Determinants and Their Applications in Mathematical Physics

(Chris Devlin) #1
4.9 Hankelians 2 119

4.9.2 The Derivatives of Turanians with Appell and Other


Elements


Let


T=T

(n,r)
=


∣C

rCr+1Cr+2···Cr+n− 1



n

, (4.9.5)

where


Cj=

[

φjφj+1φj+2···φj+n− 1

]T

,

φ


m=mF φm−^1.

Theorem 4.34.


T


=rF


∣C

r− 1 Cr+1Cr+2···Cr+n− 1


∣.

Proof.


C


j=F

(

jCj− 1 +C


j

)

,

where


C


j

=

[

0 φj 2 φj+1 3 φj+2···(n−1)φj+n− 2

]T

.

Hence,


T


=

r+n− 1

j=r



CrCr+1···Cj− 1 C


j···Cr+n−^1



=F

r+n− 1

j=r



CrCr+1···Cj− 1 (jCj− 1 +C


j)···Cr+n−^1



=rF


∣C

r− 1 Cr+1Cr+2···Cr+n− 1



+F

r+n− 1

j=r



CrCr+1···C


j···Cr+n−^1



after discarding determinants with two identical columns. The sum is zero


by Theorem 3.1 in Section 3.1 on cyclic dislocations and generalizations.


The theorem follows. 


The column parameters in the above definition ofTare consecutive. If

they are not consecutive, the notation


Tj
1 j 2 ...jn

=


∣C

j 1 Cj 2 ···Cjr···Cjn


∣ (4.9.6)

is convenient.


T


j 1 j 2 ...jn

=F

n

r=1

jr


∣C

j 1 Cj 2 ···C(jr−1)···Cjn


∣. (4.9.7)

Higher derivatives may be found by repeated application of this formula,


but no simple formula forD
k
(Tj
1 j 2 ...jn
) has been found. However, the

Free download pdf