a

r
X
i
v
:
g
r
-
q
c
/
9
2
0
7
0
0
9



2
7

J
u
l

1
9
9
2
Gravitational Waves from an Axisymmetric Source in the Nonsymmetric
Gravitational Theory
J. W. Moat and D. C. Tatarski
Department of Physics
University of Toronto
Toronto, Ontario M5S 1A7, Canada
Abstract
We examine gravitational waves in an isolated axisymmetric reexion symmetric NGT system.
The structure of the vacuum eld equations is analyzed and the exact solutions for the eld variables
in the metric tensor are found in the form of expansions in powers of a radial coordinate. We nd
that in the NGT axially symmetric case the mass of the system remains constant only if the system
is static (as it necessarily is in the case of spherical symmetry). If the system radiates, then the
mass decreases monotonically and the energy ux associated with waves is positive.
Revised version, June 1992
UTPT-92-01, gr-qc/9207009
1 Introduction
The present work examines gravitational radiation in the Nonsymmetric Gravitational Theory
(NGT) (for a recent detailed review see [1]). We probe the NGT asymptotic behaviour in the wave
zone using an exact solution.
In General Relativity (GR) gravitational radiation from bounded sources has been studied not
only through the linearized theory but also with the use of exact solutions. The latter was done
for the general case of a bounded source in asymptotically at spacetime [2]. It was found that
conning the arguments to the axially symmetric case did not cause any essential loss of generality.
Since even the relevant GR calculations are very tedious and the level of technical diculty in
the case of NGT increases considerably, we limit ourselves to the axisymmetric case. The GR
gravitational waves from isolated axially symmetric reexionsymmetric systems were studied in
detail in [3]. Since our treatment of the axisymmetric NGT case is rather parallel, familiarity with
this analysis is strongly recommended.
Since the NGT was introduced [4] there have been few analytic solutions of the eld equations
published. The exact solutions known to date include the spherically symmetric vacuum case [5],
the spherically symmetric interior case [6, 7] and Bianchi type I cosmological solutions with and
without matter [8, 9].
This, at least in part, follows from the fact that deriving NGT eld equations relevant for
particular cases of interest is not as technically simple as may be suggested by its supercial
similarity to the corresponding GR situations. Firstly, since the underlying geometry is non-
Riemannian, neither the fundamental metric tensor g

nor the ane connection is symmetric.

This does not constitute a serious problem for the choice of the form of g

, since we can always

assume that its nonsymmetric part takes on the isometries of the symmetric part, which in turn
has a well dened GR limit. On the other hand, calculating the connection coecients proves
to be a tedious and time consuming exercise, independently of the method chosen. Secondly, the
resultant formulae for the nonsymmetric connection coecients are extremely complicated for all
but the simplest forms of the metric, thus becoming unwieldy to use in the derivation of still
more complicated eld equations.
The NGT quantities presented in this paper were derived with the use of symbolic algebraic
computation procedures. To this end we have used the symbolic computation system Maple.
1
In Section 2 we briey summarize the necessary fundamentals of NGT. Section 3 deals with
a coordinate system and generalization of the GR metric to the NGT case. Then in Section 4 we
expand the metric in negative powers of a suitably chosen radial coordinate and analyze the eld
equations. The closing section contains conclusions.
Due to their unwieldiness, we retain most of the formulae in the appendices and present
the actual calculations only if their tediousness is not forbidding. However, all the results and
intermediate steps of the calculations can be made available to the interested reader. In the case
of extremely unwieldy quantities the most advisable form of doing that would be the transfer of
computer les in the internal Maple format.
Throughout this paper we use units in which G = c = 1.
2 NGT Vacuum Field Equations
The NGT Lagrangian without sources takes the form:
L =

gg

(W), (1)
with g the determinant of g

. The NGT Ricci tensor is dened as:

R

(W) = W

,

1
2
(W

,
+ W

,
) W

+ W

, (2)
and W

is an unconstrained nonsymmetric connection :

W

= W

()
+ W

[]
. (3)
(Throughout this paper parentheses and square brackets enclosing indices stand for symmetrization
and antisymmetrization, respectively.) The contravariant nonsymmetric tensor g

is dened in
terms of the equation:
g

= g

. (4)
If we dene the torsion vector as:
W

[]
=
1
2
_
W

_
, (5)
then the connection:

= W

+
2
3

(6)
2
is torsion free:

[]
= 0. (7)
Dening now:
R

() =

,

1
2
(

(),
+

(),
)

()

, (8)
we can write:
R

(W) = R

() +
2
3
W
[,]
, (9)
where W
[,]
=
1
2
(W
,
W
,
). Finally, the NGT vacuum eld equations can be expressed as:
g
,
g

= 0, (10a)
(

gg
[]
)
,
= 0, (10b)
R

() =
2
3
W
[,]
. (10c)
For the purpose of the analysis of Section 4, it is convenient to decompose R

into standard
symmetric and antisymmetric parts: R
()
, R
[]
, and then rewrite the eld equation (10c) in the
following form:
R
()
() = 0, (11a)
R
[,]
() = 0, (11b)
where we used equations (6), (7) and the notation:
R
[,]
= R
[]
,
+R
[]
,
+R
[]
,
. (12)
3 The Metric
Similarly to GR, the simplest NGT eld due to a bounded source would be spherically symmetric.
However, the NGT equivalent of Birkhos theorem (see e.g. [1]) shows that a spherically symmetric
gravitational eld in an empty space must be static. Hence, no gravitational radiation escapes into
empty space from a pulsating spherically symmetric source.
Following [3] we consider the next simplest case: the eld which was initially static and spher-
ically symmetric and eventually becomes such, but undergoes an intermediate nonspherical wave
emitting period. Also, spacetime is assumed to be axially symmetric and reexionsymmetric at
all times. Even now, due to the complexity of the eld equations, we are forced to use the method
3
of expansion to examine the problem. This approach, namely expanding in negative powers of a
radial coordinate, was also used in the GR analysis [3] and seems to naturally suit a wave problem.
Due to the physical picture sketched above and to the fact that we are interested in the
asymptotic behaviour of the eld at spatial innity (in an arbitrary direction from our isolated
source) polar coordinates x
0
= u, x = (r, , ) are the natural choice. The retarded time u has
the property that the hypersurfaces u = const. are lightlike. Detailed discussion of the coordinate
systems permissible for investigation of outgoing gravitational waves from isolated systems can be
found in [2, 3].
The covariant GR metric tensor corresponding to the situation described above:
GR
g

=
_
_
_
_
_
_
_
_
_
V r
1
e
2
U
2
r
2
e
2
e
2
Ur
2
e
2
0
e
2
0 0 0
Ur
2
e
2
0 r
2
e
2
0
0 0 0 r
2
e
2
sin
2

_
_
_
_
_
_
_
_
_
, (13)
with U, V, , being functions of u, r and was rst given in [10].
For any metric in polar coordinates form conditions must be imposed in the neighbourhood
of the polar axis, sin = 0, to ensure regularity. In the case under consideration we have that, as
sin 0,
V, , U/ sin, / sin
2
(14)
each is a function of cos regular at cos = 1.
The NGT generalization of the metric tensor (13) is:
g

=
_
_
_
_
_
_
_
_
_
V r
1
e
2
U
2
r
2
e
2
e
2
+ Ur
2
e
2
+ 0
e
2
0 0 0
Ur
2
e
2
0 r
2
e
2
0
0 0 0 r
2
e
2
sin
2

_
_
_
_
_
_
_
_
_
, (15)
where and are functions of u, r and , and the eld equation (10b) imposes the following
conditions:
= 0 : 2
3
e
4
[2(1 2r
,r
) +r
,r
] = 0, (16a)
= 1 : e
2
[(cot 2
,
) +
,
]
4
+r
2
e
2
[(2
,u
+ 2U
,
U
,
U cot )
,u

,
]
+e
2
_
2(
,

,
)
2

2
(
,
+ cot ) +
,

_
+r
2
e
2()
(U
,
+ U cot )
3
= 0, (16b)
= 2 : e
2
(
,r
2
,r
) re
2
[(2U 2rU
,r
+ rU
,r
) + rU
,r
]
+e
2
_
2(
,r

,r
)
2

2

,r
+
,r
_
+re
2()
(U + rU
,r
)
3
= 0. (16c)
We rst solve the equation (16a) for , then substitute the solution into (16c) and solve it for .
The solutions are:
= e
2
_
1 +C

r
4
_
1/2
, (17a)
= e
2
(U +C

)r
2
_
1 + C

r
4
_
1/2
, (17b)
where C

(u, ) and C

(u, ) are functions of integration satisfying the condition following from the
equation (16b):
C
,u
C

C
,
+ 2C

(C
,
+ C

cot ) = 0. (18)
It may be noted here that the functions and must both be present in the metric (15) to ensure
its regularity. The case = 0, = 0 leads to a singular g

, while the case = 0, = 0 leads to a

solution for singular on the polar axis.
Since we consider an asymptotically at spacetime, it follows from (17b) that to satisfy the
boundary conditions we must have C

= 0. This in turn, due to the condition (18), restricts C

to
be a constant. To maintain a notation consistent with other NGT results, we write this constant
as:
C

= l
4
.
The skew components of the metric are now:
=
e
2
l
2
(l
4
+ r
4
)
1/2
, (19a)
=
e
2
Ur
2
l
2
(l
4
+ r
4
)
1/2
. (19b)
5
Finally, the NGT covariant metric tensor we work with takes the form:
g

=
_
_
_
_
_
_
_
_
_
_
_
V r
1
e
2
U
2
r
2
e
2
e
2
_
1 +
l
2
(l
4
+r
4
)
1/2
_
e
2
Ur
2
_
1 +
l
2
(l
4
+r
4
)
1/2
_
0
e
2
_
1
l
2
(l
4
+r
4
)
1/2
_
0 0 0
e
2
Ur
2
_
1
l
2
(l
4
+r
4
)
1/2
_
0 r
2
e
2
0
0 0 0 r
2
e
2
sin
2

_
_
_
_
_
_
_
_
_
_
_
.
(20)
The contravariant metric tensor obtained through (4) is given by:
g

=
_
_
_
_
_
_
_
_
_
_
_
0 e
2
_
1 +
l
2
(l
4
+r
4
)
1/2
_
1
0 0
e
2
_
1
l
2
(l
4
+r
4
)
1/2
_
1
V r
1
e
2 l
4
+r
4
r
4
+ U
2
e
4
e
2 l
4
r
2
e
2
U 0
0 e
2
U r
2
e
2
0
0 0 0 r
2
e
2
sin
2

_
_
_
_
_
_
_
_
_
_
_
.
(21)
4 The Field Equations
Ane connection components for the metric (20) are obtained by solving the system of 64 equations
(10a). The list of nonzero components is given in Appendix A. The Ricci tensor is then calculated
with the use of (8) and the standard decomposition of its nonzero components into symmetric and
antisymmetric parts is performed. The symmetric components R
()
relevant to our calculation
can be found in Appendix B. For the sake of convenience in comparing our results to the GR case
[3], we have split each R
()
into the GR part
GR
R

and the NGT contributions.

The only nonzero antisymmetric Ricci tensor components for the case at hand are R
[01]
, R
[02]
, R
[12]
.
However, the antisymmetric eld equation (11b):
R
[01,2]
= R
[01],2
+R
[12],0
+ R
[20],1
= 0, (22)
as will be demonstrated shortly, must be trivially satised as a consequence of the NGT Bianchi
identities. For this reason we do not present these skew components (except for their expanded
forms) in this paper.
The seven nonzero symmetric components of the Ricci tensor belong to three categories. The
symmetric eld equations (11a) for four of them:
R
11
= R
(12)
= R
22
= R
33
= 0, (23)
6
will be called following the terminology of [3] the main equations. The other two nontrivial
symmetric equations:
R
00
= R
(02)
= 0, (24)
we will call again following [3] the supplementary conditions. The remaining symmetric eld
equation
R
(01)
= 0
is, similarily to the equation (22), identically satised due to the Bianchi identities.
The generalized Bianchi identities of NGT can be written [11] (also [12]) in the form:

gg

(R
+;
R
++;
R
;
) = 0, (25)
where we used Einsteins notation:
A
+
;
= A

,
+ A

, (26a)
A

;
= A

,
+ A

. (26b)
We rewrite (25) in the form:

gg

_
R
,
R
,
R
,
+ 2R
()

_
= 0. (27)
If we now recall the form of the metric (21) and suppose that the four main equations (23) are
satised, then the Bianchi identities reduce to:
= 0 : g

R
(02)
g
(01)
R
00,1
g
11
R
(01),1
g
22
R
(02),2
+g
[12]
(R
(01),2
+ R
(02),1
) + g
(12)
(R
[01],2
+ R
[12],0
+R
[20],1
) = 0, (28a)
= 1 : g

R
(01)
= 0, (28b)
= 2 : g

R
(02)
+ g
[01]
(R
(01),2
+ R
(02),1
)
+g
(01)
(R
[01],2
+ R
[12],0
+R
[20],1
) = 0. (28c)
From (28b), we see that R
(01)
vanishes as a consequence of the main equations. Also, if supple-
mentary conditions are satised, then the only nontrivial antisymmetric eld equation (22) must
hold.
7
We now rewrite the main equations in the form:
0 = R
11
=
4r
4
l
4
+ r
4
__

,r
2
2


,r
r
_

l
4
r
2
(l
4
+r
4
)
_
, (29a)
0 = 2r
2
R
(12)
=
_
r
4
e
2()
U
,r
_
,r
2r
2
_
2
,r
(
,
cot )
2
,
r
+
,r

,r
_
(29b)

2l
4
l
4
+r
4
e
2()
r
_
3U(1 + 2r
,r
) rU
,r
2r
2
U
,r
(1 + r
,r
)
_
,
0 = R
22
e
2()
r
2
R
3
3
e
2
= 2V
,r
+
r
4
2
e
2()
U
,r
2
r
2
(U cot +U
,
)
,r
4r(U cot + U
,
) 2e
2()
[1 + (3
,

,
) cot
+ 2
,
(
,

,
)
,
+
,

,
2
_
(29c)

2l
4
l
4
+r
4
e
2()
r
2
_
U
2
(1 + 2r
,r
r
2

,r
2
) + rUU
,r
_
+
2l
4
l
4
+r
4
_
r(U cot + U
,
)
2V
r
_
,
0 = r
2
R
3
3
e
2
= 2r(r)
,ur
+ (1 r
,r
)(V
,r
rU
,
) (r
,rr
+
,r
)V
2e
2()
[1 + (3
,
2
,
) cot + 2
,
(
,

,
) +
,
]
+r
_
2(r
,r
+ )
,
+ (r
,r
3) cot
_
U
+r
2
(
,
cot )U
,r
(29d)

2l
4
l
4
+r
4
e
2
1
r
_

,u
+U(
,r
cot ) +
V
r
2
(1 r
,r
)
_
.
The structure of the above equations is analogous to the GR case. The rst three equations
include only derivatives on the hypersurface u = const. The last equation, however, contains
dierentiation with respect to u. Thus, if is given for some value of u, then (29a) determines
. In turn (29b) gives U and then V follows from (29c). Having all those functions, one can use
(29d) to nd at the later instant of u. One can then repeat the whole cycle. Thus, if is given
for some value of u, the equations (29) determine the temporal evolution of our system except for
functions of integration. The latter are arbitrary functions of u and , but are independent of r.
The equations (29) are independent. This is to be expected, since the number of independent
NGT eld equations must correspond to the number of independent eld variables, when the NGT
Bianchi identities are taken into account (see e.g. [13]). It may also be noted that the system (29)
reduces to its GR counterpart of ref. [3] on substitution l
2
= 0.
The analysis leading to determining the form of the functions U, V, , in our case duplicates
8
the one given in detail in [3]. The reason is straightforward. Upon inspection of the equations, we
nd that if the solutions U, V, , have forms of an expansion in powers of r, then the NGT terms
of the equations do not contribute to the leading terms. Further, the functions of integration in
(29) must remain unchanged, if we are to recover the GR limit upon setting l
2
= 0. This can be
easiest seen in the case of the equation (29a). If were to be the same in GR and NGT cases, then
the NGT solution for would have the form:

NGT
=
1
4
ln
_
C

(u, ) +
l
4
r
4
_
+
GR
=
1
4
ln
_
1 +
l
4
r
4
_
+
GR
,
where the second equation takes into account the boundary condition. The leading term in the
expansion of
GR
is of order r
2
and the highest order NGT contribution goes like r
4
. Therefore,
any NGT eect must fall o faster than the GR eects in the wave zone.
The reader interested in the details of determining the functions U, V, , is referred to [3].
Here we only summarize the results.
The requirement that the eld contain only outgoing radiation at large distances from the
source gives the form of :
=
f(t r)
r
+
g(t r)
r
2
+...
Substituting this into (29), isolating the functions of integration and applying suitable coordinate
transformations, gives the following:
=
1
4
c
2
r
2
+...
+
1
4
l
4
r
4

1
8
l
8
r
8
+..., (30a)
= cr
1
+
_
C
1
6
c
3
_
r
3
+..., (30b)
U = (c
,
+ 2c cot )r
2
+ (2N + 3cc
,
+ 4c
2
cot )r
3
+
_
3
2
C
,
+ 3C cot 3cN 4c
2
c
,
4c
3
cot
_
r
4
+...

2
3
(c
,
+ 2c cot )l
4
r
6
+..., (30c)
V = r 2M
_
N
,
+ N cot c
,
2
4cc
,
cot
1
2
c
2
(1 + 8 cot
2
)
_
r
1

1
2
[C
,
+ 3C
,
cot 2C + 6N(c
,
+ 2c cot )
+ 8c(c
,
2
+ 3cc
,
+ 2c
2
cot
2
)
_
r
2
+...
+
1
2
l
4
r
3

1
6
_
6M +
1
4
c
,
+
3
4
c
,
cot +c
_
l
4
r
4
+ ..., (30d)
9
where c(u, ), N(u, ), M(u, ) are the functions of integration and C(u, ) satises:
4C
,u
= 2c
2
c
,u
+ 2cM +N cot N
,
.
In (30), we isolated the NGT contributions from the rest of the expansion. Clearly, the leading
terms are purely GR ones.
One can readily verify the expansions (30) by a direct substitution into (29). The expanded
nonzero ane connection components are given in the Appendix C. Again, for the sake of conve-
nience of comparison to the GR results, we split the expansions into the NGT and GR parts. We
draw the readers attention to certain errors occurring in the expanded forms of
0
22
and
1
12
, as
published in [3].
The expansions (30) considerably simplify the form of the supplementary conditions (24).
These extremely complex equations (see Appendix B) reduce now to the leading inverse square
law terms involving only the relations of the functions c, M, N and the lower order terms including
pure NGT contributions and the coupling of the two. We see that, as we anticipated earlier from
the form of the equations (29), the NGT skew contributions decay rapidly with increasing distance
from the source and the dominant asymptotic behaviour of the system at spatial innity is given
by r
2
terms. From the latter, we get:
M
,u
= c
,u
2
+
1
2
(c
,
+ 3c
,
cot 2c)
,u
, (31a)
3N
,u
= M
,
+ 3cc
,u
+ 4cc
,u
cot + c
,u
c
,
, (31b)
in full agreement with the GR results. We note that the equations (31) are exact , if the series
expansions are valid.
To complete the discussion of the asymptotic behaviour of the NGT quantities, we expand the
only nonzero components of R
[]
:
R
[01]
= l
2
_
(4M + 4cc
,u
+c
,
+ 3c
,
cot 2c) r
5
+...
_
, (32a)
R
[02]
= l
2
_
3(c
,
+ 2c cot )
,u
r
3
(32b)
+[4c
,u
(c cot c
,
) + 3c(c
,
+ 2c cot ) + 3c
,
M
,
] r
4
+ ...
_
,
R
[12]
= l
2
_
3(c
,
+ 2c cot )r
4
2 [c(3c
,
+ 2c cot ) + 3N] r
5
+ ...
_
. (32c)
From (10c), it now follows that the torsion vector W

has the maximum asymptotic behaviour

1/r
3
. This clearly contradicts the result of [14] obtained through a perturbation expansion of the
10
NGT eld equations about a purely Einstein local vacuum background. The reader is referred to
[15] for a discussion of the inconsistencies of the latter result.
In order to identify M and N one has to turn to the static metrics. We once again refer
the reader interested in the details to [3]. After connecting the metric (20) to the static axially
symmetric metric in the Weyl form (through a lengthy transformation), one nds that in the static
case M(u, ) reduces to the mass m of the system and is independent of both its arguments. The
functions N and C are related to the dipole and quadrupole moments, respectively. Also, c is an
arbitrary function of in the static case. In the timedependent case the entire information about
the temporal evolution of the system is contained in c(u, ).
Since, in this paper, we are interested only in the physical interpretation of the function M
(N and C can be examined with the use of the above mentioned transformation), we limit our
analysis here to the following simple exercise. First let us write the nonzero components of the
metric tensor (20) using the expansions (30):
g
00
= 1
2M
r

1
r
2
(N
,
+N cot ) +...
+
l
4
r
4

l
4
r
5
_
2M +
1
8
c
,
cot +
1
6
c +
1
24
c
,
_
+ ..., (33a)
g
(01)
= 1
c
2
2r
2
+
1
2r
4
_
3Cc
c
4
2
_
+
l
4
2r
4
+ ..., (33b)
g
[01]
=
l
2
r
2

c
2
l
2
2r
4
+..., (33c)
g
(02)
= c
,
2c cot +
2N +cc
,
r
+..., (33d)
g
[02]
=
l
2
r
2
(c
,
+ 2c cot ) +
l
2
r
3
(2N +cc
,
) +..., (33e)
g
22
= r
2
c
2
2cr
1
r
_
2C
1
3
c
3
_
+..., (33f)
g
33
= sin
2

_
r
2
c
2
+ 2cr +
1
r
_
2C
1
3
c
3
_
+...
_
. (33g)
We conne ourselves to initial and nal static systems. Suppose that the system is in a dynamic
period, which can be physically interpreted only as the emission of waves, between the moments u
i
and u
f
. Thus, for u u
i
and u u
f
the function c does not vary. The static limit c(u, ) c
s
()
of (33) gives the equivalent of the static Weyl metric in our case. However, to demonstrate the
physical interpretation of M, we simplify the situation further. We can scale the function c for
either one of the static periods to be c = 0 (forsaking the dependence of c limits us here to a
static spherically symmetric system). We now remove the terms containing the functions N and C
11
from (33). We have a post factum justication for doing that from the interpretation of N and C
as multipole moments. Since there is no radiation during the static period, then N = C = 0. The
metric (33) tends now to its static spherically symmetric limit:
g
00
= 1
2M
s
r
+
l
4
r
4

2M
s
l
4
r
5
, (34a)
g
(01)
= 1 +
l
4
2r
4
, (34b)
g
[01]
=
l
2
r
2
, (34c)
g
(02)
= g
[02]
= 0, (34d)
g
22
= r
2
, (34e)
g
33
= r
2
sin
2
, (34f)
where by M
s
we denote the static limit of M.
Now a coordinate transformation fromour retarded time u to the usual time coordinate t = u+r
converts (34) into the NGT static spherically symmetric metric [1, 5]:
ds
2
= (1 +
l
4
r
4
)(1
2M
s
r
)dt
2
(1
2M
s
r
)
1
dr
2
r
2
_
d
2
+ sin
2
d
2
_
, (35)
with
g
[01]
=
l
2
r
2
.
Thus, the static spherically symmetric limit M
s
of the mass aspect M(u, ) can only be interpreted
as the mass of the system.
If we dene the mass m(u) of the system as the mean value of M(u, ) over the sphere:
m(u) =
1
2
_

0
M(u, ) sind, (36)
then c(u, ) completely determines the time evolution of the mass m(u). Integrating (31a) and
noticing that the second term in it does not contribute to the integral due to the conditions (14),
we get:
m
,u
=
dm
du
=
1
2
_

0
c
,u
2
sind. (37)
Since we discussed here systems whose initial and nal states are static, the physical interpretation
of m(u) as the mass of the system is unambiguous. Analogously to the GR case the main result is
as follows:
12
The mass of an axially symmetric NGT system is constant only if the system remains static.
If the system evolves in time (emits waves), the mass decreases monotonically.
Since radiation is the only energy loss mechanism available to the system, the above proves
that gravitational waves emitted by an axisymmetric reexion symmetric NGT source compatible
with the metric (20) carry positive energy or, in other words, the ux of gravitational energy in
NGT is positive.
5 Conclusions
We have proved that an NGT axisymmetric system emitting gravitational waves has the usual
GR-like asymptotic behaviour in the wave zone. The NGT contributions to the physical quantities
decay rapidly with the distance from the source and the energy ux at spatial innity is necessarily
positive.
Whether a similar result could be obtained in the general NGT case without any symmetries
remains to be veried, although it is expected that a proof of this would be similar to the one in
[2].
Acknowledgements
This work was supported by the Natural Sciences and Engineering Research Council of Canada.
We are grateful to N. Cornish and J. Spencer for helpful discussions.
13
A List of nonzero ane connection components

0
00
=
GR

0
00
+
4l
4
r
2
(l
4
+ r
4
)
_
U
2
r
3
e
2()
V
_
,
GR

0
00
= 2
,u
+r
2
e
2()
U
_
U
,r
+
U
r
+
,r
U
_

,r
V
r

V
,r
2r
+
V
2r
3
, (38)

0
01
=
2l
2
r(l
4
+ r
4
)
_
l
2

_
l
4
+r
4
_
, (39)

0
02
=
GR

0
02

l
2
r
l
4
+ r
4
_
3l
2
+
_
l
4
+r
4
_
Ue
2()
,
GR

0
02
=
,
r
2
e
2()
_
U
,r
2
+
U
r
+
,r
U
_
, (40)

0
10
=
2l
2
r(l
4
+ r
4
)
_
l
2
+
_
l
4
+r
4
_
, (41)

0
20
=
GR

0
02

l
2
r
l
4
+ r
4
_
3l
2

_
l
4
+r
4
_
Ue
2()
, (42)

0
22
=
GR

0
22
= re
2()
(1 +r
,r
) , (43)

0
33
=
GR

0
33
= re
2(+)
sin
2
(1 r
,r
) , (44)

1
00
=
GR

1
00
+
4l
4
r
3
(l
4
+ r
4
)
V
2
+
l
4
r(2l
4
+ r
4
)
UV
,
+
2l
4
r
3
l
4
+ r
4
U
4
e
4()
+
l
4
r
4
2l
4
+ r
4
U
3
U
,r
e
4()

3l
4
(3l
4
+r
4
)
(l
4
+ r
4
)(2l
4
+r
4
)
V U
2
e
2()
+
rl
4
2l
4
+ r
4
e
2()
_
2V U
2
( )
,r
+ U(U
,r
V UV
,r
)
+2rU
3
( )
,
+ 4rU
2
( )
,u
+ 2rU(U
,u
+UU
,
)
_
,
GR

1
00
=
V
,u
2r


,u
V
r

UV
,
2r


,
UV
r
+
V V
,r
2r
2
+

,r
V
2
r
2

V
2
2r
3
+r
2
e
2()
_
U
2
_
U
,
+
,
U
V
r
2


,r
V
r
+
,u
_

UU
,r
V
r
_
, (45)

1
01
=
GR

1
01
+
2l
4
l
4
+ r
4
_
2V
r
2
rU
2
e
2()
_
+
l
2

l
4
+ r
4
_
U
,

2V
r
2
_

l
2

l
4
+r
4
re
2()
_
l
4
l
4
+ r
4
U
2
+rU
2

,r
+
1
2
rUU
,r
_
,
GR

1
01
=
V
,r
2r

V
2r
2
+

,r
V
r

,
U
1
2
r
2
e
2()
UU
,r
, (46)

1
02
=
GR

1
02

l
4
2l
4
+r
4
V
,

l
4
2l
4
+ r
4
_
2
l
4
+ r
4
r
7
U
3
+
1
2
U
2
U
,r
_
e
4()
+
l
4
2(2l
4
+r
4
)
e
2()
_
UV [2r( )
,r
3] + r(UV
,r
U
,r
V )
+2r
2
[U
2
( )
,
+ U
,u
+ UU
,
2U( )
,u
]
_
14
+
l
4
(l
4
+ r
4
)(2l
4
+r
4
)
r
3
UV e
2()

l
2

l
4
+ r
4
2l
4
+ r
4
V
,
+
l
6
(2l
4
+ r
4
)

l
4
+r
4
r
4
U
2
(rU
,r
+ 2U) e
4()
+
l
2

l
4
+r
4
2l
4
+ r
4
e
2()
_
2r
3
[U(U
,
+U
,
) +U( 2)
,u
+U
,u
]
+r
2
(UV
,r
U
,r
V ) + 2r
2
UV
,r
rUV
_
+
l
6
(2l
4
+ r
4
)

l
4
+r
4
e
2()
_
2U(V +r
2
V
,r
) 2r
3
U(U
,

,u
)
_
+
2l
2

l
4
+r
4
rUV e
2()
+
4l
2
(l
4
+ r
4
)
3/2
r(2l
4
+ r
4
)
U
2

,
e
2()
,
GR

1
02
=
V
,
2r
+r
2
e
2()
_
U
_
V
r
2
+

,r
V
r

,u
U
,

,
U
_
+
U
,r
V
2r
_
, (47)

1
10
=
GR

1
01
+
2l
4
l
4
+ r
4
_
2V
r
2
rU
2
e
2()
_

l
2

l
4
+ r
4
_
U
,

2V
r
2
_
+
l
2

l
4
+r
4
re
2()
_
l
4
l
4
+ r
4
U
2
+rU
2

,r
+
1
2
rUU
,r
_
, (48)

1
11
=
GR

1
11
+
4l
4
(l
4
+r
4
)r
,
GR

1
11
= 2
,r
, (49)

1
12
=
GR

1
12
+ e
2()
Ur
_
l
2

l
4
+ r
4
r
,r

3l
4
l
4
+ r
4
_
,
GR

1
12
=
,
+
1
2
r
2
e
2()
U
,r
, (50)

1
20
=
GR

1
02

l
4
2l
4
+r
4
V
,

l
4
2l
4
+ r
4
_
2
l
4
+ r
4
r
7
U
3
+
1
2
U
2
U
,r
_
e
4()
+
l
4
2(2l
4
+r
4
)
e
2()
_
UV [2r( )
,r
3] + r(UV
,r
U
,r
V )
+2r
2
[U
2
( )
,
+ U
,u
+ UU
,
2U( )
,u
]
_
+
l
4
(l
4
+ r
4
)(2l
4
+r
4
)
r
3
UV e
2()
+
l
2

l
4
+ r
4
2l
4
+ r
4
V
,

l
6
(2l
4
+ r
4
)

l
4
+r
4
r
4
U
2
(rU
,r
+ 2U) e
4()

l
2

l
4
+r
4
2l
4
+ r
4
e
2()
_
2r
3
[U(U
,
+U
,
) +U( 2)
,u
+U
,u
]
+r
2
(UV
,r
U
,r
V ) + 2r
2
UV
,r
rUV
_

l
6
(2l
4
+ r
4
)

l
4
+r
4
e
2()
_
2U(V +r
2
V
,r
) 2r
3
U(U
,

,u
)
_
15

2l
2

l
4
+r
4
rUV e
2()

4l
2
(l
4
+ r
4
)
3/2
r(2l
4
+ r
4
)
U
2

,
e
2()
, (51)

1
21
=
GR

1
12
e
2()
Ur
_
l
2

l
4
+ r
4
r
,r

3l
4
l
4
+ r
4
_
, (52)

1
22
=
GR

1
22
+
l
4
l
4
+ r
4
U
2
r
3
e
4()
,
GR

1
22
= r
2
e
2()
_

,u
+U
,
+
,
U
V
r
2


,r
V
r
_
, (53)

1
33
=
GR

1
33
= r
2
sin
2
e
2(+)
_

,u
+U cot
,
U
V
r
2
+

,r
V
r
_
, (54)

2
00
=
GR

2
00

l
4
l
4
+ r
4
2UV
r
2

l
4
2l
4
+r
4
V
,
r
3
e
2()
+
_
l
4
l
4
+r
4
rU
3

l
4
2l
4
+ r
4
r
2
U
2
U
,r
_
e
2()

l
4
(2l
4
+ r
4
)r
2
_
3UV r(UV
,r
U
,r
V ) 2rUV ( )
,r
+4r
2
U( )
,u
+ 2r
2
U
2
( )
,
2r
2
(U
,u
+UU
,
)
_
,
GR

2
00
= U
,u
+U
_
2
,u
2
,u
U
,

,
U +
V
2r
2

V
,r
2r


,r
V
r
_
+e
2()
V
,
+ 2
,
V
2r
3
+r
2
e
2()
U
2
_
U
,r
+
U
r
+
,r
U
_
, (55)

2
01
=
GR

2
01
+
l
2

l
4
+r
4
2
,
r
2
e
2()

l
2

l
4
+r
4
_
2U
r
(1 + r
,r
) +U
,r
_

l
2

l
4
+r
4
_
l
2

l
4
+r
4
1
_
2
2U
r
,
GR

2
01
=
U
,r
2

U
r

,r
U +

,
r
2
e
2()
, (56)

2
02
=
GR

2
02

l
2

l
4
+r
4
_
V
,r
r

,u
U
,
+ V
_
+
l
2
2l
4
+ r
4
r
2
UU
,r
e
2()

l
2

l
4
+r
4
_
l
2

l
4
+r
4
+ 1
_
2
rU
2
e
2()
,
GR

2
02
=
,u
+
,
U r
2
e
2()
U
_
U
,r
2
+
U
r
+
,r
U
_
, (57)

2
10
=
GR

2
01

l
2

l
4
+r
4
2
,
r
2
e
2()
+
l
2

l
4
+r
4
_
2U
r
(1 + r
,r
) +U
,r
_
+
l
2

l
4
+r
4
_
l
2

l
4
+r
4
1
_
2
2U
r
, (58)

2
12
=
GR

2
12

l
2

l
4
+r
4
_
1
r
+
,r
_
,
16
GR

2
12
=
1
r
+
,r
, (59)

2
20
=
GR

2
02
+
l
2

l
4
+r
4
_
V
,r
r

,u
U
,
+ V
_

l
2
2l
4
+ r
4
r
2
UU
,r
e
2()
+
l
2

l
4
+r
4
_
l
2

l
4
+r
4
1
_
2
rU
2
e
2()
, (60)

2
21
=
GR

2
12
+
l
2

l
4
+r
4
_
1
r
+
,r
_
, (61)

2
22
=
GR

2
22
=
,
+r
2
e
2()
U
_
1
r
+
,r
_
, (62)

2
33
=
GR

2
33
= r
2
sin
2
e
2()
U
_
1
r

,r
_
e
4
sin
2
(cot
,
) , (63)

3
03
=
GR

3
03

l
2

l
4
+r
4
_

,u
+
_
1
r

,r
__
V
r
r
2
U
2
e
2()
__
,
GR

3
03
=
,u
, (64)

3
13
=
GR

3
13

l
2

l
4
+r
4
_
1
r

,r
_
,
GR

3
13
=
1
r

,r
, (65)

3
23
=
GR

3
23

l
2

l
4
+r
4
_
1
r

,r
_
r
2
Ue
2()
,
GR

3
23
= cot
,
, (66)

3
30
=
GR

3
03
+
l
2

l
4
+r
4
_

,u
+
_
1
r

,r
__
V
r
r
2
U
2
e
2()
__
, (67)

3
31
=
GR

3
13
+
l
2

l
4
+r
4
_
1
r

,r
_
, (68)

3
32
=
GR

3
23
+
l
2

l
4
+r
4
_
1
r

,r
_
r
2
Ue
2()
. (69)
17
B List of nonzero symmetric Ricci tensor components
R
00
=
GR
R
00

l
4
2l
4
+ r
4
e
2()
1
r
3
_
2V
,
( )
,
+ V
,
cot + V
,
_
+
l
4
r(2l
4
+r
4
)
_
2UV [( )
,
+ ( ) cot ]
,r
4r(U
,
+U cot )( )
,u
2r(2UU
,
+ U
2
cot )( )
,
+ 2r(U
,u
+ U
,
2
+UU
,
+2(U
,
V + 2UV
,
)( )
,r
+ (UV
,r
U
,r
V ) cot
+2r(U
,u
+ UU
,
) cot 4rU( )
,u
2rU
2
( )
,
(U
,r
V )
,
+U
,
V
,r
+ 2UV
,r
_

l
4
(l
4
+ r
4
)
2
l
4
+ 9r
4
r
4
V
2

l
4
(l
4
+ r
4
)(2l
4
+r
4
)
7l
4
+ 5r
4
r
2
(U cot + U
,
)V

l
4
(l
4
+ r
4
)(2l
4
+r
4
)
2
8l
8
+ 14l
4
r
4
+ 7r
8
r
2
UV
,

l
4
r
2
(l
4
+r
4
)
_
V
,u
2r
2

,u
2
+ 2V (2r
,u

,r

,u
V
,r
2
+U
,
)
+
V (2V
,r
3V
,r
)
r
_
+
l
4
2l
4
+ r
4
r
2
Ue
2()
_
6UU
,r
( )
,
+ 4
_
U( )
,u
_
,r
+ 2U
2
( )
,r
U(3U
,
+ U cot )
,r
2(U
,ur
+ 2U
,r
U
,
)
_

l
4
2l
4
+ r
4
rUe
2()
_
UV
,rr
U
,rr
V +
_
UV ( )
,r
_
,r
_
+
2l
4
l
4
+r
4
rU
2
e
2()
_
2r
,u

,r
+ U cot 2V
,r
2
_
+
2l
4
(l
4
+ r
4
)
2
(2l
4
+ r
4
)
2
_
r
12
+ 10l
12
+ 17l
4
r
8
+ 29l
8
r
4
_
V
r
U
2
e
2()

2l
4
(l
4
+ r
4
)
2
(2l
4
+ r
4
)
2
_
l
12
3r
12
+ l
4
r
8
+ 8l
8
r
4
_
r
,u
U
2
e
2()
+
2l
4
(l
4
+ r
4
)(2l
4
+r
4
)
2
_
2r
9
(UU
,
+ U
,u
2U
,u
) r(4l
8
+ 4l
4
r
4
r
8
)U
2

,
(4l
8
+ 2l
4
r
4
r
8
)UV
,r
+ (10l
8
+ 11l
4
r
4
+ 5r
8
)UV
,r
+l
4
(2l
4
+ 3r
4
)U
,r
V + (2l
8
+l
4
r
4
2r
8
)UV
,r
_
Ue
2()
+
16l
8
(2l
4
+ r
4
)
2
rU
3

,
e
2()

2l
4
l
4
+r
4
r
3
U
4
e
4()
_
2
,r
r
,r
2
_
+
_
2l
4
(2l
4
+ r
4
)
2
(6l
4
+ r
4
)r
3
U
,r

2l
4
(2l
4
+ r
4
)
(l
4
+r
4
)
2
rU
_
U
3
e
4()
18
+
l
4
2l
4
+ r
4
r
4
U
2
e
4()
_
UU
,rr
+ 2U
,r
2
2UU
,r
( )
,r
_
, (70)
GR
R
00
= 2U(
,u

,u
) 2U
,
(
,u

,u
) + U
,u
+UU
,
+U
,
2
2UU
,
(
,

,
) + U
2
(2
,
2
2
,

,
+
,
) + 2
,u
2
cot
_
2U(
,u

,u
) U
,u
UU
,
U
2

,
_
+ e
2()
_
4UU
,r
V +U
2
(V
,r
+V
,r
)
_

1
2r
(U
,r
V
,
U
,
V
,r
UV
,r
cot )
+
1
r
[2V (
,ur
+ U
,r
) + 2UV
,
(
,r

,r
) + U
,
V
,r
+U(V
,r

,
+ V
,r
+ V
,r
cot )]
1
2r
2
V (U
,
+ V
,rr
+U cot )

1
r
2
[2U(V
,
V
,
) + V (V
,rr
+ V
,r

,r
2
,u
) + V
,u
]

1
2r
3
e
2()
[V
,
+ 2V
,
+ (2
,
2
,
+ cot ) (V
,
+ 2V
,
)]
+ re
2()
[UV (U
,rr
2U
,r

,r
+ 2U
,r

,r
)
+U
2
(V
,rr
+ V
,r

,r
3U
,
2
,u
) U
3
(cot + 2
,
) +
U
2
,r
V
2
_
+ r
2
e
2()
_
2U
3

,r
+U
2
(2U
,r

,
3U
,r

,
U
,

,r
2
,ur
2U
,r
)
U
2
(U
,r
+ U
,r
) cot U(U
,ur
+ 2U
,r
U
,
) + 2UU
,r
(
,u

,u
)
_
+
1
2
r
4
e
4()
U
2
U
,r
2
, (71)
R
(01)
=
GR
R
01

l
4
l
4
+r
4
_
2U
,
r
+
2U
r
(
,
+ cot )
2V
,r
2
r
+ 2
,u

,r
+
2V
,r
3V
,r
r
2
+
V
r
3
_

8l
4
(l
4
+ r
4
)
2
rV
2l
8
(l
4
+r
4
)
2
U
2
e
2()

l
4
l
4
+r
4
e
2()
_
2U
2
(r
2

,r
2
2r
,r
1) +rUU
,r
_
, (72)
GR
R
01
= U(
,r

,r
)
_
U
,r
2
+ U
,r
_
cot U
,r

,
U
,

,r
2
,u

,r

U
,r
2
2
,ur
+
V
,r
r
2
+
1
r
2
e
2()
_
2
,
2
2
,

,
+
,
cot +
,
_

1
r
_
2U
,
+ U cot + U
,
+V
,rr
V
,r

,r

V
,rr
2
_
r
2
e
2()
_
UU
,r
_
2
r

,r
+
,r
_
+
UU
,rr
2
+
U
,r
2
2
_
, (73)
R
(02)
=
GR
R
02
+
4l
4
(2l
4
+r
4
)
2
r
2
V
,
+
l
4
(2l
4
+ r
4
)r
_
V
,
(
,r
2
,r
)
V
,r
2
_
19

2l
4
l
4
+ r
4
r
3
U
3
_
r
,r
2
2
,r
_
e
4()
+
2(2l
4
+r
4
)l
4
(l
4
+ r
4
)
2
r
2
U
3
e
4()

l
4
2l
4
+r
4
r
4
U
_
UU
,r
(
,r

,r
) + U
,r
2
+
UU
,rr
2
_
e
4()

_
4l
8
(2l
4
+r
4
)
2
+
l
8
(2l
4
+ r
4
)(l
4
+r
4
)
_
r
3
U
2
U
,r
e
4()

2l
4
l
4
+ r
4
e
2()
rU
_
r
,u

,r
V
,r
2
+U cot
_

l
4
2l
4
+r
4
r
_
U
,rr
V UV
,rr
2
U
,ur
UV ( )
,rr
+r
_
U
2
( )
,)
_
,r
+ 2r
_
U( )
,u
_
,r
r(UU
,r
+ U
,r
U
,
) (U
,r
V + UV
,r
)( )
,r
_
e
2()
+
l
4
(2l
4
+r
4
)
2
e
2()
_
(2l
4
+ 5r
4
)UV
,r
(2l
4
r
4
)U
,r
V 4r
5
U
2

,
_
+
l
4
(l
4
+r
4
)(2l
4
+ r
4
)
2
_
(4l
8
3r
8
)(2rU
,u
+rU
,u
2rU
,u
)
(10l
8
+ 15l
4
r
4
+ 7r
8
)UV
,r
+ (8l
8
+ 4l
4
r
4
r
8
)UV
,r
r(4l
8
+ 6l
4
r
4
+ 3r
8
)UU
,
_
e
2()

l
8
(13r
8
2l
8
+ 19l
4
r
4
)
r(l
4
+r
4
)(2l
4
+r
4
)
2
UV, (74)
GR
R
02
=
,u

,u
+ 2
,u
(
,
cot ) U(
,
+ 2
,
2
2
,

,
+
,
cot )
e
2()
(UV
,r
+ UV
,r
+ 2U
,r
V )
V
,r
2r
(
,r

,r
)
V
,
r
+
V
,
2r
2
+ re
2()
_
U (3U
,
+ 2
,r
V
,rr
V
,r

,r
) +U
2
(2
,
+ cot )

U
,rr
V
2
+ U
,r
V (
,r

,r
)
_
+r
2
e
2()
_
U
,ur
2
+ U
,r
U
,
U
,r
(
,u

,u
) +U
_
3
2
U
,r
+ 2
,ur
+ U
,

,r
+
U
,r
2
cot
_
UU
,r
(
,

,
) +U
2
(2
,r
+
,r
cot )
_

1
2
r
4
e
4()
UU
,r
2
, (75)
R
11
=
GR
R
11

4l
4
l
4
+r
4
_

,r
2
2


,r
r
_

4l
4
r
2
(l
4
+r
4
)
2
, (76)
GR
R
11
= 4
_

,r
2
2


,r
r
_
, (77)
R
(12)
=
GR
R
12
+
l
4
l
4
+r
4
e
2()
[2rU
,r
(r
,r
+ 1) 3U(2r
,r
+ 1) + rU
,r
] , (78)
20
GR
R
12
= 2
_

,r
(cot
,
) +

,
r
_

,r
+
,r
+ r
2
e
2()
_
U
,r

,r
U
,r

,r
+
2U
,r
r
+
U
,rr
2
_
, (79)
R
22
=
GR
R
22

6l
8
(l
4
+ r
4
)
2
U
2
r
2
e
4()
+
2l
4
l
4
+ r
4
r
2
_
U
2
(r
2

,r
2
2r
,r
1) rUU
,r
_
e
4()
+
2l
4
l
4
+ r
4
_
r(U
,
+ U
,
+
,u
) V
,r

V
r
_
e
2()
, (80)
GR
R
22
= 1 + 2
_

,

,
2

,
2

,
+

,
2
_
+ 3
,
cot e
2()
(V
,r
+ V
,r
)
+ re
2()
(2
,u
+ 2U
,
+U cot + 3U
,
V
,rr
V
,r

,r
)
+ r
2
e
2()
(2
,ur
+ U
,

,r
+U
,r
cot +U
,r
2U
,r
)
+ r
2
U
,r
2

,
r
4
e
4()
U
,r
2
2
, (81)
R
33
/ sin
2
=
GR
R
33
/ sin
2

2l
4
l
4
+ r
4
_
r(U
,
U cot +
,u
) V
,r
+
V
r
_
e
2()
, (82)
GR
R
33
/ sin
2
= e
2()
(V
,r
V
,r
)
+ e
4
_
1 + 2
,
(
,
cot ) + 3
,
cot +
,
2
,
2
_
+ re
2()
(U
,
2
,u
2U
,
+ 3U cot + V
,rr
+V
,r

,r
)
+ r
2
[(U
,r
U
,r
) cot U
,

,r
U
,r

,
2
,ur
2U
,r
] . (83)
21
C Expanded nonzero ane connection components

0
00
=
1
r
2
(M + cc
,u
)
1
r
3
(N cot +N
,
) +...

2l
4
r
5
+
l
4
2r
6
_
7M
5
2
_
c
3
+
c
,
cot
4
+
c
,
12
__
+ ..., (84)

0
01
=
2l
2
r
3

2l
4
r
5
+..., (85)

0
02
=
N
r
2
+
1
r
3
_
3C
,
2
+ Nc + 3C cot
_
+...
+
l
2
r
3
(c
,
+ 2c cot )
l
2
r
4
(2N + cc
,
) + ..., (86)

0
10
=
2l
2
r
3

2l
4
r
5
+..., (87)

0
20
=
N
r
2
+
1
r
3
_
3C
,
2
+ Nc + 3C cot
_
+...

l
2
r
3
(c
,
+ 2c cot ) +
l
2
r
4
(2N + cc
,
) + ..., (88)

0
22
= c +r
c
2
2r

1
r
2
_
c
3
3
+C
_
+ ...

l
4
2r
3

l
4
2r
4
c +
3l
4
8r
5
c
2
+..., (89)

0
33
/ sin
2
= r c +
c
2
2r

1
r
2
_
c
3
3
+C
_
+ ...

l
4
2r
3
+
l
4
2r
4
c +
3l
4
8r
5
c
2
+..., (90)

1
00
=
M
,u
r
+
1
r
2
_
M + cc
,u
(1 + 4 cot
2
)
1
2
(N cot +N
,
)
,u
+ c
,
c
,u
+ 2(c
,u
c
,
+cc
,u
) cot ] +...
+
2l
2
r
5

l
4
2r
5
_
M
,u
+
1
2
_
c
3
+
c
,
cot
4
+
c
,
12
_
,u
_
+ ..., (91)

1
01
=
M
r
2
+
1
r
3
(N cot +N
,
) +...
+
2l
2
r
3

4l
2
r
4
M +
2l
4
r
5
+..., (92)

1
02
=
1
r
[c
,u
(c
,
+ 2c cot ) M
,
]

1
2r
2
[2c
,u
(cc
,
+ 2N) + N + N
,
+ (N
,
N cot ) cot ] + ...

l
2
r
2
(c
,
+ 2c cot )
,u
(93)
+
l
2
r
3
[cc
,u
+ 2c cot (c
,u
2) + 2c
,
(c
,u
1) + 2N
,u
+M
,
] + ...,
22

1
10
=
M
r
2
+
1
r
3
(N cot +N
,
) +...

2l
2
r
3
+
4l
2
r
4
M +
2l
4
r
5
+..., (94)

1
11
=
c
2
r
3
+... +
2l
4
r
5
+..., (95)

1
12
=
1
r
(c
,
+ 2c cot )
1
r
2
(3N + 3cc
,
2c
2
cot ) + ...
+
l
2
r
4
_
cc
,
+ 2c
2
cot
_

l
2
r
5
c (cc
,
+ 2N) +..., (96)

1
20
=
1
r
[c
,u
(c
,
+ 2c cot ) M
,
]

1
2r
2
[2c
,u
(cc
,
+ 2N) + N + N
,
+ (N
,
N cot ) cot ] + ...
+
l
2
r
2
(c
,
+ 2c cot )
,u
(97)

l
2
r
3
[cc
,u
+ 2c cot (c
,u
2) + 2c
,
(c
,u
1) + 2N
,u
+M
,
] + ...,

1
21
=
1
r
(c
,
+ 2c cot )
1
r
2
(3N + 3cc
,
2c
2
cot ) + ...

l
2
r
4
_
cc
,
+ 2c
2
cot
_
+
l
2
r
5
c (cc
,
+ 2N) +..., (98)

1
22
= r(1 c
,u
) + 2M 2c
,
cot +c(1 + 2c
,u
+ 2 cot
2
) c
,
+ ...

l
4
2r
3
c
,u
+ ..., (99)

1
33
/ sin
2
= r(1 + c
,u
) + 2M +c(1 + 2c
,u
2 cot
2
) c
,
cot +...
+
l
4
2r
3
c
,u
+ ..., (100)

2
00
=
1
r
2
(c
,
+ 2c cot )
,u

1
r
3
(2N
,u
+ 3cc
,u
+M
,
+ 4cc
,u
cot + c
,u
c
,
) + ...

4l
4
3r
6
(c
,
+ 2c cot )
,u
+ ..., (101)

2
01
=
N
r
4
+... +
l
2
r
5
(c
,
+ 2c cot )
,u
+ ..., (102)

2
02
=
c
,u
r

1
2r
3
(c
2
c
,u
2C
,u
) +...
+
l
2
r
3
(c
,u
1) +
l
2
r
4
(2M +c) +..., (103)

2
10
=
N
r
4
+...
l
2
r
5
(c
,
+ 2c cot )
,u
+ ..., (104)

2
12
=
1
r

c
r
2

1
2r
4
(6C c
2
) + ...

l
2
r
3
+
l
2
r
4
c + ..., (105)
23

2
20
=
c
,u
r

1
2r
3
(c
2
c
,u
2C
,u
) +...

l
2
r
3
(c
,u
1)
l
2
r
4
(2M +c) +..., (106)

2
21
=
1
r

c
r
2

1
2r
4
(6C c
2
) + ...
+
l
2
r
3

l
2
r
4
c + ..., (107)

2
22
=
2
r
c cot +
2
r
2
(N + cc
,
+ c
2
cot ) + ...

l
4
6r
5
(c
,
+ 2c cot ) + ..., (108)

2
33
= sin cos (1
2c
r

2
r
2
N cot +...)

l
4
3r
5
sin
2
(c
,
+ 2c cot ) + ..., (109)

3
03
=
c
,u
r

C
,u
r
3
+
c
2
c
,u
2r
3
+ ...

l
2
r
3
(c
,u
+ 1) +
l
2
r
4
(2M c) +..., (110)

3
13
=
1
r
+
c
r
2
+
3C
r
4

c
3
2r
4
+...

l
2
r
3

l
2
r
4
c + ..., (111)

3
23
= cot
c
,
r

C
,
r
3
+
c
2
c
,
2r
3
+...
+
l
2
r
3
(c
,
+ 2c cot ) +
2l
2
r
4
(N +c
2
cot ) +..., (112)

3
30
=
c
,u
r

C
,u
r
3
+
c
2
c
,u
2r
3
+ ...
+
l
2
r
3
(c
,u
+ 1)
l
2
r
4
(2M c) +..., (113)

3
31
=
1
r
+
c
r
2
+
3C
r
4

c
3
2r
4
+...
+
l
2
r
3
+
l
2
r
4
c + ..., (114)

3
32
= cot
c
,
r

C
,
r
3
+
c
2
c
,
2r
3
+...

l
2
r
3
(c
,
+ 2c cot )
2l
2
r
4
(N +c
2
cot ) +..., (115)
24
References
[1] Moat J. W. in Gravitation A Ban Summer Institute, eds. Mann R. B. & Wesson P.,
World Scientic, Singapore, 1991, p. 523.
[2] Sachs R. K., Proc. Roy. Soc. A, 270, 103, 1962.
[3] Bondi H., van der Burg M. G. J. & Metzner A. W. K., Proc. Roy. Soc. A, 269, 21, 1962.
[4] Moat J. W., Phys. Rev. D 19, 3554, 1979.
[5] Moat J. W., Phys. Rev. D 19, 3562, 1979.
[6] Savaria P., Class. Quantum Grav. 6, 1003, 1989.
[7] Savaria P., Class. Quantum Grav., to be published.
[8] Kunstatter G., Moat J. W. & Savaria P., Phys. Rev. D 19, 3559, 1979.
[9] Kunstatter G., Moat J. W. & Savaria P., Can. J. Phys. 58, 729, 1980.
[10] Bondi H., Nature, 186, 535, 1960.
[11] Einstein A., Can. J. Math. 2, 120, 1950.
[12] Kunstatter G. & Moat J. W., Phys. Rev. D 19, 1084, 1979.
[13] Moat J. W., J. Math. Phys. 21, 1798, 1980.
[14] Damour T., Deser S. & McCarthy J., Phys. Rev. D 45, R3289, 1992.
[15] Cornish N. J. & Moat J. W., University of Toronto preprint UTPT-91-33, 1991. (revised
version, June 1992).
25