4.7 Wronskians 99
Hence,
W(ty 1 ,ty 2 ,...,tyn)=
∣
∣
(tC)(tC
′
)(tC
′′
)···(tC
(n−1)
)
∣
∣
=t
n
∣
∣CC′C′′···C(n−1)
∣
∣.
The theorem follows.
Exercise.Prove that
d
n
x
dyn
=
(−1)
n+1
W{y
′′
,(y
2
)
′′
,(y
3
)
′′
...(y
n− 1
)
′′
}
1!2!3!···(n−1)!(y
′
)
n(n+1)/ 2
,
wherey
′
=dy/dx,n≥2. (Mina)
4.7.2 The Derivatives of a Wronskian
The derivative ofWnwith respect tox, when evaluated in column vector
notation, consists of the sum ofndeterminants, only one of which has
distinct columns and is therefore nonzero. That determinant is the one
obtained by differentiating the last column:
W
′
n
=
∣
∣CC′C′′···C(n−3)C(n−2)C(n)
∣
∣.
Differentiating again,
W
′′
n=
∣
∣
CC
′
C
′′
···C
(n−3)
C
(n−1)
C
(n)
∣
∣
+
∣
∣
CC
′
C
′′
···C
(n−3)
C
(n−2)
C
(n+1)
∣
∣
, (4.7.3)
etc. There is no simple formula forW
(r)
n. In some detail,
W
′
n=
∣
∣
∣
∣
∣
∣
∣
∣
y 1 y
′
1 ··· y
(n−2)
1
y
(n)
1
y 2 y
′
2
··· y
(n−2)
2
y
(n)
2
..........................
yn y
′
n ··· y
(n−2)
n y
(n)
n
∣
∣
∣
∣
∣
∣
∣
∣
n
. (4.7.4)
The first (n−1) columns ofW
′
n are identical with the corresponding
columns ofWn. Hence, expandingW
′
nby elements from its last column,
W
′
n
=
n
∑
r=1
y
(n)
r
W
(n)
rn
. (4.7.5)
Each of the cofactors in the sum is itself a Wronskian of order (n−1):
W
(n)
rn
=(−1)
r+n
W(y 1 ,y 2 ,...,yr− 1 ,yr+1,...,yn). (4.7.6)
W
′
nis a cofactor ofWn+1:
W
′
n
=−W
(n+1)
n+1,n
. (4.7.7)
Repeated differentiation of a Wronskian of ordernis facilitated by adopting
the notation
Wijk...r=
∣
∣C(i)C(j)C(k)···C(r)