254 23. HEAT KERNEL FOR STATIC METRICS
we have
(23.121)
.. 8 (al/2)
-lJ.
8xJ
= (bij + ~Ripqj (y) xPxq + 0 (r^3 ))
x (~Rjq (y) xq +
2
~ (V'jRpq + 2\i'pRjq) (y) xPxq + 0 (r (x)^3 ))
= ~Rir (y) xr +
2
1
4
(Y'iRpq + 2\i'p~q) (y) xPxq + 0 (r (x)^3 ).
Thus, by (23.119) and by a-^1 a~i applied to (23.121) with a~i 0 (r (x)^3 ) =
0 (r (x)^2 ), we have
.6.x (a-1/2) = -a-1 _!!_, (gij 8 ( al~2))
8xi 8xJ
-1 -1
= a
6
~i + ~
2
(23.122) (2\i'iRiq + V' qRii) xq + 0 (r (x)^2 ),
where the Ricci tensors and their derivatives on the RHS are evaluated at y
and where we used
1 8
24
(\i'iRpq + 2\i'pRiq)
8
xi (xPxq)
1 1
=
24
(\i'iRiq + 2\i'iRiq) xq +
24
(Y'iilpi + 2\i' pRii) xP
1
=
12
(2\i'iRiq + V'qRii) xq.
LEMMA 23.32. The norm squared of the gradient and the Laplacian of
</>o = a-^1 /^2 are given by
(1)
(23.123)
(2)
(23.124)
I" ,.1-. 12 ( ) _ ij 8</>o 8</>o _^0 ( ( )
(^2) )
V Xl/-'0 X - g 8xi 8Xj - r X l
(.6.x</>o) (x) = .6.x ( a-^1!^2 ) (x) = R ~y) + ~ (Y'rR) (y) xr + 0 (r (x)^2 ).
PROOF. (1) By (23.120) we have
8</>o _ 1 q ( 2)
8 xj - f,Riq(y)x +0 r(x).
(2) This follows from applying to (23.122) the equations a= 1+0 (r^2 ),
gij = bij + 0 (r^2 ), and gij\i'iRjk = !VkR· D