162 4. Particular Determinants
EvaluatingEnfor small values ofn, it is found that
E 1 =
x
1 −x
,E 2 =
1!
2
x
3
(1−x)
4
,E 3 =
[1! 2!]
2
x
6
(1−x)
9
. (4.12.42)
The solution which satisfies (4.12.41) and (4.12.42) is as given in the the-
orem. It is now a simple exercise to evaluateFnandGn.Gnis found in
terms ofEn+1by replacingnbyn+ 1 in (4.12.32) and thenFnis given
in terms ofGnby (4.12.40). The results are as given in the theorem. The
proof of the formula forHnfollows from (4.12.14).
Hn=|Sm|n
=|(1−x)
m+1
ψm|n
=(1−x)
n
2
|ψm|n
=(1−x)
n
2
En. (4.12.43)
The given formula follows. The formula forJnis proved as follows:
Jn=|Am|n.
Since
A 0 =1=(1−x)(ψ 0 +1), (4.12.44)
it follows by applying the second line of (4.12.31) that
Jn=
∣ ∣ ∣ ∣ ∣ ∣ ∣
(1−x)(ψ 0 +1) (1−x)
2
ψ 1 (1−x)
3
ψ 2 ···
(1−x)
2
ψ 1 (1−x)
3
ψ 2 (1−x)
4
ψ 3 ···
(1−x)
3
ψ 2 (1−x)
4
ψ 3 (1−x)
5
ψ 4 ···
∣ ∣ ∣ ∣ ∣ ∣ ∣ n
=(1−x)
n
2
∣ ∣ ∣ ∣ ∣ ∣ ∣
ψ 0 +1 ψ 1 ψ 2 ···
ψ 1 ψ 2 ψ 3 ···
ψ 2 ψ 3 ψ 4 ···
∣ ∣ ∣ ∣ ∣ ∣ ∣ n
=(1−x)
n
2
x
−n
En
=x
−n
Hn, (4.12.45)
which yields the given formula and completes the proofs of all five parts of
Lawden’s theorem.
4.12.3 A Further Generalization of the Geometric Series
LetAndenote the Hankel–Wronskian defined as
An=|D
i+j− 2
f|n,D=
d
dt