29.2 Differential Equations29.4 Graphics

§29.3 Definitions and Basic Properties

Contents

§29.3(i) Eigenvalues

For each pair of values of \nu and k there are four infinite unbounded sets of real eigenvalues h for which equation (29.2.1) has even or odd solutions with periods 2\!\mathop{K\/}\nolimits\! or 4\!\mathop{K\/}\nolimits\!. They are denoted by \mathop{a^{{2m}}_{{\nu}}\/}\nolimits\!\left(k^{2}\right), \mathop{a^{{2m+1}}_{{\nu}}\/}\nolimits\!\left(k^{2}\right), \mathop{b^{{2m+1}}_{{\nu}}\/}\nolimits\!\left(k^{2}\right), \mathop{b^{{2m+2}}_{{\nu}}\/}\nolimits\!\left(k^{2}\right), where m=0,1,2,\ldots; see Table 29.3.1.

§29.3(ii) Distribution

The eigenvalues interlace according to

The eigenvalues coalesce according to

29.3.5\mathop{a^{{m}}_{{\nu}}\/}\nolimits\!\left(k^{2}\right)=\mathop{b^{{m}}_{{\nu}}\/}\nolimits\!\left(k^{2}\right),\nu=0,1,\dots,m-1.

If \nu is distinct from 0,1,\dots,m-1, then

29.3.6\left(\mathop{a^{{m}}_{{\nu}}\/}\nolimits\!\left(k^{2}\right)-\mathop{b^{{m}}_{{\nu}}\/}\nolimits\!\left(k^{2}\right)\right)\nu(\nu-1)\cdots(\nu-m+1)>0.

If \nu is a nonnegative integer, then

29.3.7\mathop{a^{{m}}_{{\nu}}\/}\nolimits\!\left(k^{2}\right)+\mathop{a^{{\nu-m}}_{{\nu}}\/}\nolimits\!\left(1-k^{2}\right)=\nu(\nu+1),m=0,1,\dots,\nu,
29.3.8{\mathop{b^{{m}}_{{\nu}}\/}\nolimits\!\left(k^{2}\right)+\mathop{b^{{\nu-m+1}}_{{\nu}}\/}\nolimits\!\left(1-k^{2}\right)=\nu(\nu+1)},m=1,2,\dots,\nu.

For the special case k=k^{{\prime}}=\ifrac{1}{\sqrt{2}} see Erdélyi et al. (1955, §15.5.2).

§29.3(iii) Continued Fractions

The quantity

29.3.9H=2\!\mathop{a^{{2m}}_{{\nu}}\/}\nolimits\!\left(k^{2}\right)-\nu(\nu+1)k^{2}

satisfies the continued-fraction equation

29.3.10\beta _{p}-H-\cfrac{\alpha _{{p-1}}\gamma _{p}}{\beta _{{p-1}}-H-\cfrac{\alpha _{{p-2}}\gamma _{{p-1}}}{\beta _{{p-2}}-H-}}\dots=\cfrac{\alpha _{p}\gamma _{{p+1}}}{\beta _{{p+1}}-H-\cfrac{\alpha _{{p+1}}\gamma _{{p+2}}}{\beta _{{p+2}}-H-}}\cdots,

where p is any nonnegative integer, and

29.3.11\alpha _{p}=\begin{cases}(\nu-1)(\nu+2)k^{2},&p=0,\\
\tfrac{1}{2}(\nu-2p-1)(\nu+2p+2)k^{2},&p\geq 1,\end{cases}
29.3.12
\beta _{p}=4p^{2}(2-k^{2}),
\gamma _{p}=\tfrac{1}{2}(\nu-2p+2)(\nu+2p-1)k^{2}.

The continued fraction following the second negative sign on the left-hand side of (29.3.10) is finite: it equals 0 if p=0, and if p>0, then the last denominator is \beta _{0}-H. If \nu is a nonnegative integer and 2p\leq\nu, then the continued fraction on the right-hand side of (29.3.10) terminates, and (29.3.10) has only the solutions (29.3.9) with 2m\leq\nu. If \nu is a nonnegative integer and 2p>\nu, then (29.3.10) has only the solutions (29.3.9) with 2m>\nu.

The quantity H=2\!\mathop{a^{{2m+1}}_{{\nu}}\/}\nolimits\!\left(k^{2}\right)-\nu(\nu+1)k^{2} satisfies equation (29.3.10) with

29.3.13\beta _{p}=\begin{cases}2-k^{2}+\tfrac{1}{2}\nu(\nu+1)k^{2},&p=0,\\
(2p+1)^{2}(2-k^{2}),&p\geq 1,\end{cases}
29.3.14
\alpha _{p}=\tfrac{1}{2}(\nu-2p-2)(\nu+2p+3)k^{2},
\gamma _{p}=\tfrac{1}{2}(\nu-2p+1)(\nu+2p)k^{2}.

The quantity H=2\!\mathop{b^{{2m+1}}_{{\nu}}\/}\nolimits\!\left(k^{2}\right)-\nu(\nu+1)k^{2} satisfies equation (29.3.10) with

29.3.15\beta _{p}=\begin{cases}2-k^{2}-\tfrac{1}{2}\nu(\nu+1)k^{2},&p=0,\\
(2p+1)^{2}(2-k^{2}),&p\geq 1,\end{cases}
29.3.16
\alpha _{p}=\tfrac{1}{2}(\nu-2p-2)(\nu+2p+3)k^{2},
\gamma _{p}=\tfrac{1}{2}(\nu-2p+1)(\nu+2p)k^{2}.

The quantity H=2\!\mathop{b^{{2m+2}}_{{\nu}}\/}\nolimits\!\left(k^{2}\right)-\nu(\nu+1)k^{2} satisfies equation (29.3.10) with

29.3.17
\alpha _{p}=\tfrac{1}{2}(\nu-2p-3)(\nu+2p+4)k^{2},
\beta _{p}=(2p+2)^{2}(2-k^{2}),
\gamma _{p}=\tfrac{1}{2}(\nu-2p)(\nu+2p+1)k^{2}.

§29.3(iv) Lamé Functions

The eigenfunctions corresponding to the eigenvalues of §29.3(i) are denoted by \mathop{\mathit{Ec}^{{2m}}_{{\nu}}\/}\nolimits\!\left(z,k^{2}\right), \mathop{\mathit{Ec}^{{2m+1}}_{{\nu}}\/}\nolimits\!\left(z,k^{2}\right), \mathop{\mathit{Es}^{{2m+1}}_{{\nu}}\/}\nolimits\!\left(z,k^{2}\right), \mathop{\mathit{Es}^{{2m+2}}_{{\nu}}\/}\nolimits\!\left(z,k^{2}\right). They are called Lamé functions with real periods and of order \nu, or more simply, Lamé functions. See Table 29.3.2. In this table the nonnegative integer m corresponds to the number of zeros of each Lamé function in (0,\!\mathop{K\/}\nolimits\!), whereas the superscripts 2m, 2m+1, or 2m+2 correspond to the number of zeros in [0,2\!\mathop{K\/}\nolimits\!).

Table 29.3.2: Lamé functions.
boundary conditions
eigenvalue
h
eigenfunction
w(z)
parity of
w(z)
parity of
w(z-\!\mathop{K\/}\nolimits\!)
period of
w(z)
\left.\ifrac{dw}{dz}\right|_{{z=0}}=\left.\ifrac{dw}{dz}\right|_{{z=\!\mathop{K\/}\nolimits\!}}=0 \mathop{a^{{2m}}_{{\nu}}\/}\nolimits\!\left(k^{2}\right) \mathop{\mathit{Ec}^{{2m}}_{{\nu}}\/}\nolimits\!\left(z,k^{2}\right) even even 2\!\mathop{K\/}\nolimits\!
w(0)=\left.\ifrac{dw}{dz}\right|_{{z=\!\mathop{K\/}\nolimits\!}}=0 \mathop{a^{{2m+1}}_{{\nu}}\/}\nolimits\!\left(k^{2}\right) \mathop{\mathit{Ec}^{{2m+1}}_{{\nu}}\/}\nolimits\!\left(z,k^{2}\right) odd even 4\!\mathop{K\/}\nolimits\!
\left.\ifrac{dw}{dz}\right|_{{z=0}}=w(\!\mathop{K\/}\nolimits\!)=0 \mathop{b^{{2m+1}}_{{\nu}}\/}\nolimits\!\left(k^{2}\right) \mathop{\mathit{Es}^{{2m+1}}_{{\nu}}\/}\nolimits\!\left(z,k^{2}\right) even odd 4\!\mathop{K\/}\nolimits\!
w(0)=w(\!\mathop{K\/}\nolimits\!)=0 \mathop{b^{{2m+2}}_{{\nu}}\/}\nolimits\!\left(k^{2}\right) \mathop{\mathit{Es}^{{2m+2}}_{{\nu}}\/}\nolimits\!\left(z,k^{2}\right) odd odd 2\!\mathop{K\/}\nolimits\!

§29.3(v) Normalization

To complete the definitions, \mathop{\mathit{Ec}^{{m}}_{{\nu}}\/}\nolimits\!\left(\!\mathop{K\/}\nolimits\!,k^{2}\right) is positive and \left.\ifrac{d\mathop{\mathit{Es}^{{m}}_{{\nu}}\/}\nolimits\!\left(z,k^{2}\right)}{dz}\right|_{{z=\!\mathop{K\/}\nolimits\!}} is negative.

§29.3(vi) Orthogonality

§29.3(vii) Power Series

For power-series expansions of the eigenvalues see Volkmer (2004b).