6.4 The Kay–Moses Equation 249
=− 4 c(c−1)α
2
1
. (6.3.20)
Referring to (6.3.16),
(
y+4n
2
1+x
)′
=4
[
(c−1)xA
11
+n
2
1+x
]′
=− 4 cλ
2
. (6.3.21)
Differentiating (6.3.19) and using (6.3.17),
y
′′
=8c(c−1)λα 1 β 1. (6.3.22)
The theorem follows from (6.3.19) and (6.3.22).
6.4 The Kay–Moses Equation...................
Theorem.The Kay–Moses equation, namely
[
D
2
+ε
2
+2D
2
(logA)
]
y= 0 (6.4.1)
is satisfied by the equation
y=e
−ωεx
1 −
n
∑
i,j=1
e
(ci+cj)ωεx
A
ij
cj− 1
,ω^2 =− 1 ,
where
A=|ars|n,
ars=δrsbr+
e
(cr+cs)ωεx
cr+cs
.
Thebr,r≥ 1 , are arbitrary constants and thecr,r≥ 1 , are constants such
thatcj=1, 1 ≤j≤nandcr+cs=0, 1 ≤r, s≤n, but are otherwise
arbitrary.
The analysis which follows differs from the original both in the form of
the solution and the method by which it is obtained.
Proof. LetA=|ars(u)|ndenote the symmetric determinant in which
ars=δrsbr+
e
(cr+cs)u
cr+cs
=asr,
a
′
rs=e
(cr+cs)u
. (6.4.2)
Then the double-sum relations (A)–(D) in Section 3.4 withfr=gr=cr
become
(logA)
′
=
∑
r,s
e
(cr+cs)u
A
rs
, (6.4.3)