182 5. Further Determinant Theory
LetH
∗
n− 1
andK
∗
n− 1
denote the determinants obtained fromHn− 1 and
Kn− 1 , respectively, by again replacing eachφrby (φr−xφr+1) and by
replacing eachψrby (ψr−xψr+1). In the notation of the second and
fourth lines of (5.2.6),
H
∗
n− 1 =
∣
∣φ
m−xφm+1
∣
∣
n
, 1 ≤m≤ 2 n− 3 ,
K
∗
n− 1
=
∣
∣ψ
m−xψm+1
∣
∣
n
, 1 ≤m≤ 2 n− 3. (5.2.18)
Lemma 5.4.
a.
n
∑
i=1
H
(n)
in
x
n−i
=H
∗
n− 1
,
b.
n
∑
i=1
K
(n)
in
x
n−i
=K
∗
n− 1
.
Proof of (a).
n
∑
i=1
H
(n)
in
x
n−i
=
∣ ∣ ∣ ∣ ∣ ∣ ∣ ∣ ∣
φ 1 φ 2 ··· φn− 1 x
n− 1
φ 2 φ 3 ··· φn x
n− 2
...............................
φn− 1 φn ··· φ 2 n− 3 x
φn φn+1 ··· φ 2 n− 2 1
∣ ∣ ∣ ∣ ∣ ∣ ∣ ∣ ∣ n
.
The result follows by eliminating thex’s from the last column by means of
the row operations:
R
′
i
=Ri−xRi+1, 1 ≤i≤n− 1.
Part (b) is proved in a similar manner.
Lemma 5.5.
(−1)
i+1
Pf
(n)
i
=
∑n
j=1
H
(n)
jn
K
(n)
i−j+1,n
, 1 ≤i≤ 2 n− 1.
SinceK
(n)
mn=0whenm<^1 and whenm>n, the true upper limit in the
sum isi, but it is convenient to retainnin order to simplify the analysis
involved in its application.
Proof. It follows from the inductive assumption (5.2.10) that
Pf
∗
n− 1 =H
∗
n− 1 K
∗
n− 1. (5.2.19)
Hence, applying Lemmas 5.3 and 5.4,
2 n− 1
∑
i=1
(−1)
i+1
Pf
(n)
i x
2 n−i− 1
=
[
n
∑
i=1
H
(n)
inx
n−i
][
n
∑
s=1
K
(n)
snx
n−s
]
=
n
∑
j=1
n
∑
s=1
H
(n)
jnK
(n)
snx
2 n−j−s
[
s=i−j+1
s→i