Elsevier Editorial System(tm) for Journal of Electrostatics

Manuscript Draft


Manuscript Number: ELSTAT-D-11-00096R3

Title: A Generalization Of Gauss's Theorem In Electrostatics

Article Type: Full Length Article

Keywords: Forces and Fields; Gauss's Theorem; Surface Charge; Method of Images; Conductors and
Dielectrics in Constant Electric Field

Corresponding Author: Mr. Ishnath Pathak,

Corresponding Author's Institution: IIT Guwahati

First Author: Ishnath Pathak

Order of Authors: Ishnath Pathak

Abstract: Gauss's theorem of electrostatics states that the flux of the electrostatic field over a closed
surface equals the net charge enclosed divided by the permittivity of free space. In this paper we prove
a generalization of Gauss's theorem which allows charges to lie on the surface of integration. For the
majority of cases the statement of our generalized Gauss's theorem can be assumed to be this: the flux
of electrostatic field over a closed surface equals 1/vacuum permittivity times the sum of the net
charge enclosed by S and half the value of net charge contained by S.




1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
61
62
63
64
65
A Generalization Of Gausss Theorem In Electrostatics
Ishnath Pathak

Department of Civil Engineering, Indian Institute of Technology, Guwahati 781039, India.

ABSTRACT
Gausss Theorem of electrostatics states that the ux of the electrostatic eld over a closed surface equals the net charge
enclosed divided by the permittivity of free space. In the derivation it is assumed that no charge lies on the surface in
question. Consider the problem of evaluation of the ux of the electrostatic eld due to a uniformly charged spherical
surface of radius R, carrying a net charge Q, through the surface itself. Due to the reason mentioned just now, we cannot
apply Gausss Theorem. In this paper we prove a generalization of Gausss Theorem which allows charges to lie on the
surface of integration. For the majority of cases the statement of our Generalized Gausss Theorem can be assumed to
be this: the ux of the electrostatic eld over a closed surface equals 1/
0
times the sum of the net charge enclosed by S
and half the value of net charge contained by S. Applying this theorem to the uniformly charged spherical surface we nd
at once that the ux equals Q/2
0
, a result we can corroborate by considering the surface as a hollow spherical shell of
innitesimal thickness carrying an appropriate innite volume charge density.
1. Introduction
One of the most important theorems of electrostatics is Gausss Theorem concerning the ux of the electrostatic eld
(of a set of source charges) over a boundary to a connected region. It is explicitly stated sometimes[1], that the closed
surface of integration must not contain any charges. But what happens if the surface contains charges in any problem in
hand? To the best of the authors knowledge, the generalization of Gausss Theorem has not been discussed anywhere till
now, and it is analysed here. For the following sections in this paper we discuss what missing terms appear in Gausss
Theorem if we permit presence of charges on the surface of integration.
One faces a problem with Gausss Theorem only when point, line or surface charges are present on the surface - the
surface may contain volume charges and it wont cause any trouble. If, for instance, there is a volume charge distribution
(

r ) in space, then the ux

da on any closed surface equals

EdV ; and as for a volume charge distribution,



E is nite and equal to (

r )/
0
, we see that the ux equals Q
enc
/
0
. In essence, why Gausss Theorem holds well
even in case of volume charges lying on the surface, and not in case of point, line, or surface charges, is simply because
even when a volume charge distribution exists on a surface, the net charge on the surface remains zero, while this is not so

Electronic mail: pathak.ishnath@gmail.com

1
Manuscript
Click here to view linked References
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
61
62
63
64
65
for point, line and surface charges. We, in this paper, will frequently distinguish between point, line, surface and volume
charges - the meanings of which are obviously understood.
2. Flux of the Electrostatic Field of a Uniformly Charged Spherical Surface
There isnt any trouble with Gausss Theorem - the only issue is that it assumes that no charge lies on the surface
over which the ux is to be evaluated. If charges lie on the surface, then - as a rule - Gausss Theorem cannot be
applied. We consider as an example, the case of a net charge Q smeared uniformly over the spherical surface r=R. By
using Coulombs law and invoking the principle of superposition, we can evaluate

E(r z), which for r < R and r > R
will denitely provide us the values 0 and
Q
4
0
r
2
z, respectively. Here, z denotes the unit vector pointing in the direction
of increase of the cartesian coordinate z and throughout this paper we adopt a convention of denoting the unit vector
pointing in the direction of steepest increase of any scalar function u of coordinates by the symbol u, which has a hat over
the variable. Also we can take note here that since the situation exhibits spherical symmetry, the initial choice of the axis
of z is arbitrary and our nal results about the magnitude of the eld strength stand as equally applicable to any radial
position at an equal distance. Now, we expect the process of evaluation to boil the integral representing the superposition
down to some nite value E
0
z even in case of r = R, but we cannot conclude (from this evidence alone) that the ux
described in the abstract equals 4R
2
E
0
.
For the value of the integral doesnt have any physical relevance as the electrostatic eld undergoes an abrupt disconti-
nuity at a surface charge and has no well dened value on the surface itself. We shall evaluate the ux through the surface
by considering the surface charge to be a spherical shell of inner radius R and outer radius R + carrying a uniform
volume charge of density given by Q = (4/3)((R + )
3
R
3
), i.e.
=
3Q
4(3R
2
+ 3R +
2
)
(1)
and then let tend to zero. The electrostatic eld for the original spherical surface as well as this spherical shell is
radial and spherically symmetric and a single scalar function E(r) determines the eld:

E(

r ) = E(r) r (2)
The ux of any such spherically symmetric central eld through a spherical surface S centered at the origin is given
by =

S
Eda. When we picture the spherical surface as a thin spherical shell, the ux through the spherical surface
must be understood to stand for

S
< E > da, where
< E >=

EdV

dV
(3)
is the mean value of the scalar eld E over the spherical shell. Now, E(r) is equal to Q(r)/4
0
r
2
, where Q(r) is
the charge lying inside the spherical surface of radius r centered at the origin. So,

EdV =
_
R+
R
E(r)4r
2
dr =
(
_
R+
R
Q(r)dr)/
0
. So, the ux equals
=
(1/
0
)(
_
R+
R
Q(r)dr)
(4/3)((R + )
3
R
3
)
(4R
2
) =
3R
2

_
R+
R
Q(r)dr
(3R
2
+ 3R +
2
)
(4)
2
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
61
62
63
64
65
Now, Q(r) = (4/3)(r
3
R
3
) =
Q(r
3
R
3
)
(3R
2
+ 3R +
2
)
. So,
=
3R
2
Q

_
R+
R
(r
3
R
3
)dr

2
(3R
2
+ 3R +
2
)
2
=
3R
2
Q
4
0

6R
2
+ 4R +
2
(3R
2
+ 3R +
2
)
2
(5)
Letting 0 in this expression, we get = Q/2
0
. As we shall see in the coming sections, the Generalized Gausss
Theorem, which is to be enunciated, states that the ux of the electrostatic eld, through any surface on which source
charges might as well lie, due to a source charge conguration consisting of only volume charges and surface charges,
equals

da = Q
enc
/
0
+ Q
con
/2
0
, where Q
enc
is the net charge lying inside the open region to which our surface is
the boundary, and Q
con
is the net charge lying on the closed surface itself. In case of our spherical surface, Q
enc
= 0 and
Q
con
= Q, so the ux is = Q/2
0
. In conclusion, our plan in this paper is to generalize Gausss Theorem so that we
can tell the ux of the electrostatic eld through a surface carrying charges as easily as Gausss Theorem empowers us to
tell the same for surfaces free of charges; and thus avoid operating with actual integrals and limits as weve just done.
It can be mentioned here in passing that the ux 4R
2
E
0
, E
0
being the constant (having the dimensions of electrostatic
eld) described in the beginning of this section, also equals Q/2
0
. A general area element on the surface is R
2
sindd,
and it carries a charge Qsindd/4. This element of area, placed at the point R r, sees the point r z at (r Rcos) z
Rsincos x Rsinsin y, and hence, is distant

r
2
+ R
2
2rRcos from r z. The situation is shown in Figure-1. The
electrostatic eld at r z due to this element is
1
4
0
Qsindd
4
((r Rcos) z Rsincos x Rsinsin y)
(r
2
+ R
2
2rRcos)
3/2
(6)
Integrating using the principle of superposition and noticing that cos and sin integrate to zero over (0, 2), we see

E(r z) =
Q z
8
0

_
0
(r Rcos)sind
(r
2
+ R
2
2rRcos)
3/2
(7)
Making the change of variable p =

r
2
+ R
2
2rRcos in equation (7) gives

E(r z) =
Q z
16
0
r
2
R
R+r
_
|Rr|
_
r
2
R
2
p
2
+ 1
_
dp (8)
=
Q z
16
0
r
2
R
_
(r + R)(r R)
_
1
| R r |

1
R + r
_
+ (R + r) | R r |
_
(9)
This equation frequents in literature as a common textbook example[2]. If we take the case r > R, the above equation
gives

E(r z) =
Q
4
0
r
2
z, and letting r < R in the same equation gives

E(r z) = 0: the exact expression for the electrostatic
eld of the spherical surface. However, putting r=R in equation (7) gives
E
0
z =
Q z
8
0

_
0
4Rsin
3
(/2)cos(/2)d
8R
3
sin
3
(/2)
=
Q
8
0
R
2
z (10)
so that 4R
2
E
0
turns out to be Q/2
0
. We remind ourselves again that this is just an observation with admittedly no
physical relevance, as discussed in the beginning of this section.
3
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
61
62
63
64
65
We evaluated the ux of a net charge Q smeared uniformly over the spherical surface r = R to be 4R
2
E
0
, where
E
0
=
Q
8
0
R
2
. Let us now model the surface charge as a net surface charge Q
0
smeared uniformly over the spherical
surface r = R plus a net surface charge Q

0
smeared uniformly over the spherical surface r = R+ plus a net volume charge
Q
1
spilled uniformly in the enclosed spherical shell, i.e. the open region R < r < R+, bounded by the two surfaces. This
kind of describing a surface charge in terms of surface charge may be considered recursive and while evaluating the uxes
of Q
0
and Q

0
, as we shall do now, we shall use the Generalized Gausss Theorem derived by describing a surface charge
as the limit of a volume charge. That other descriptions of a surface charge cannot lead to dierent results concerning
the ux is corroborated in the next three paragraphs. Let us set Q
0
+Q
1
+Q

0
to be Q and then let 0. Now, we can
write

E(

r ) = E
0
(r) r + E
1
(r) r + E

0
(r) r, where E
0
(r) r, E
1
(r) r and E

0
(r) r are the electrostatic elds of Q
0
, Q
1
and Q

0
.
Now, the ux of the net charge Q through the spherical surface r = R, when modeled by R r R + carrying the
three charge distributions, takes the value
0
+
1
+

0
= 4R
2
E
0
+4R
2
< E
1
> +4(R+)
2
E

0
, where < E
1
> is the
mean value of the scalar eld E
1
(r) over the spherical surface. Now as 0,
1
4R
2
< E
1
>= Q
1
/2
0
(cf. equation
(4)). Also R + R as 0 and we get
0
Q
0
/2
0
and

0
Q

0
/2
0
. Thus, the ux, according to this model too,
is = Q/2
0
.
Many who have been in touch with me and have known Gausss Theorem and even the caveat that it doesnt apply to
surfaces containing surface charges, tried to convince me that the ux of the electrostatic eld due to a uniform surface
charge =
Q
4R
2
over the spherical surface r = R, through the surface itself, is Q/
0
and not Q/2
0
, where Q = 4R
2

is the net charge contained by the spherical surface. Such suggestions are what I am going to falsify next by means of
novel arguments.
We can model a net surface charge Q uniformly distributed at r = R as two surface charges Q
0
at r = R and Q

0
at
r = R + plus a net volume charge Q
1
uniformly distributed in R < r < R + and then let 0. We can make an
assumption that the statement =
Q
k
0
is true for some value of k. By assumption, we have the uxes
0
and

0
as
Q
0
/k
0
and Q

0
/k
0
. Thus, by assumption
Q
0
+ Q
1
+ Q

0
k
0
=
Q
0
+ Q

0
k
0
+ 4R
2
< E
1
>, where 4R
2
< E
1
>=
Q
1
2
0
(by
equation (4)). Thus, we get 2(Q
0
+ Q
1
+ Q

0
) = 2(Q
0
+ Q

0
) + kQ
1
, which gives (2 k)Q
1
= 0. Now as Q
1
can be set
at any arbitrary value dierent from Q but close to Q, we can set (k 2) equal to zero. Thus the statement, which we
assumed at the outset of our arguments, leads to a contradiction for any value of k other than 2 (and surely for k = 1).
Thus, we see that by modeling the net surface charge Q distributed uniformly at r = R, as a limiting case of a net
volume charge Q distributed uniformly in a spherical shell R < r < R+ of volume V, and then letting 0, we get the
result =
Q
2
0
. I nd it worth mentioning here that physically thinking it seems as convincing as a mathematical proof to
me that if the result is what I have derived modeling a surface charge by the limit of a uniform volume charge, then it will
be essentially the same - if surface charge can in principle be modeled as the limit of a volume charge - when the surface
charge Q is modeled by the limit of any spherically symmetric non-uniform volume charge distribution of a net charge Q
in the spherical shell R < r < R + . I think the above physical argument is convincing enough for a physicist like me to
believe the truth of the equivalent mathematical statement (equation ( 11 ) below) concerning the functional dependence
of the spherically symmetric volume charge density on the distance from the center of the surface. It will be no less than
satiating to see an expert of integral calculus of one variable analyse it and device for it, a completely mathematical proof.
4
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
61
62
63
64
65
1
2
R+
_
R
r
2
(r)dr
3R
2
(R + )
3
R
3
R+
_
R
_
_
r
_
R
u
2
(u)du
_
_
dr as 0 (11)
3. Flux of the Electrostatic Field of a Point Charge
We prove in this section that the ux of the electrostatic eld

E
i
of a charge q
i
over a boundary S to a connected
region R
enc
is given by

S

E
i

da =
_

_
0, if q
i
lies outside S;
q
i

i
4
0
, if q
i
lies on S at a point

r
i
and

S
(

da )/
2
,

being

r

r
i
, equals
i
;
q
i

0
, if q
i
is enclosed by S.
(12)
This statement is axiomatic but still we will write a satisfactory proof soon. Yet, before doing that, we will try to
expatiate a little. An analogy is often drawn[3] between the actual situation and an imaginary situation with the point
charge radiating particles at a constant rate in a spherically symmetric manner. Since the electrostatic eld is given by a
radial inverse square law equation, the total number of particles passing per unit time through a surface can be thought
of to be the ux of the electrostatic eld through that surface. Now if we consider a charge lying on the closed surface
under consideration, the total number of particles passing per unit time through the closed surface should be obtained by
multiplying the total number of particles emitted by the charge per unit time by 1/4 times the inner solid angle formed
at the surface at the place where the point charge lies inasmuch as any particle radiated outside this solid angle will not
leave the interior without entering. Alternatively, as is done in many conventional texts[4], it may be pointed out that
the ux of the electrostatic eld of a point charge q
i
through any innitesimal surface (with the normal to the surface
pointing away from the charge) is equal to q
i
d/4
0
, where d is the solid angle subtended by the innitesimal surface
at the position of the charge. Once this result is known, one can say at once that the ux of the eld of a single point
charge through a closed surface is given by equation (12). If the reader is convinced by the above arguments, then the
proof which now follows in this section (and ends at equation (17)) can be skipped.
We denote by

r
i
the position of q
i
. Let us rst consider the case A when the charge lies outside the boundary S. In
this case

E
i
is dierentiable properly over an open connected region containing the surface S along with its enclosure.
Application of divergence theorem will yield the result. Let us now consider the case B when

r
i
lies on S. The situation
is shown in Figure-2. Let the inner solid angle formed on S at the point

r
i
be
i
in measure. We chose an arbitrary
non-zero radius , suciently small, say less than a critical radius
c
, such that the part S

of the spherical surface of

radius centered at

r
i
, which lies not outside the region R
enc
enclosed by S, satises the following two conditions:
1. It divides R
enc
into two parts R
1
and R
2
such that R
2
, the one not adjacent to

r
i
, is enclosed by a closed surface
S

2
which has the inner surfaces of S as its inner surfaces in case R
enc
has cavities;
2. It has no missing patches, i.e. S

is bounded by a single closed curve P

and not by a group of closed curves.

Hence, the closed curve P

divides the outer surface of S into two parts, S

1
and S
2
, where S
2
and S

together with
the inner surfaces of S form the closed surface S

2
which encloses R
2
. Also S
1
and S

together form the closed surface

S

1
which encloses R
1
. Now, we can write the ux of the electrostatic eld through S as
5
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
61
62
63
64
65

S

E
i

da =

1

E
i

da +

2

E
i

da (13)
because when we add the ux
=
q
i
4
0
area ofS

2
(14)
of

E
i
through S

to the second integral on the right of equality and subtract the same from the rst one, we get
identity. Now according to the result of case A, the second integral must vanish. The ux of

E
i
through S can now be
easily obtained

S

E
i

da = lim
0

S

E
i

da = lim
0

1

E
i

da = lim
0
=
q
i

i
4
0
(15)
Finally we consider the case C when

r
i
lies enclosed by S. We consider any plane passing through

r
i
and denote by
S
0
the part of the plane that is enclosed by S. We denote again by R
1
and R
2
the regions into which S
0
divides R
enc
, and
by S

1
and S

2
the closed surfaces that enclose these regions. So we have again

S

E
i

da =

1

E
i

da +

2

E
i

da (16)
Only this time we are ready with an identity, for this time the ux of

E
i
through S
0
is zero. Now according to the
result of case B, both the terms on the right of equality are equal to q
i
/2
0
, so that the ux is

S

E
i

da =
q
i

0
(17)
We conclude this section with a few examples. In a common problem in electrostatics[5], one is asked to evaluate the
ux of the electrostatic eld of a point charge Q lying at the corner of a cube having edges of length a, through a face of
the cube not adjacent to Q. The situation is as shown in Figure-3. There are three such faces, and by symmetry, the ux
through each one of them is equal to a common value . In addition, the ux through any of the remaining three faces
vanishes as the eld in any of these faces lies in the plane of the face. Equation (12), which is the Generalized Gausss
Theorem for an electrostatic eld due to a single point charge, gives us the equation 3 = Q(/2)/(4
0
), which yields
= Q/24
0
. This result is conventionally derived using the reasoning that must be one fourth of the value of the
ux of the electrostatic eld of Q through a face of a cube, as shown in Figure-4, which has edges of length 2a, and its
geometric centre coinciding with Q. This method of solution using Gausss Theorem suggests 46 = Q/
0
, which reads
= Q/24
0
- the same result.
The two most well known applications of Gausss Theorem are to evaluate the eld of a uniformly charged sphere and
that of an innite at plane surface carrying a uniformly distributed charge per unit area . The rst problem was taken
in section 2, where the statement of the Generalized Gausss Theorem for a particular case was assumed. We consider
the second problem here. If the result is to be obtained using Gausss Theorem, one has to argue using symmetry that
the magnitude of the electrostatic eld at a distance d from the plane is the same on both sides and its direction on both
sides is either away from the plane or towards the plane. Choosing as Gaussian surface, a pillbox of the shape of a cuboid
6
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
61
62
63
64
65
(as shown in Figure-5), of base area A and height 2d extending an equal distance d above and below the plane, we get

E = E(d) n, where n is a unit vector normal to the plane and directed away from the plane, and E(d) satises Gausss
Theorem 2 E(d) A = A/
0
. So, in this manner, the electrostatic eld is evaluated to be

E =

2
0
n using Gausss
Theorem. If the Generalized Gausss Theorem is applied to a Gaussian surface of the shape of a cuboid (as shown in
Figure-6), which has a base of area A lying on the plane and extends to a distance d above the plane, and symmetry is
invoked to argue that the normal component of the eld is E(d) throughout the face on the top and zero throughout the
face at the bottom, we get E(d) A = A/2
0
, i.e.

E =

2
0
n, which is the same result.
If the surface charge density on the innite at plane is not uniform, then we cannot argue using symmetry that the
electrostatic eld near the plane is normal to the plane. But, still, at any point on the plane, the component of the eld
near the plane directed along the normal to the plane pointing away from the plane, can be shown, on identical lines as
that of last example, to be

2
0
, where is the local surface charge density. Inherent in this statement is the assumption
that just near the plane, the upward normal component of the eld is the same in magnitude on both sides of the plane,
but opposite in sign. Symmetry is invoked to deduce this assumption, riding on which our statement can be derived
using Gausss Theorem. If the derivation is to be done using the Generalized Gausss Theorem, symmetry justies the
assumption that right on the plane, the normal component of the electrostatic eld vanishes.
4. The Generalized Gausss Theorem
The principal of superposition for electrostatic eld allows us to insist that the ux of the electrostatic eld over a
closed surface S is

(

E
i
)

da, where the summation is done over all charges in the source charge conguration. The
linearity of the operation of ux enables us to insist that the ux is

E
i

da). Since the ux

E
i

da of any charge
q
i
enclosed by S is q
i
/
0
, the sum of the isolated uxes of all charges enclosed by S is Q
enc
/
0
, where Q
enc
is the net
charge enclosed by S. Similarly we get that the sum of the isolated uxes of all charges lying outside S is zero, and that
of all charges residing on the continuities of S is Q
con
/2
0
, where Q
con
is the net charge residing on the continuities of S.
By continuities of S, we mean a point on S where the principle curvatures of S vary smoothly so that the inner solid angle
formed is equal to 2. The sum of the isolated uxes of the remaining charges (all of which lie at the discontinuities of S)
is left as a summation

d
=

discon
q
i

i
4
0
(18)
for we cannot say anything about the inner solid angles
i
without a particular knowledge of the geometry of the closed
surface and the locations of the charges lying at the discontinuities. Hence, we have the Generalized Gausss Theorem:
The ux of the electrostatic eld

E over any closed surface S is equal to

S

E

da =
Q
enc

0
+
1
2
Q
con

0
+
d
(19)
where Q
enc
equals the net charge enclosed by S, Q
con
equals the net charge residing on the continuities of S and
d
,
as described by equation (18), equals the ux of the electrostatic eld of the charges lying at the discontinuities of S.
7
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
61
62
63
64
65
5. The Simplest Form of the Generalized Gausss Theorem
We in this section consider those cases in which the Generalized Gausss Theorem takes a beautiful form which we
shall call as the simplest form of the theorem. We begin by conning ourselves to an electrostatic eld caused by a
conguration of charges not containing any point or line charges. In this special case Q
con
of equation (19) can also be
interpreted as the net charge contained by S, for the net charge lying on the discontinuities of S is zero anyway, because
S now contains no point or line charges - which were the only varieties which could accumulate to a nite amount by
assembling only at discontinuities. Also
d
vanishes in case of an electrostatic eld caused by such a conguration. This
needs some explanation. Let us imagine a dierent source charge conguration - the one in which each charge is replaced
by a corresponding positive one of an equal magnitude. The fact that the net charge lying on the discontinuities of S
is zero whenever the charge conguration does not contain any point or line charges implies that for the original charge
conguration well have

discon
| q
i
|= 0.
Now, since | q
i
| q
i
| q
i
|, we have

discon
| q
i
|
i
4
0

discon
q
i

i
4
0

discon
| q
i
|
i
4
0
(20)
The central term in the inequalities is
d
. So,
|
d
|

discon
| q
i
|
i
4
0

discon
| q
i
|
i
4
0
(21)
And since 0
i
4, 0 | q
i
|
i
| q
i
| 4. Hence,
0

discon
| q
i
|
i
4

discon
| q
i
| (22)
As the rightmost term in the inequalities is zero, as argued earlier, we have the middle term equal to zero. From here
we conclude that
d
vanishes. Therefore, if the source charge conguration contains only surface charges and volume
charges, then
The ux of the electrostatic eld

E over any closed surface S is equal to

S

E

da =
Q
enc

0
+
1
2
Q
con

0
(23)
where Q
enc
is the net charge enclosed by S and Q
con
is the net charge contained by S.
Superposing the uxes we get as a corollary that for any kind of charge conguration

S

E

da =
Q
enc

0
+
1
2
Q
sur

0
+
pl
(24)
where Q
enc
is the net charge enclosed by S, Q
sur
equals the net surface charge residing on S and
pl
is the ux of the
electrostatic eld of the point and line charges residing on S.
Till now equation (23) was referred to as applicable only in case of charge congurations containing only volume
charges and surface charges. From equation (24) it can be seen that (23) holds whenever no point or line charge lies
anywhere on the discontinuities of S, for in that case we see (by referring to equation (19)) that
pl
= Q
pl
/2
0
, where Q
pl
8
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
61
62
63
64
65
is the algebraic sum (with due consideration to the signs) of all point and line charges residing on S. And then, since the
recent most restriction holds if S has no discontinuity at all, (23) holds whenever S is throughout continuous. We shall
call equation (23) as the simplest form of the Generalized Gausss Theorem and see that it is almost always applicable.
6. Results
In the abstract it was mentioned that for the majority of cases the statement of our Generalized Gausss Theorem can
be assumed to be this: the ux of electrostatic eld over a closed surface equals 1/
0
times the sum of the net charge
enclosed by S and half the value of net charge contained by S. Let us now enlist the cases which we claim to be in majority:
1. The surface of integration does not contain point or line charges at any of the corners or edges;
2. The surface of integration does not have any edge or corner, i.e. it is throughout continuous;
3. The source charge conguration consists of only volume charges and surface charges.
7. Discussion
Usually, in real life problems of macroscopic electrostatics, there are xed free charges, some induced charges in the
conductors and bound charges in the dielectrics. To complete the analysis of the electrostatic conguration, it sometimes
helps to apply Gausss Theorem. In such situations one needs only to draw Gaussian surfaces within or around the
conductors and the dielectrics because application of Gausss Theorem in free space will only determine the distribution
of free charges there, which is already known. The induced charges are a surface charge distribution on the boundary of
the conductors and the bound charges are a surface charge distribution on the boundary of the dielectrics along with some
volume charge distribution inside the dielectrics in case of non-uniform polarization eld.
Here the term polarization eld, which is an indispensable and widely accepted term in electrostatics, refers to the
volume density of the dipole moment of the charges appearing in the dielectric as a result of the process of polarization.
These charges, which obviously add up to zero, are in the literature called bound charges or polarization charges, and
they reside on the boundary of the dielectric as a surface charge equal to the normal component of the polarization eld
and inside the volume, if there is a non-uniform polarization eld, as the negative of the divergence of the polarization
eld. These are very fundamental concepts at the heart of electrostatics of insulators[6].
As the relevant charge conguration is composed only of volume charges and surface charges, we can even let our
Gaussian surface meet in part with the surfaces of the conductors and the dielectrics and apply equation (23), i.e. the
simplest form of the Generalized Gausss Theorem. This remark is in consonance with our earlier claim of a wide
applicability of the simplest form.
Consider, for example, the conducting plane image problem of Figure-7 in which a point charge q is held at a height
h from an innite conducting plane and were required to gure out what is the charge density at a point P on the plane
which lies at a distance r from the foot O of the perpendicular to the plane from the point charge q. Let the charge q be
at h z and the upper surface of the conducting plane be the X-Y plane. On the plane,

r = x x +y y is the vector from the
origin and r = (x
2
+y
2
)
1/2
is the distance from the origin. At any point in space the total eld is due in part to q and in
part to the surface charges induced on the plane:

E =

E
q
+

E
S
. We draw as a Gaussian surface, a closed surface of the
form of a very thin geometry-box with the plane top of an extremely small area A lying on the surface of the conductor
9
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
61
62
63
64
65
and the remaining part extending inside the conductor to an innitesimal distance from the surface. The area A drawn
around

r is supposed so small that the surface charge density in that area is essentially uniform and equal to (

r ). The
application of the Generalized Gausss Theorem, which applies here in its simplest form, gives us

S

E

da =
Q
enc

0
+
1
2
Q
con

0
=
(

r )A
2
0
(25)
Now the ux

da is simply

E(

r ) A z as the eld inside the conductor is zero[7]. In addition, the tangential
component of electrostatic eld is zero on the surface[8], and it follows that

E(

r ) = ((

r )/2
0
) z. It follows straight from
Coulombs law (see Figure-8) that for any point

r on the conducting plane, we have

E
q
(

r ) = q(

r h z)/4
0
(r
2
+h
2
)
3/2
.
Hence,
q
4
0
(

r h z)
(r
2
+ h
2
)
3/2
+

E
S
(

r ) =
(

r ) z
2
0
(26)
As

r is orthogonal to z, we can take a dot product of both sides of the above equation with z to get
(

r )
2
0
=
qh
4
0
(r
2
+ h
2
)
3/2
+

E
S
(

r ) z (27)
Now, since the electrostatic eld of any charge points radially outwards/inwards from the charge, we must have

E
S
lying in the plane, i.e. orthogonal to z. This leads us straight to the solution of the conducting plane image problem:
(r) =
1
2
qh
(r
2
+ h
2
)
3/2
(28)
This is a well-known equation[9], and until recently it was held by the physics community that it can be derived only
using the method of images. This and other solutions without using the method of images were recently presented by the
author[10]. The other solutions dont use the Generalized Gausss Theorem either. For instance, from the last example
that we considered in section 3, we have the knowledge (which can be derived premising only Gausss Theorem) that
the z component of

E
S
is (

r )/2
0
above the plane and (

r )/2
0
below the plane. Also, the z component of

E
q
is
qh/(4
0
(r
2
+ h
2
)
3/2
) both above and below the plane. As the net electrostatic eld below the plane, i.e. inside the
conductor, must vanish, we must have (

r )/2
0
+ qh/(4
0
(r
2
+ h
2
)
3/2
) equal to zero, which again gives equation (28).
Now, lets say we have a dielectric of a weird shape occupying a region R bounded by the surface S. If we apply the
Generalized Gausss Theorem on the boundary S, we get

S

E

da =
Q
fenc

0
+
Q
benc

0
+
1
2
Q
fcon

0
+
1
2
Q
bcon

0
(29)
where Q
fenc
denotes the net free charge enclosed by S, Q
fcon
denotes the net free charge contained by S, Q
benc
denotes
the net bound charge enclosed by S and Q
bcon
denotes the net bound charge contained by S. Whenever we have an interface
between two dielectrics, then if at a point

r on the boundary, the polarization eld takes the value

P
1
, near

r inside
the rst dielectric on one side, and

P
2
inside the second dielectric on the other side, then its value exactly on the surface
at the point

r must be understood to be

P
1
+

P
2
2
.
10
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
61
62
63
64
65
For any minute sphere of radius , drawn centered at

r , carries a net dipole moment equal to (2/3)
3

P
1
+
(2/3)
3

P
2
. If the interface is between a dielectric and vacuum, the polarization at any point

r on the surface is

P =

P
in
/2, where

P
in
equals the polarization just inside the dielectric at the point

r on the surface of the dielectric.
So, for our dielectric placed at R in vacuum[11],

S

P

da =
1
2

S

P
in

da =
Q
bcon
2
(30)
From the above two equations we see

S

D

da = Q
fenc
+ Q
benc
+
1
2
Q
fcon
+ Q
bcon
(31)
where

D =
0

E +

P denotes the electric displacement. As the net bound charge Q
benc
+ Q
bcon
must vanish, we get

S

D

da = Q
fenc
+
1
2
Q
fcon
(32)
where Q
fenc
is the net free charge enclosed by S and Q
fcon
is the net free charge contained by S. This is the Generalized
Gausss Theorem for the electric displacement. It states that the ux of the electric displacement through a closed surface
equals the sum of the net free charge lying inside the open region bounded by the surface and one half of the net free
charge lying on the surface itself.
As an example, we shall now consider a problem (similar to the classic image problem) involving a dielectric instead
of the conductor. Here we have a point charge q at h z and the region below the X-Y plane is lled with a linear isotropic
dielectric of susceptibility
e
. We have to evaluate the surface charge density (

r ) at a point

r = x x + y y on the
dielectric surface. We will now write the two solutions analogous to those presented for our last example involving the
conductor. We again draw as a Gaussian surface, a cuboid of the shape of a very thin geometry-box, with a top of a
minute area A lying on the X-Y plane at

r , and the remaining volume of the box extending to an innitesimal distance
inside the dielectric. As there is no free charge lying on the surface of the dielectric or embedded inside the dielectric, the
net ux of

D through the surfaces of this box must vanish. As the sides are of innitesimal height, the z component D
z1
of
0

E +

P at the face on the top must be equal in magnitude as well as in sign to the z component D
z2
of
0

E +

P at
the face at the bottom. Now, D
z1
=
0
E
z
+
1
2

P
in
z, where E
z
is the z component of the electrostatic eld at the point

r
due to the point charge q, and

P
in
is the value of the polarization eld just inside the dielectric at the point

r . Here, we
havent included the z component of the electrostatic eld due to the bound charges on the X-Y plane as, by symmetry,
it must vanish on the plane. We have also used the fact that there cannot be any volume density of bound charges inside
a linear isotropic dielectric which has no embedded free charge inside it[12]. Based on this fact, we maintain that the eld
in this example is only due to the point charge q and the surface distribution of bound charges on the X-Y plane.
Now, as

P
in
z is (

r )[11], we have D
z1
equal to
0
E
z
+ (

r )/2. Also, the z component E
z2
of the eld inside
the dielectric is E
z
(

r )/2
0
, so that we have D
z2
=
0
(1 +
e
)(E
z
(

r )/2
0
). As D
z1
= D
z2
, we get
0

e
E
z
=
(
e
+ 2)(

r )/2. Thus we have (

r ) = 2
0
(
e
/(
e
+ 2))E
z
, which, since E
z
= qh/(4
0
(r
2
+ h
2
)
3/2
), gives us
(r) =
1
2
_

e

e
+ 2
_
qh
(r
2
+ h
2
)
3/2
(33)
11
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
61
62
63
64
65
To arrive at this result using Gausss Theorem for electric displacement, we have to argue that as there is no free
charge enclosed by the pillbox, the z component of the electric displacement just above

r is equal in magnitude as
well as sign to that just below

r . If we denote these components by D
z1
and D
z2
respectively, well once again have
D
z2
=
0
(1 +
e
)(E
z
(

r )/2
0
). Only this time, well have no contribution to D
z1
from the polarization eld and
D
z1
/
0
will simply be the z component of the electrostatic eld just above

r . Just as in the image problem, this value is
E
z
+(

r )/2
0
, so that once again we have D
z1
equal to
0
E
z
+(

r )/2, and get the same result as that of equation (33).
Acknowledgements
Its a pleasure for the author to acknowledge that the inception of the simplest form of the Generalized Gausss Theorem
in the mind of the author occurred in 2006 in a conversation with his younger brother Sharvanath where the author was
integrating the ux of the electrostatic eld of a point charge placed on an ellipsoid at the point nearest to one of the
foci. Therefore, from the viewpoint of the author - as opposed to his brothers - this work belongs to his brother. He also
takes pleasure in thanking the reviewers for many fruitful suggestions, which in each critical review impelled the author
to create new arguments to include in the discourses answers to curious ambiguities left earlier unexamined.
12
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
61
62
63
64
65
References
[1] E.M.Purcell, Electricity and Magnetism, rst ed., McGraw-Hill, New York, 1965, p. 22.
[2] M.Zahn, Electromagnetic Field Theory: a problem solving approach, John Wiley and Sons, 1979, p. 78, eq. (19).
[3] See, for example, the last paragraph on p.24, reference [1].
[4] J.D.Jackson, Classical Electrodynamics, third ed., John Wiley and Sons, 1999, pp. 27-28.
[5] D.J.Griths, Introduction to Electrodynamics, third ed., Prentice-Hall, Upper Saddle River, NJ, 1999, p. 69, Prob.
2.10.
[6] R.P.Feynman, R.B.Leighton, and M.Sands, The Feynman Lectures on Physics, Addison-Wesley, New York, 1964,
Vol. 2, sections 10-2 and 10-3.
[7] See, for example, reference [5], property (i) in section 2.5.1 on p. 97.
[8] L.D.Landau, E.M.Lifshitz, and L.P.Pitaevskii, Electrodynamics of Continuous Media, second ed., Elsevier-
Butterworth-Heinemann, Oxford, 1984, eq. (1.7).
[9] J.C.Maxwell, Treatise on Electricity and Magnetism Vol. 1, third ed.(1891), reprint, Dover, New York, 1954, art. 161,
eq. (16).
[10] I.Pathak, A Generalization Of Gausss Theorem In Electrostatics, Proc. 2011 Annual Meeting of the Electrostatics
Society of America.
[11] W.K.H.Panofsky and M.Phillips, Classical Electricity and Magnetism, second ed., Addison-Wesley, Reading, MA,
1962, eq. (1-71).
[12] See, for example, reference [5], p. 186, eq. (4.39).
13
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
61
62
63
64
65
FIGURES
x
y
z
R

z r
Figure-1
d 1
R
i

d
S
i
q
d 1
S
d 1
S
2
S
2
S
d 2
R
Figure-2
14
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
61
62
63
64
65
a
a
a
Q
Figure-3
a a
a
a
a
a
Q
Figure-4
15
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
61
62
63
64
65
A
E(d)
E(d)
Figure-5
A
E(d)
E=0
Figure-6
16
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
61
62
63
64
65
r
r
q
P
O
x
y
z
h
Figure-7
O P
q

z
r
q
E
r
h
Figure-8
17
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
61
62
63
64
65
Figure Captions
Figure-1: The eld of a uniformly charged spherical surface
_
x
2
+ y
2
+ z
2
= R is to be evaluated at the point r z. Shown
here in the gure is the case where r is greater than R.
Figure-2: A cross section (of the closed surface S) showing the point charge q
i
and the inner solid angle
i
. The thick
solid curve is the outer surface of S and the thick dotted curves are the inner surfaces of S enclosing the shaded regions
which represent the cavities. The part of the outer surface drawn with a thin dotted curve underneath represents the
surface S
2
, while the remaining part on the left represents S
1
. The thin solid arc represents the segment S

of a spherical
surface of radius centered at q
i
.
Figure-3: A point charge Q lies at a corner of a cube. The ux through the shaded face is to be evaluated.
Figure-4: The same charge shown at the centre of a cube of sides of length twice as that of the original cube. In this
bigger cube, the original surface on which the ux was obtained, lies as a quadrant of one face.
Figure-5: A Gaussian pillbox of base area A straddling the innite at plane. The cuboid extends to a distance d on both
sides of the plane.
Figure-6: Gaussian surface obtained by removing half the volume of the original pillbox. This cuboid has the base of area
A lying on the plane and extends to a distance d only on one side of the plane.
Figure-7: A point charge q is held at a height h above an innite conducting plane. Here q lies at h z and the region z < 0
is lled with grounded conducting material. The induced surface charge density is to be evaluated at a general point P in
the plane.
Figure-8: The plane containing the z-axis and the point P. It can be seen from this gure that according to Coulombs
law

E
q
(

r ) is given by q(

r h z)/4
0
(r
2
+ h
2
)
3/2
.
18
Dear Reviewer,
I have read the reviewers review and found many questions, which the reviewer had
raised, helpful in improving the presentation of ideas. The issues raised in the review were subtle
and I have tried my best to resolve them in this revised manuscript. I am keen to learn if the
reviewer finds them appropriate. The point-by-point response to the reviewers comments
follows below:
In response to earlier critical reviews by this reviewer, the author has extended his
paper but this has not improved the content or clarity of the work. Perhaps it is a matter
of definition but the last sentence of the abstract is the essence of this paper's
contribution and it is incorrect.
Yes, the assertions of the paper may turn out to be incorrect if we define the net
charge contained by a set of points in a manner different from intuition. In addition, as I
do not think it meaningful to examine the situations, which may arise from such unusual
conventions that will go on affecting many major equations of science, I do not make any
change in my definitions of the net charge enclosed by a surface and the net charge
contained by a surface. Since I do not make any such non-standard definition, my
assertions are correct. For a uniformly charged spherical surface with surface charge per
unit area at r =R, I am considering the net charge contained by the sphere as
r Q
2
4 = , rather than Q/2. I invite the reviewer to suggest which set of words will
better mean what I want my words to mean as the reviewer may be remembering how my
lines in the abstract lead to the misunderstanding.
In particular, for the case of a uniformly surface charged sphere of radius R in free
space with surface charge per unit area sigma at r=R, the electric field E is zero inside
the sphere, r<R, and is radially directed with magnitude Q/(4*pi*epsilon0*r^2) outside
the sphere, r>R, where Q=sigma*4*pi*R^2.
Yes, it is true, as is known to the whole world. I concur with this remark.
In this example, the charge enclosed by the sphere is zero and the total charge on the
sphere surface is Q, rather than Q/2 as given in the abstract and paper.
In the case described in the abstract, the net charge enclosed is zero and the total
charge on the surface, i.e. the net charge contained by the surface, is r Q
2
4 = and the
flux through the surface is
0
2 Q/ . The simplest form of the Generalized Gausss
Theorem states
0 0
2
+

=
con enc
Q Q
. Here we have
enc
Q equal to zero and Q Q
con
= so that
0
2
=
Q
. It is not that the Generalized Gausss Theorem can wrongly be assumed to be
stating
0 0

+

=
con enc
Q Q
and we can go on rightly asserting
enc
Q to be zero and awkwardly
interpreting
con
Q by asserting 2 Q/ Q
con
= so as to conclude rightly that
0
2
=
Q
. The
statement
0 0

+

=
con enc
Q Q
is wrong as according to the intuitive definition of
enc
Q and
*Detailed Response to Reviewers
con
Q , we have
enc
Q equal to zero and Q Q
con
= . In the non-standard definition, which
somehow occurred to the reviewer and which states
enc
Q as equal to zero and 2 Q/ Q
con
= ,
the statement
0 0

+

=
con enc
Q Q
shall turn out to be correct. I could not second-guess that
such confusion could occur to anyones mind and after having pointed out this possibility
to me, the reviewer must have realized by now why the Generalized Gausss Theorem is
needed.
The paradoxical result most likely comes from not being definite on how the surface
charge is distributed on the surface of zero thickness. One can assume that the volume
charge density rho(r) is uniformly distributed from R to R+delta so that
rho(r)=3Q/[4*pi{(R+delta)^3-R^3}] and the total charge in the layer is Q. If we do not
allow any surface charge at r=R and r=(R+delta) the electric fields in the three regions
are:
E(r<R)=0
E(R<r<R+delta)=Q(r^3-R^3)/ [4*pi*epsilon*r^2*{(R+delta)^3-R^3}]
E(r>R+delta)=Q/(4*pi*epsilon*r^2)
where the electric fields are continuous at r=R, E(r=R)=0, and at r=R+delta,
E(r=R+delta)= Q/[4*pi*epsilon*( R+delta)^2]
However, if surface charge is allowed at either or both of E(r=R) and E(r=R+delta),
then the electric field within the layer is different.
These remarks have been analyzed in 2 beginning with the paragraph
following equation (10). I have compared the solutions above with my equations 1 5. In
the line following equation (3), I have written that
2
0
4 r
Q(r)
E(r) = and in the line
following equation (4), I have written that ) / ( ) R (r Q(r) = 3 4
3 3
, where
according to equation (1) is
) R ) ((R
Q
3 3
4
3
+
. So, E(r) comes out to be the same
E(R<r<R+delta) in the above lines from the review. Also when R r < , Q(r) is zero; and
when R r > , Q Q(r) = . Thus, we see that my equations are the same as the equations in
the above lines from the review, when the equations are compared.
) R E(r & ) R E(r ), R r E(R R), E(r + = + > + < < < are correctly evaluated in
the equations above, which come from your review. But, when we let tend to 0, R +
tends to R . But, the only meaningful quantities in the limit at 0 are
) R E(r and R) E(r + > < . ) R E(r and R) E(r ), R r E(R + = = + < < have values,
which must be used to analyze the flux through the surface and themselves have no
meaning in the limit at 0 . For the electrostatic field on the surface is discontinuous
and has no well-defined value on the surface.
In mathematical terms, the surface charge density is an impulse function of volume
charge density being zero everywhere except at r=R and r=R+delta and the volume
distribution can have any functional form between R and R+delta. This, I believe, is why
the author gets the erroneous result that the surface charge distribution is that of half the
value of the net charge contained by S. Depending on the assumptions at the interface,
any value of surface charge density can be obtained. The author should compare my
solutions above with his in eqs. 1-5.
These remarks have been analyzed next in 2 from the beginning of the last
paragraph of 2 to equation (11). That the assumptions regarding the surface charge
densities at R r = and R r + = do not alter the statement of the theorem has been
discussed prior to the last paragraph of 2. In the last paragraph, it is corroborated that
non-uniform spherically symmetric volume density functions, when made to tend to zero
thickness with the same net charge, also give the same Generalized Gausss Theorem,
which uniform spherically symmetric volume density gives. Physically thinking it seems
as convincing as a mathematical proof to me that if the result is what I have derived
modeling a surface charge by the limit of a uniform volume charge, then it will be
essentially the same - if surface charge can in principle be modeled as the limit of a
volume charge - when the surface charge Q is modeled by the limit of any spherically
symmetric non-uniform volume charge distribution of a net charge Q in the spherical
shell. I think it is convincing enough to state this line based on the belief that to a
physicist like me it would feel exhilarating and astonishing to learn that an expert
operator of this tool called integral calculus has turned him down, by proving it incorrect
by means of a proof by contradiction or by an example, when he was sure of his assertion
based on sound physical arguments.
I would like to see this paper published if it correctly describes these issues, but in its
present form, I don't believe it is generally correct, only valid for assumptions, which are
not specified. In addition, I believe that the author is often unclear and so I list some
suggestions for improvement below
Even I shall be happy to see this paper correctly describe the issues, if it
wrongly describes any, and I am trying to alter the presentation so that no reader may
form incorrect understanding of the assertions in my paper, which I currently believe are
all correct. However, I will be happier to learn if my theoretical beliefs are wrong, if any
of them actually are, so that the journal, if it wishes to present my ideas, presents correct
ideas to the wider audience. I firmly believe and have always believed that a series of
questions and answers and counter-questions can always finally lead to a correct idea on
which both the asserting side and believing side are unanimous. I will be happy to see
this paper published once nobody doubts that it presents only correct ideas as can be
seen at this time. In this particular revision, I have happily made all the changes
suggested by the reviewer to the best of my abilities and the only boundary condition on
my efforts was the belief that no assertion in the paper or its reviews especially the
incorrect ones shall by any means be cut and dried. By substantial healthy discourse,
the correct ideas should be preserved and the incorrect ones should be dropped. In my
revised manuscript, just as in the original manuscript, I had pondered only over the
correctness of ideas and tried my best to get the methodologies right. However, I do not
remember having given even one word a second thought for the use of a synonym or
having tried to improve the wordiness. It seems wise to me to assume that the reviewer
can contend with the issue better. No issues of possibilities of enhancement of theoretical
completeness that I had overlooked were pointed out and I shall not change them now. At
the same time some presentation issues, which I had missed, were pointed out. I happily
make all these suggested changes, which concern the choice of words and do not at the
same time make incorrect scientific ideas enter the paper. For the sake of mere
completeness, I stipulate here that I have preferred to choose to make an assumption that
it is free from errors to state that the suggestions of the reviewer concerning the selection
of words only enhance the presentation of ideas.
Throughout the paper the permittivity of the outside medium is that of free space,
epsilon0. However, the results are equally valid if the outside medium has a general
permittivity of epsilon
Yes, I have considered microscopic electrostatics until 6 and do not wish to
make the discussions prior to that section count only free charges and forget out bound
charges. When I want to talk of Gausss Theorem for the electrostatic field and consider
both free charges and bound charges as simply charges without even distinguishing
between different kinds of media and without considering all the charges, in any medium,
not just as charges in all space, I have to consider permittivity of free space. Even if the
outside medium is filled with a dielectric, Gausss Theorem or the Generalized Gausss
Theorem (whichever is applicable) will contain only permittivity of free space as by Q in
enc
Q or
con
Q we are referring to the net charge, which is not just the net free charge but
the net free charge plus the net bound charge. This is just what anyone understands
whenever microscopic electrostatics is discussed. The results using the permittivity of the
medium are valid. I know that but since I have seen many great scientists warn me as
textbook authors that including the permittivity of the medium doesnt generalize the
presentation and considering permittivity of free space is no less an insight, I have to use
it everywhere until 6. It is only in 7 that free charges and bound charges begin to be
considered separately and the flux of the electric displacement is considered. Thus,
susceptibility and the actual permittivity of the dielectric media are considered only since
then. The flux of the electrostatic field, even in case of linear media, is given by Gausss
Theorem or the Generalized Gausss Theorem (whichever is applicable) and contains the
permittivity of free space, as the electrostatic field is due in parts to both free and bound
charges, which are both called just charges as far as microscopic electrostatics is
concerned.
First paragraph-It is quite ordinary for problems to have surface charge on interfaces,
particularly cylindrical and spherical interfaces, especially at interfaces of dissimilar
electrical properties such as permittivity or ohmic conductivity
Yes, it is quite ordinary and I agree with the remark that that it is a fact.
However, as is included in the statement of Gausss Theorem that the closed surface of
integration must not contain any charges, we cannot apply Gausss Theorem over any
surface containing a part or whole of the interface. In addition, since until now no one
tried arriving at a Generalized Gausss Theorem for calculating flux over any surface
containing charges, the situations that need the Generalized Gausss Theorem are not
quite ordinary.
Section 1, second paragraph, first sentence-Change the first 6 words to "One faces a
problem with Gauss's."; last sentence on page 1- When volume charge is on a surface it
really means that it has some finite thickness. If it has zero thickness then it is better to
describe it as a surface charge density with units of coulombs/meter^2 rather than as a
volume charge density with units of coulombs/meter^3. Another way to write it is:
rho(x)=sigma*delta(x) where Rho(x) is the volume charge density, sigma is the surface
charge density (charge per unit area) and delta(x) is the Dirac Delta impulse function
with units of 1/distance
The first six words of the sentence, which was pointed out, have been changed,
as suggested. Whenever I say, when a volume charge distribution exists on a surface, I
do not mean the assertion when a volume charge distribution exists nowhere except on
a surface. A surface cannot have a finite thickness and by definition the only finite
value, which the thickness of a surface can take is zero. Since I mean a volume charge
distribution to exist on a surface, its volume density is an ordinary mathematical function
taking finite values and there is no need to describe its density using a Dirac Delta
volume density function. By definition, a volume charge distribution having finite
volume charge density cannot reside only on a surface and nowhere else in the proximity
of the surface. Volume charge must have some finite thickness and a surface does not. A
nice example will be the problem of evaluation of the flux of a volume charge
distribution

< <

=
b r
b r a
a r
for
for
for
r
k
(r)
0
0

over the surface r =R, where a <R <b. A volume charge exists on the surface r =R and
the net charge on the surface remains zero and Gausss Theorem can be applied.
Section 2, line 1-delete "the" before Gauss's Theorem
This change has been made.
The author identifies unit vectors with a caret above the variable, such as z, but does not
define this convention. It should be mentioned in the first paragraph of section 2. Often
the z with a caret is used rather than r with a caret and yet the result is valid for any
radial position and not just that along the z-axis
The convention has now been mentioned in the first paragraph of 2. That the
result is valid for any choice of the z-axis since the situation exhibits spherical
symmetry, has been mentioned now.
Line after eq. 15-Correct "already" to "ready".
This correction has been made. Equation (15) appears as equation (16) in this
revised manuscript.
Fig. 1 caption-Change r and R at the end of the second sentence to italics
The required changes have been made.

I think that I have resolved all the issues, which the reviewer had raised. Nevertheless, I am open
to any further suggestions, which can help improve the manuscript still. I even welcome
criticisms that may seem scientifically important to the reviewer.

Sincerely,
Ishnath Pathak.
HIGHLIGHTS

Gausss theorem applies only for a surface free of any charges
Here the idea is that it can be generalized
The generalized theorem only contains some new terms in the equation
It is shown that for most of the cases only one missing term appears
The equation
0 enc
S
/ Q =

da E is changed to
0
con
0
enc
S

Q
2
1

Q
+ =

da E
This equation is shown to have wide applicability
*Highlights (for review)