CHAPTER 3 - FIVE-DIMENSIONAL SYSTEMS

During the preceding development of the Dynamic Theory, there did

not appear to be anything that approached a description of nuclear effects.
Of course quantum theorists may respond that the nuclear effects lie
within the realm of quantum theory. This, however, does not seem to be a
strong argument since current nuclear theory appears to depend upon a
number of ad hoc postulates.
If it is supposed that nuclear theory cannot be extracted from some
aspect of the preceding four-dimensional world view, then how might the
Dynamic Theory produce a foundation for nuclear theory? At this point
there may appear to be no obvious way. Therefore, let us proceed on a
different tack.
Thus far we have constantly adhered to the policy of dividing
systems into two types: thermodynamic systems with only a work term of
the pdv type and mechanical systems with three mechanical, or spatial,
work terms. The generality of the adopted laws places no restrictions upon
the number or type of variables used. In particular, there is no restriction
coming from the laws themselves which says we cannot use four work
terms, one the thermodynamic pdv term and three mechanical Fdq terms.
Obviously pdv itself is just another Fdq type term with the pressure as the
generalized force and the volume as the generalized displacement.
When we teach thermodynamics we write the First Law with the
right hand side of the equation being the change in internal energy (system
energy), the thermodynamic work term, and three spatial work terms. We
then tell the students that since the right hand side of the equation
involves five unknowns we must have five independent equations in order
to have a solvable system. The first equation offered is the conservation of
mass which we state guarantees that we may write mass density as a
function of space and time. But is this really true for all space and for all
time?
The rub comes in attempting to visualize a world description in five
dimensions. Many arguments may be envisioned which tend to imply only
a four-dimensional manifold is needed. The kinetic theory of gases relates
the pressure to the average velocities of the particles contained. Does that
not imply that thermodynamics ultimately rests on a four-dimensional
manifold? Recall that the system in the kinetic theory is basically in
equilibrium.
Statistical thermodynamists may claim that thermodynamics is
basically statistical in nature and is fundamentally tied to order and
disorder and hence to the four-dimensional world of quantum theory. But
remember that the overall system, to which the statistical approach is
applicable, is a composite system made up of many subsystems each in an
equilibrium state. What happens to this argument when the number of
individual particles is not infinitely large?

90
 Still there seems to be no substantial support for a five-dimensional
world from the point of view of current theories. This is to be expected
though in view of the difficulties experienced in the transition from the
classical three-dimensional world to the four-dimensional space-time of
Einstein's theories. Obviously had the extension of the universe been
restricted on a priori grounds to three-dimensional Euclidean space,
Einstein's theory would have been rejected on first principles. On the
other hand, as soon as we recognize that the fundamental continuum of
the universe and its geometry cannot be posited a priori and can only be
disclosed to us from place to place by experiment and measurement, a vast
number of possibilities are thrown open. Among these the
four-dimensional space-time of relativity, with its varying degrees of
non-Euclideanism, has found a place. So also may the five-dimensional
view of the Dynamic Theory be found within the possibilities. Ultimate
judgment upon its necessity, or applicability, should rest upon a
comparison of the theory's predictions with reality.

A. Systems Near an Equilibrium State

The metric coefficients are made up of the second partial derivatives

of the system energy function and, therefore, if the system remains near an
equilibrium state, then the value of these derivatives evaluated at the
equilibrium state may be used as a first approximation for the metric
coefficients. In this case the geometry will be Euclidean and, from the
preceding four-dimensional development, the Euclidean manifold produced
by applying the E-conservative restriction was Minkowski's space-time
continuum of Special Relativity.
Therefore, suppose we begin an investigation of the five-dimensional
world by staying very near an equilibrium state so as to simplify the
description to a five-dimensional generalization of Minkowski's space-time
manifold.

3.1 Equations of Motion

Suppose that we consider some sort of system requiring four work

terms and for the moment not concern ourselves as to exactly what this
system might be. Thus, for our system we will have thermodynamic as
well as mechanical variables and the First Law becomes
~ ~
_E = dU + Pdv - F~α dqα ; α = 1,2,3.

Where the v and are considered as specific quantities. That is, these
quantities are related to a unit of mass such as is customary in
thermodynamics.
The specific volume is the reciprocal of the mass density γ. Using the
mass density instead of the specific volume the First Law becomes
91
 ~ ~ P
d E = dU -  2 dγ - F~α dqα ; α = 1,2,3.
γ 

This law now requires that the system's specific energy U be a function of
five independent variables so that U = U(S,q1,q2,q3). Thus, the First Law
requires a five-dimensional manifold of specific entropy, space, and mass
density for a general system. Since the system under consideration needs
both thermodynamic and mechanical variables, we can no longer refer to
the entropy as mechanical or thermodynamic; however, the limiting case
where the mass is held fixed must produce the mechanical entropy.
The procedure established by the Dynamic Theory is to take the
stability condition quadratic form as the metric for a stable system. Thus,
the coefficients of the metric become the second partial derivatives of the
energy function. In order to simplify the metric, suppose for the present
that we restrict our system to be very near an equilibrium state so that we
may consider the second partial derivatives to be constants. This is in
essence considering a local Euclidean manifold; the symmetry of the
geometric connections guarantees that we may do this.
Since the metric coefficients are constants, a transformation may be
found such that the cross terms are zero. Then in this coordinate system
and when

0 S γ
q ≡ and q 4 ≡ ,
F0 a0

the metric becomes

α α
2 0 2 4 2
c (dt ) = ( dq ) + dq dq + ( dq ) ; α = 1,2,3.
2
(3.1)

If we again consider the restriction d = 0 so that we are talking of an

E-conservative system for which the principle of increasing entropy holds,
then we have the variational principle given by

δ ∫ (dS )2 = 0. (3.2)

Solving Eqn. (3.1) for dq0 and squaring we get

( dq0 )2 = c 2 (dt )2 - dqα dqα - ( dq 4 )2 ; α = 1,2,3, (3.3)

or
2
 dq0   α  β 
  = c 2 - g αβ  dq  dq  ; α , β = 1,2,3,4.
 dt   dt  dt 
    

The entropy manifold given by Eqn. (3.3) is a five-dimensional

Minkowski-type manifold with coordinates of space-time-mass. We may,

92
therefore, follow the procedure Minkowski and Einstein used in the Special
Theory of Relativity.
First, to avoid confusion, let us rename the coordinates as
1 2 3 4
x ≡ ct, x ≡ q , x ≡ q , x ≡ q , and x ≡ q .
0 1 2 3 4

Then define the five-dimensional velocity vector as

i
dx
u ≡
i
0
; i = 0,1,2,3,4
dq

and define the five-dimensional acceleration vector as

δ ui d x  i  dx dx
2 i j k
i
f ≡ ≡ +  jk  dq0 dq0 .
δ q0 dq
02
 

Now the specific entropy is the arc length and the variational
principle is based upon the entropy. Therefore, if we multiply the specific
entropy by the mass density, we have the entropy density. The variational
problem becomes

δ ∫ γ 2 ( dq0 )2 = δ ∫ γ ( dq0 )2 = 0. (3.4)

Notice how the mass has entered our variational problem. It has entered
because our metric was in terms of the "specific entropy", or entropy per
unit mass. The variational problem is based upon the entropy, not the
specific entropy. Thus, the mass density is required in the variational
problem to correct this difference. The importance of this lies in the fact
that this is the origin of the "inertia" which appears in the following
equations of motion.
The Euler equations for this problem are
 ∂ g ik i j 
 j 
d  γ g ij u  ∂γ γ ∂ k u u 
- g ij u u - 
i j x =0
dq  g ij u i u j  ∂ x k
0
 g ij u i u j 
   

or
 ∂ g ik i j 
 g j    g j  i u u 
u ∂γ  d u - ∂x 
a u  - 
0 4 ij i j ij
g u u +   = 0.
 g u i u j  ∂ xi ij
 dq
0
 i j 
g ij u u i j
g ij u u 
 ij     

Using the fact that gijuiuj = 1, the Euler equations become

93
 ∂γ
γ fi = - a0 u 4 g ij u j ≡ F i (3.5)
∂ xi

where the Fi are force densities.

Obviously if we hold the mass density fixed, u4=0, then the volume
integral of this equation becomes the force-mass- acceleration relationship
of Special Relativity.
Now Since
2
δ ui  dq0 
i
 
 dt  = c - u u ; α = 1,2,3,4
2 α α
f = and
δ q0  

then

δ ui δ u i dt
F =γ =γ
i

δq 0
δt dq0
γ δ  dxi 
=  0  where v 2 = uα uα : α = 1,2,3,4.
 
c - v δt  dq 
2 2

Then

γ δ  1 dx 
α
α
F = 
2 2 δt  2 - 2 dt 
c -v  c v 
γ δ  1 α 
dx 
= c2  ,
1 - β 2 δt  1 - β 2 dt 

where ß = v/c with v being the four-dimensional speed.

The force density equation may now be written as

γ δ  1 α 
1 - β 2 Fα =  dx 
.
c δt
2  2 dt 
 1 - β 

Consider

δ  γ α
dx  δγ  1 α
dx  δ  1 α
dx 
= +γ ;
δt  1 - β 2 dt  dt  1 - β 2 dt  δt  1 - β 2 dt 
     

but δγ/δt=a0v4, so that the force density equations may be written as

94
 1 δ  γ α 4 
dx  a0 u  1
α
dx 
1 - β 2 Fα = - .
c δt  1 - β 2 dt  c  1 - β 2 dt 
2 2
   

We may define

1 γ
γ ≡
1- β 2

as the effective mass density or "relativistic" mass density; then

1 δ  1 dxα  a0 v 4 vα
1 - β 2 Fα = γ - .
c δt
2
 dt  c 2 1 - β 2

By defining F α ≡ c 2 1 - β 2 Fˆ α 23 so that

α δ  1 dxα  a0 v 4 vα
Fˆ = γ dt  - (3.6)
δt   1- β 2

we see that this force density becomes Einstein's special relativistic force
density when v4 = 0, or for constant "rest mass." Thus, the equations of
motion, Eqn.s (3.6), reduce to Einstein's special relativistic equations of
motion when dγ/dt=0.

3.2 Energy Equation

Now for our system the restriction that

~ ~ P
_E = 0 = dU - 2 dγ - F~α dxα , α = 1,2,3
γ

requires that
~ p
dU = 2 dγ + F~α dxα , α = 1,2,3,
γ

or if p/γ2 is considered as another generalized force density, then

~
dU = F~α dxα , α = 1,2,3,4.

Thus, by integrating the expression for the system's specific energy

change, we should arrive at the Einstein energy equation if we hold dx4/dt
/ 0. Therefore, we shall perform the integration using the force densities
given by Eqn. (3.6) to get the system's energy, or
95
   α 4 α
~ d γ dx  a0 u u  α
U - U~ 0 = ∫ pp0 F~α dxα = ∫ pp0   -  dx
 dt  1 - β dt  1 - β 
2 2

  α 
α du 
    u 
 d
= ∫ tt0 γ  
1  α α
+ dt  dt.
 u u 
  dt  1 - β  1 - β 2 
2

   

But c2ß2 = uαuα and c2ß(d/dt) = uα duα/dt; therefore,

   
~  d 1  2 2 c 2 ββ&  
U - U~ 0 = ∫ tt0 γ   c β + dt
2 
  dt  1 - β  1 - β 2  
 
βdβ
= c 2 ∫ ββ 0 γ 3
.
2 2
(1 - β )

Now ß depends upon u and not upon x4 or γ; therefore

~ γ c2
U - U~ 0 =
1- β 2

or
~ γ c2
U = + constant.
1- β 2

If the internal energy is considered to be the system's energy when

the spatial velocities uα; α = 1, 2, 3 are taken as zero, then the internal
energy density given by
~ γ c2
U = + constant.
 4
1 -  u 
 c

At the condition where u4 is also zero the internal energy density is then
~
U = γ c 2 + constant.

By taking the constant of integration to be zero, this internal energy

density then is seen to correspond to Einstein's "rest energy" where here
the "rest energy" is in terms of a four-dimensional "at rest" state.
If we make the usual approximation of allowing ß2<<1, then the
system's energy density is approximately given by

96
 ~ 1 1 γ
U = γ c2 + γ v2 + ( γ&)2 ,
2 2 ( a0 )2

where here u4 = dγ/dt is used. This displays the classical limit system
energy density for an E-conservative system very near equilibrium.

B. Systems With Non-Euclidean Manifold

Suppose now we relax the assumption that the system is very near
an equilibrium point so that the second partial derivatives are no longer
constants but are functions. This is essentially the same transition as
Einstein made going from his Special to General theory; however, the logic
of the transition is much simpler here. The only change in the logic
appears in the relaxation of the assumption of nearness. There is, of
course, a drastic increase in mathematical difficulty since the metric com-
ponents are no longer constants.

3.3 General Variational Principle

We shall consider a system that must be described by both

thermodynamic and mechanical variables. When written in terms of the
mass density, the First Law for this system may be written as
~ ~ P
_E = dU - 2 dγ - F~α dqα , α = 1,2,3,
γ

where the tilde denotes specific quantities.

Following the prescribed procedures of the Dynamic Theory we shall
take the stability condition quadratic form as the metric for our system.
Thus, the metric coefficients will be given by the second partial derivatives
2
∂ U
hij = , i, j = 0,1,2,3,4,
∂ qi ∂ q j

where q4=γ/a0. The metric may then be written as

α α β
2 2 0 2 0
c (dt ) = h00 ( dq ) + 2 h0α dq dq + hαβ dq dq , (3.3A)

where α,ß=1,2,3,4.
Imposing the restriction that the system be E-conservative, πE=0,
results in the principle of increasing entropy, so that

δ ∫ (dS )2 = 0.

In terms of the specific entropy the variational principle may be

written as

97
 δ ∫ ( γ dq0 )2 = δ ∫ γ ( dq0 )2 = 0. (3.4)

Solving the metric given by Eqn. (3.3A) and squaring yields the
expression
 1 
( dq0 )2 =  { c 2 (dt )2 + 2 h0α Adtdqα - hαβ dqα dq β }, α , β = 1,2,3,4, (3.7)
 h00 

with
η 2
c hην &η &ν ( h0η q& )
2
h0η η
A= q& + _ - q q + .
h00 h00 h00 ( h00 )2

This metric in a five-dimensional manifold of space-time-mass may

be rewritten as
 1 
( dq0 )2 =  (dσ )2 ,
 h00 

where

( dq0 )2 ≡ gˆ ij dxi dx j , i, j = 0,1,2,3,4,

and

(dσ )2 ≡ q ′ij dxi dx j , i, j = 0,1,2,3,4,

with x0=ct, x1=q1, x2=q2, x3=q3, and x4=γ/a0. Thus we may write

1 1
( dq0 )2 = qˆ ij dxi dx j =  (dσ )2 =   g ′ij dxi dx j . (3.8)
 f  f

Having established the metrics in Eqn. (3.8) in the manner

prescribed by the Dynamic Theory, the geometry must be Weyl geometry;
wherein the potential five-vector is defined as
1
∂ lnf 2
φi ≡ + _ (3.9)
∂ xi

and the field tensor is given by

F ij ≡ φ i, j - φ j,i . (3.10)

98
3.4 Gauge Function Field Equations

In order to isolate the field equations resulting from a gauge function

from the field equations produced by a vector curvature, let us consider a
local Euclidean manifold for (dσ)2.
Now the field tensor given by Eqn. (3.10) has 25 components. We
would like to determine the field equations for these components. The
quickest, though not the only, way is to consider the five dimensions to be
x0=ict, xα=xα, α=1,2,3,4. The field tensor is then defined to be
0 iE 1 iE 2 iE 3 iV 4
- iE 1 0 B3 - B2 V1
F ij = - iE 2 - B 3 0 B1 V2.
- iE 3 B2 - B1 0 V3
- iV 4 -V 1 -V 2 -V 3 0

Using Bianchi's identities

∂ F ij ∂ F jk ∂ F ki
+ + =0
∂x k
∂x i
∂ xj

and the various combinations of the indices 0, 1, 2, 3, 4 we obtain the field

equations

ƥB = 0
1 ∂B
∆x E + =0
c ∂t
∂B (3.11)
∆xV + a0 =0
∂γ
1 ∂V ∂E
∆V 4 + + a0 = 0.
c ∂t ∂γ

The definition of the five-vector current density

∂ F ij 4π
≡ Ji (3.12)
∂x i
c

yields the equations

∂V 4
∆ • E + a0 = 4πρ
∂γ
1 ∂E ∂V 4π J
∆x B - + a0 = (3.13)
c ∂t ∂γ c
1 ∂V 4 4π J 4
∆ •V + = - .
c ∂t c

99
 In addition to these field equations there is the statement of
conservation of charge where
∂ Ji
= 0, i = 0,1,2,3,4,
∂ xi

so that
∂ρ ∂ J4
+ ∆ • J + a0 = 0. (3.14)
∂t ∂γ

For ease in future reference to these eight field equations they may
be rewritten as

ƥB = 0 [a]
1 ∂B
+ ∆x E = 0 [b]
c ∂t
1 ∂E ∂V 4π J
∆x B - + a0 = [c]
c ∂t ∂γ c
∂V 4
∆ • E + a0 = 4πρ [d]
∂γ
∂ρ ∂ J4 (3.15)
+ ∆ • J + a0 =0 [e]
∂t ∂γ
∂B
∆xV + a0 =0 [f]
∂γ
1 ∂V ∂E
∆V 0 + = a0 [g]
c ∂t ∂γ
1 ∂V 4 4π J 4
∆ •V + = - [h]
c ∂t c

3.5 Energy-Momentum Tensor

If we follow the approach of relativistic electrodynamics, we may

define the tensor {T} in terms of the field tensor {F} according to

 1  1 
T jk ≡    F jλ F λk + δ jk F st F st 
 4π  4 

Using the field tensor to calculate the components of the

energy-momentum tensor we find that the components are given by

100
 -1
T 0α = [( E x B )α + V 4 V α ] , α = 1, 2, 3 ,
4π
1
T 00 = [ E 2 + B 2 + V 42 + V 2 ] ,
8π
i
T 04 = [ E •V ] ,
4π
1
T 4α = [ V 4 Eα + ( V x B )α ] , α = 1, 2, 3 ,
4π
1
T 44 = [ V 42 + B 2 - E 2 - V 2 ] ,
8π

and
1 1
T αβ = { Eα E β + Bα B β - V α V β - δαβ [ E 2 + B 2 + V 42 - V 2 ]} ,
4π 2

where
α ,β = 1, 2, 3.

Equations (3.15) form a set of eight Maxwell-type equations which

obviously reduce to Maxwell's four equations.
The wave equations for the new field quantities may be derived using
standard assumptions.

∂ 1 ∂ 2 V 4 4π ∂ J 4 ∂V 1 ∂ 2 V 4
( ∆ •V )+ = =∆• +
∂t c ∂t 2
c dt dt c ∂ t 2

and

1 ∂V ∂E 1 ∂V
∆ • ( ∆V 4 )+ ∆ • = - a0 ∆ • = ∆• + ∆ • ∆V .
c ∂t ∂γ c dt

Therefore,

1 ∂ 2 V 4 4π ∂ J 4 ∂E
2
∆ V4- = 2 - a0 ∆ • .
c ∂t c ∂t ∂γ
2 2

For the vector field we have:

∆ ∂V 4 4π
∆ ( ∆ •V )+ = - ∆( J4 )
c ∂t c

and
1 ∂ 4π
∆x( ∆xV ) + ∆ 2 V + ∆V4= ∆ J4 ;
c ∂t c

101
therefore

1 ∂ 2 V 4π a ∂ E
2
∂B
2
∆ V- = ∆ J4+ 0 + a0 ( ∆ x ) .
c ∂t c ∂ t ∂γ ∂γ
2 2
c

∂
But ∆ • E = 4πρ - a0 V 4 70, so that
∂γ

1 ∂ 2 V 4 4π ∂ J 4 ∂ ∂V 4
2
∆ V4- = 2 - a0 4πρ - a0
c ∂t c ∂t ∂γ ∂t
2 2

1 ∂ E 4π ∂V
and ∆x B - J - a0 72 , so that
c ∂t c ∂γ

1 ∂ 2 V 4π ∂ 4π ∂E ∂V
2
∆ V- = ∆ J 4 + a0 J +2 - a0 .
c
2
∂ t
2
c ∂γ c ∂t ∂γ

Now the wave equations for the usual vector and scalar potentials
are

2 1 ∂2 A 4π
∆ A- =- J
c ∂t
2 2
c

and
1 ∂2 φ
∆ φ- = - 4πρ .
2

c ∂t
2 2

We may differentiate these with respect to the mass density and substitute
them into our wave equations and get
1 ∂2 4π ∂ J 4 2
∂ V4
2
∆ V4 - 2 V4
= 2 + a02
c ∂t c ∂t ∂γ 2
2

and

1 ∂ 2 V 4π ∂  ∂E ∂V 
2
∆ V- = ∆ J 4 + a0 2 - a0  ,
c ∂t ∂γ  ∂t ∂γ 
2 2
c

where

∂φ ∂A
V 4 ≡ V 4 + a0 and V = V - a0 .
∂γ ∂γ

The field energy density may be defined by

1
ξ≡ [ E • E + B • B + V • V + V 42 ] ,
8π

102
and the electrical Poynting vector may be defined by
c
SE ≡ (E x B) .
4π

Now the electrical Poynting vector represents the outward flow of the
electromagnetic field energy through a surface. If we take the total vector,
whose components are T0α, to be the total flow of energy, then the vector
c
with components V 4 V α 81must be the outward flow of energy due to changes of the
4π
mass density within the surface. Let us designate the mass energy vector as
c
Sm ≡ (V 4V ) ,
4π

so that the total energy vector is

S = SE + Sm

whose components are

c c
Sα = [ ( E x B )α + V 4 V α ] = - ( ) T 0α .
4π i

The Dynamic stress tensor may be defined as the three-dimensional

tensor whose elements are
1 1
D
T αβ = { Eα E β + Bα B β - V α V β - δ αβ [ ξ - 2 V 2 ]}
4π 2

The Maxwell stress tensor is defined in electrodynamics as the

three-dimensional tensor with elements

M 1 1
T αβ = { Eα E β + Bα B β - δ αβ [ E 2 + B 2 ]}
4π 2

In terms of the Maxwell stress tensor, the Dynamic stress tensor may be
written as

D M 1 2 2
T αβ = T αβ - { V α V β - δ αβ [ V - V 4 ]} .
2

Then in terms of the above defined quantities

103
  ξ 1 i 
 - S ( E • V )
 c 4π 
 
 i {T D } i 
{T} =  - S V 4 E + ( V x B )] .
 c 4π 
 
 i [ V 4 E + ( V x B )] 1 [ 2 + 2 - 2 - 2 ] 
 ( E •V ) V4 B E V 
 4π 8π 

Suppose we calculate the trace of the energy-momentum tensor:

1
t R {T} = T jj = T jj + ξ + [ V 42 + B 2 - E 2 - V 2 ]
D
8π
1
= [ B 2 + V 42 ] + T αα
4
4π
1 1 3
= [ B 2 + V 42 ] + [ E 2 + B 2 - V 2 - ( E 2 + B 2 + V 42 - V 2 )]
4π 4π 2
1 1 2 1 2 1 2
= [ B - V - V4]
4π 2 2 2
1
= [ B 2 + V 2 - E 2 - V 42 ] .
8π

3.6 Force Density Vector.

The force density vector may be defined in terms of the divergence of

the energy-momentum tensor. Therefore, suppose we calculate the
five-dimensional divergence of the tensor {T}, or

∂ T jk 1 ∂ 1
= [ F jλ F λk + δ jk F st F st ] .
∂x k
4π ∂ x k
4

Because of the antisymmetry of Fjk, the first term may be written as

∂ F jλ ∂ F λj
k F λk
= F kλ .
∂x ∂ xk

By interchanging the indices k and l

∂ F jλ ∂ F kj 1  ∂ F j λ ∂ F jλ 
k F λk
= λ F λk
=  +  F λk .
∂x ∂x 2  ∂ xk ∂ x λ 

Using the Bianchi identity

∂ F jλ ∂ F kj ∂ F λk
+ + =0 ,
∂x k
∂x λ
∂ xj

the terms contained within the parentheses may be written as

104
 ∂ F jλ 1 ∂ F λk 1 ∂( F λk F λk )
F λk = j F λk
=- .
∂x k
2 ∂x 4 ∂ xi

Substituting this back into the expression for the divergence, the last term
will be canceled because l, k, s, and t are dummy indices. Then the
divergence becomes
∂ T jk 1 ∂ F λk
= F jλ .
∂x k
4π ∂ xk

By interchanging the indices k and l on the right-hand side we obtain

∂ T jk 1 ∂ F kλ
= F jk . (3.16)
∂ x k 4π ∂ xλ

The Dynamic force density five-vector may now be defined as

K ≡ Div 5 {T} .

Therefore, the components of K are given by

1 ∂ F kλ
K j= F jk .
4π ∂ xλ

But the five-vector current density is given by

∂ F kλ -4π
= Jk .
∂ xλ c

The components of the five-vector force density become

-1  4π  - 1
K j= F kj  J k  = J k F kj .
4π  c  c

Now, since Jk = (icρ, J, J4), then

i
K0 = [ J • E + J 4V 4 ] ,
c
, (3.17)
1 J
K = ρ [ E + ( u x B )] + 4 V
c c

and

J •V
K4= ρV 4 - ,
c

where J = ρu. These then are the components of the force density
five-vector resulting from a gauge field in the Dynamic Theory. These
components reduce to the four components of the Lorentz force density
should V4 = V = 0.
105
 With the interpretation that the four force density components with
subscript 1 through 4 are the force density vectors which appear in the
First Law as Fα, then the force density vector provides the connection
between the First Law and the geometry of the sigma manifold discussed in
section 2.9. Thus, the existence of the vector field ϕi is also demanded by
the Dynamic Theory and need not exist as a separate assumption.

3.7 Equation of Energy Flow.

Consider the zeroth component of the Dynamic force density

five-vector
∂ T 0k ∂ T 00 ∂ T 0α ∂ T 04
K0= = + + .
∂ xk ∂ x0 ∂ xα ∂ x4

Then

i ∂ξ i ∂ Sα i ∂( E • V )
[ J • E + J 4V 4 ] = - +
c ∂(ict) c ∂ x α
4π ∂ x4

or

1 1 ∂ξ 1 ∂ S α 1 ∂( E • V )
[ J • E + J 4V 4 ] = + - ,
c c ∂t c ∂ xα 4π ∂ x 4

or, since x4 = γ/a0,

∂ξ ∂ S α a0 c ∂( E • V )
- J • E - J 4V 4 = + - .
∂t ∂ xα 4π ∂γ

Rearranging the terms

∂ξ a c ∂( E • V )
div S + = - J • E - J 4V 4 + 0
∂t 4π ∂γ

and separating out the electrical Poynting vector leads to

∂ξ a c ∂( E • V )
div S E + = - J • E - J 4 V 4 - div S m + 0 .
∂t 4π ∂γ

This then is the five-dimensional energy flow equation.

3.8 Momentum Conservation

The expression for the conservation of momentum may be obtained

from the space portion of the force density five-vector

106
 ∂ T αk T ∂ T αβ ∂ T α 4
K= = ∂ α 00 + + , α , β = 1, 2, 3 .
∂x k
∂x ∂ xβ ∂ x4

∂
But T αββ 110 is the three-dimensional divergence of the Dynamic stress tensor {TD},
∂x
therefore,

1 ∂S a ∂
K=- + div { T D } + 0 [ V 4 E + ( V x B )] .
c
2
∂t 4π ∂γ

If we consider a volume in which all the material is contained and

outside of which the field vanishes, then integrating over this volume yields

1 ∂ S a0 ∂
∫{K + c
v
2
-
∂t 4π ∂γ
[ V 4 E + ( V x B )]} dv = ∫ div { T D } dv .
v

The integral of K gives the total force (i.e., the time derivative of the
mechanical momentum p less the vector  γ&vα / 1 - β 2  113. Now define the vector
 

S a0 ∂ γ&vα
g≡ - ∫{ [V 0 E + (V x B )+ } dt .
c
2
4π ∂γ 1- β 2

Then define

∫ g dσ ≡ G
v
,

so that
d
( p + G ) = ∫ div { T D } dv .
dt v

Using the divergence theorem the volume integral may be converted to a

surface integral so that
d
( p + G ) = ∫ { T D } • n da .
dt s

If the field vanishes outside of V, it must do so also on the boundary

surface s, hence
d
( p + G )= 0 .
dt

Therefore, it is not the mechanical momentum p but the quantity p +

G which is conserved. Therefore, we must interpret G as the momentum of
the field and

107
  ∂ 4πγ&v 
g=
S
-
a0
4π
∫ 
 ∂γ
[V • E +(V xB ) + ]  dt
1 - β 2 
2
c


as the momentum density of the field.

3.9 Gauge Field Pressure

The Dynamic stress tensor is given by

1 1
D
T αβ = { Eα E β + Bα B β - V α V β - δ αβ E 2 + B 2 + V 42 - V 2 ]}
4π 2
1 2 2 2 3 2
= - [ E + B -V ] - V4 .
8π 8π

Now separate the three-dimensional dynamic stress tensor into a traceless

and an isotropic tensor.
1 1
D
T αβ = { Eα E β + Bα B β - V α V β - γ αβ [ E 2 + B 2 + V 42 - V 2 }]
4π 2
1 1 2 2 2 2
= { Eα E β + Bα B β - V α V β - δ αβ [ E + B + V 4 - V ]}
4π 8π
1 2 1
= [ Eα E β + Bα B β - V α V β ] - ( )( )δ αβ [ E 2 + B 2 - V 2 ]
4π 3 8π
1 1
- ( )( )δ αβ [ E 2 + B 2 + 3V 42 - V 2 ]
3 8π
≡ t αβ + τ ′αβ

where
1 1
tαβ ≡ { Eα E β + Bα B β + V α V β - ( )δ αβ [ E 2 + B 2 - V 2 ]
4π 3π

and
1
τ ′αβ ≡ - ( )δ αβ [ E 2 + B 2 + 3V 42 - V 2 ] .
24π

Now
1
t r { tαβ } = ( ) [ E 2 + B 2 - V 2 - ( E 2 + B 2 - V 2 )] ≡ 0
4π

and
1
t r { τ ′αβ } = - ( ) [ E 2 + B 2 + 3 V 42 - V 2 ] .
8π

108
 Consider the definition
1 1
t δ αβ ≡ τ ′ = - ( )δ αβ [ E 2 + B 2 + 3V 42 - V 2 ] .
3 24π

Then
1
t= -( ) [ E 2 + B 2 + 3V 42 - V 2 ]
8π

and
 t 0 0
 
τ ′αβ =  0 t 0 
 
0 0 t 

The isotropic part of the stress tensor is usually called the

"pressure." Therefore, define 3p = t in accordance with customary
notation, so that
1
p= -( ) [ E 2 + B 2 + 3V 42 - V 2 ] .
24π

With the exception of the factor of 3 this reduces to the "radiation

pressure" for an electromagnetic field when V = V0 =0. Note that this
pressure may be zero since it is the sum and difference of squares, or p =
0, when
2 2 2 2
V = E + B + 3V 4 .

This may prove to be an important point when considering boundary

conditions in cosmology or the study of elementary particles.

109