200 5. Further Determinant Theory
=U
2
−V
2
+
∑
r
∂S
∂xr
, (5.4.42)
where
S=
∑
i,j
cicjE
ij
. (5.4.43)
This function is identical to the left-hand side of (5.4.26). Let
T=
∑
i,j,r,s
(ci+cj)(cr+cs)E
is
E
rj
. (5.4.44)
Then, applying (5.4.6),
T=
∑
i,s
ciE
is
∑
j,r
crE
rj
+
∑
j,r
E
rj
∑
i,s
cicsE
is
+
∑
i,s
E
is
∑
j,r
cjcrE
rj
+
∑
j,r
cjE
rj
∑
i,s
csE
is
=(U+V)
2
+2S
∑
r,s
E
rs
+(U−V)
2
=2(U
2
+V
2
)+2S
∑
r
E
rr
. (5.4.45)
EliminatingVfrom (5.4.42),
T+2R=4U
2
+2
∑
r
(
E
rr
+
∂
∂xr
)
S. (5.4.46)
To obtain a formula forQ, multiply (5.4.11) by (ci+cj), sum overiand
j, and apply (5.4.13) with the modifications (i, j)↔(r, s) on the left and
(i, r)→(r, s) on the right:
Q=
∑
i,j,r,s
(ci+cj)(cr+cs)E
is
E
rj
− 2
∑
i,j,r
cr(ci+cj)E
ir
E
rj
=T− 2
∑
r
cr
∑
i,j
(ci+cj)E
ir
E
rj
=T− 4
∑
r,s
crcsE
rs
E
sr
. (5.4.47)
EliminatingTfrom (5.4.46) and applying (5.4.26) and the fourth and sixth
lines of (5.4.4),
Q+2R=4
∑
r,s
crcs(E
rr
E
ss
−E
sr
E
rs
)
+2
∑
r
(
E
rr
+
∂
∂xr
)
[
∑
s
c
2
s
E
ss
−
1
3
∑
s,t,u
E
stu,stu