations. That's a total of eight so far.

    If we add the relations between E and D, and between B and H, that's six more, for a total of fourteen.

    And then there's the continuity equation for current density and charge density. But that's only one more equation, because the divergence of J is a scalar. That's fifteen equations so far. Where could the other five come from?
  5. Jan 22, 2005 #4
  6. Jan 22, 2005 #5


    User Avatar
    Science Advisor
    Homework Helper

    IIRC,the FUNDAMENTAL PROBLEM OF CLASSICAL LINEAR ELECTROMAGNETISM involves solving 22 diff.eqns.with 22 unknowns:

    [tex] \sigma,\overline{\overline{\chi}}_{el},\overline{\overline{\chi}}_{mag},\epsilon_{0},\mu_{0} [/tex]
    all presumably space and time dependent...
    [tex] \vec{E}(\vec{r},t),\vec{D}(\vec{r},t),\vec{P}(\vec{r},t),\vec{H}(\vec{r},t),\vec{B}(\vec{r},t),\vec{M}(\vec{r},t),\rho(\vec{r},t),\vec{j}(\vec{r},t) [/tex]

    I challenge u to come up with the 22 equations...

    NOTE:Double bar on top is a notation for 2-nd rank cartesian tensors.U can denote the contracted tensor product by ":".

    Last edited: Jan 22, 2005
  7. Jan 22, 2005 #6
    I think that first Maxwell's formulae were writted in "quaternions", but the Maxwell's editor thinked that were very difficult to understand. Heaviside changed it to vector formulae. I have curiosity with original formulation, truncated also with Lorenz gauge, i readed.

    In my university there's no oportunity to take this book. The only version is the revised version, with this ecuations in "vectors" not in original version.
  8. Jan 22, 2005 #7


    User Avatar
    Science Advisor
    Homework Helper

    Forget history for a second,and think physics,modern physics:Are u sure 100% that there are 22 equations with 22 unknowns??

  9. Jan 22, 2005 #8


    User Avatar
    Staff Emeritus
    Gold Member
    Dearly Missed

    Here you can see the original twenty equation in twenty unknowns, together with the modern interpretation of the variables and the mapping into the vector form of the equations.

    I found this site by entering maxwell original equations in google. It was about the fourth down in the resulting list. Finding things for yourself with google is a lot more rewarding than just asking someone else,
  10. Jan 22, 2005 #9

    Thanks, selfadjoint. So much times, the problem is to find the adecuate "keys" in google.

    And I tried (without success) tu find in google, writting "maxwell treatise"

  11. Apr 7, 2006 #10
    The Original Maxwell's Equations

    Maxwell, actually, was not using vectors so much as differential forms and he did (in fact) make the underlying Grassmann algebra explict at one point in his treatise. He wrote much of chapter 1 on the same topics covered by more modern treatments of differential forms and integration thereof (even touching into basic homology theory with the discussion about cyclomatic number, cyclosis, or whatever he called it).

    The fields (in more modern notation) were
    D = (D^x dy dz + D^y dz dx + D^z dx dy)
    H = (H_x dx + H_y dy + H_z dz)
    E = (E_x dx + E_y dy + E_z dz)
    B = (B^x dy dz + B^y dz dx + B^z dx dy),
    the potentials
    A = (A_x dx + A_y dy + A_z dz)
    the velocity, G, relative to the distinguished frame of reference, and
    the sources
    J = (J^x dy dz + J^y dz dx + J^z dx dy)
    rho dx dy dz
    and the "total" current C = J + dD/dt. There was no decomposition D = epsilon_0 E + P, with a separate polarization vector because a central thesis of Maxwell was that the vacuum, itself, was a dielectric medium with the ability to engender charge screening (thus effecting a distinction between bare vs. dressed charges -- a point he touched on), vacuum polarization (which he touched on in his thought experiment explaining the impossibility of point-like and line-like sources), and dielectric breakdown producing electromagnetic energy (the section on the "glow" phenomenon) which -- in the other context -- would have meant also the ability of the vacuum, itself, to break down and produce electromagnetic radiation (i.e. matter/anti-matter annihilation).

    The distinguished frame of reference is that where electromagnetic radiation proceeds in a sphere centered on a fixed point. There was no explicit discussion of which frame might be so distinguished and what the earth's motion relative to it was, because the section on the measurement and values of light speed only had the speed resolved to 100,000 km/sec, which is not precise enough for the Earth's motion to become an issue.

    The equations posed by are NOT equivalent to the modern equations. In particular, for the electric field, he had
    E = -dA/dt - grad(phi) + G x B.
    If you take the *modern* definition (E = -dA/dt - grad(phi)), what this amounts to is that Maxwell's constitutive law (D = K E) becomes in modern notation
    D = epsilon (E + G x B).
    This is the constitutive law that is uniquely identified by the requirement that the remaining part of Maxwell's equations be invariant under the Galilean transformations that define Newtonian physics. The other constitutive law (not explicitly mentioned) would then have to be B = mu (H - G x D). There was a brief mention implying the -GxD term in a footnote to a later edition of the treatise, however.

    The 4 macroscopic equations were present (div D = rho; div B = 0; curl H = C = J + dD/dt; curl E - dB/dt = 0). Maxwell wasn't clear on how the last equation meshed with the relation (E = ... + G x B). He never did a comprehensive analysis, in fact, to determine how his entire system transformed under a change in reference frame. That was the deficit Einstein picked up on and resolved.

    Ohm's law (J = sigma E) was included, as well as the continuity equation (d(rho)/dt + div J = 0), and the derivation of B from the potential (B = curl A). The continuity law is deriveable and Maxwell didn't realise that early on, so the earliest form he posed for a system was actually UNDER-determined; not perfectly matched with 20 variables and 20 independent equations. The other constitutive law (B = mu H) was listed with several variants, and no clear, definitive, statement was made on what it ought to be. That's partly linked to the point made above about Maxwell not having done a clear analysis on his system to check their Galilean invariance. Had he done so, he would have clearly seen that the only relation possible is B = mu (H - G x D) and that, right there, might have raised a few eyebrows, implying as it does, a dependence of the B field on the electric displacement in frames that are in motion relative to the distinguished frame.

    The Lorentz relations (D = epsilon_0 E, B = mu_0 H) therefore represent a completely INEQUIVALENT and fundamental departure from Maxwell's Galilean invariant (D = epsilon_0 (E + G x B), B = mu_0 (H - G x D)). As soon as you write down the former -- whether you realise it or not -- you're in Relativity, for they are only invariant under the LORENTZ transformations, not the Galilean transformations.

    So, a major historical myth is also debunked here: Einstein did not get rid of the Aether from an "otherwise-equivalent" Aether theory; he (and Lorentz, already) posed a system of constitutive relations that were fundamentally different from Maxwell's system and never had an Aether frame in it. It's Maxwell's consitutive laws that had the G in it, not Lorentz's. The Aether version (with Maxwell's G) is NOT equivalent to the Aether-less version (with Lorentz's relations with the G). In the latter, G would be 0 in every frame of reference. In the former, G would change. They are completely different, empirically, and have different empirical consequences.
  12. Apr 9, 2006 #11
Share this great discussion with others via Reddit, Google+, Twitter, or Facebook