23.5 Special Lattices23.7 Quarter Periods

§23.6 Relations to Other Functions

Contents

§23.6(i) Theta Functions

In this subsection 2\omega _{1}, 2\omega _{3} are any pair of generators of the lattice \mathbb{L}, and the lattice roots e_{1}, e_{2}, e_{3} are given by (23.3.9).

23.6.1
q=e^{{i\pi\tau}},
\tau=\omega _{3}/\omega _{1}.
23.6.2e_{1}=\frac{\pi^{2}}{12\omega _{1}^{2}}\left({\mathop{\theta _{{2}}\/}\nolimits^{{4}}}\!\left(0,q\right)+2\!{\mathop{\theta _{{4}}\/}\nolimits^{{4}}}\!\left(0,q\right)\right),
23.6.3e_{2}=\frac{\pi^{2}}{12\omega _{1}^{2}}\left({\mathop{\theta _{{2}}\/}\nolimits^{{4}}}\!\left(0,q\right)-{\mathop{\theta _{{4}}\/}\nolimits^{{4}}}\!\left(0,q\right)\right),
23.6.4e_{3}=-\frac{\pi^{2}}{12\omega _{1}^{2}}\left(2\!{\mathop{\theta _{{2}}\/}\nolimits^{{4}}}\!\left(0,q\right)+{\mathop{\theta _{{4}}\/}\nolimits^{{4}}}\!\left(0,q\right)\right).
23.6.8\eta _{1}=-\frac{\pi^{2}}{12\omega _{1}}\frac{{\mathop{\theta _{{1}}\/}\nolimits^{{\prime\prime\prime}}}\!\left(0,q\right)}{{\mathop{\theta _{{1}}\/}\nolimits^{{\prime}}}\!\left(0,q\right)}.
23.6.9\mathop{\sigma\/}\nolimits\!\left(z\right)=2\omega _{1}\mathop{\exp\/}\nolimits\!\left(\frac{\eta _{1}z^{2}}{2\omega _{1}}\right)\frac{\mathop{\theta _{{1}}\/}\nolimits\!\left(\pi z/(2\omega _{1}),q\right)}{\pi{\mathop{\theta _{{1}}\/}\nolimits^{{\prime}}}\!\left(0,q\right)},

With z=\ifrac{\pi u}{(2\omega _{1})},

23.6.13\mathop{\zeta\/}\nolimits\!\left(u\right)=\frac{\eta _{1}}{\omega _{1}}u+\frac{\pi}{2\omega _{1}}\frac{d}{dz}\mathop{\ln\/}\nolimits\mathop{\theta _{{1}}\/}\nolimits\!\left(z,q\right),
23.6.14\mathop{\wp\/}\nolimits\!\left(u\right)=\left(\frac{\pi}{2\omega _{1}}\right)^{2}\left(\frac{{\mathop{\theta _{{1}}\/}\nolimits^{{\prime\prime\prime}}}\!\left(0,q\right)}{3\!{\mathop{\theta _{{1}}\/}\nolimits^{{\prime}}}\!\left(0,q\right)}-\frac{{d}^{2}}{{dz}^{2}}\mathop{\ln\/}\nolimits\mathop{\theta _{{1}}\/}\nolimits\!\left(z,q\right)\right),
23.6.15\frac{\mathop{\sigma\/}\nolimits\!\left(u+\omega _{j}\right)}{\mathop{\sigma\/}\nolimits\!\left(\omega _{j}\right)}=\mathop{\exp\/}\nolimits\!\left(\eta _{j}u+\frac{\eta _{j}u^{2}}{2\omega _{1}}\right)\frac{\mathop{\theta _{{j+1}}\/}\nolimits\!\left(z,q\right)}{\mathop{\theta _{{j+1}}\/}\nolimits\!\left(0,q\right)},j=1,2,3.

For further results for the \mathop{\sigma\/}\nolimits-function see Lawden (1989, §6.2).

§23.6(ii) Jacobian Elliptic Functions

Again, in Equations (23.6.16)–(23.6.26), 2\omega _{1},2\omega _{3} are any pair of generators of the lattice \mathbb{L} and e_{1},e_{2},e_{3} are given by (23.3.9).

23.6.16
k^{2}=\frac{e_{2}-e_{3}}{e_{1}-e_{3}},
{k^{{\prime}}}^{2}=\frac{e_{1}-e_{2}}{e_{1}-e_{3}},
23.6.17
{\mathop{K\/}\nolimits^{{2}}}=(\mathop{K\/}\nolimits\!\left(k\right))^{2}=\omega _{1}^{2}(e_{1}-e_{3}),
{\mathop{{K^{{\prime}}}\/}\nolimits^{{2}}}=(\mathop{K\/}\nolimits\!\left(k^{{\prime}}\right))^{2}=\omega _{3}^{2}(e_{3}-e_{1}).

Similar results for the other nine Jacobi functions can be constructed with the aid of the transformations given by Table 22.4.3.

For representations of the Jacobi functions \mathop{\mathrm{sn}\/}\nolimits, \mathop{\mathrm{cn}\/}\nolimits, and \mathop{\mathrm{dn}\/}\nolimits as quotients of \mathop{\sigma\/}\nolimits-functions see Lawden (1989, §§6.2, 6.3).

§23.6(iii) General Elliptic Functions

For representations of general elliptic functions (§23.2(iii)) in terms of \mathop{\sigma\/}\nolimits\!\left(z\right) and \mathop{\wp\/}\nolimits\!\left(z\right) see Lawden (1989, §§8.9, 8.10), and for expansions in terms of \mathop{\zeta\/}\nolimits\!\left(z\right) see Lawden (1989, §8.11).

§23.6(iv) Elliptic Integrals

Rectangular Lattice

Let z be on the perimeter of the rectangle with vertices 0,2\omega _{1},2\omega _{1}+2\omega _{3},2\omega _{3}. Then t=\mathop{\wp\/}\nolimits\!\left(z\right) is real (§§23.5(i)23.5(ii)), and

23.6.30z=\frac{1}{2}\int _{t}^{\infty}\frac{du}{\sqrt{(u-e_{1})(u-e_{2})(u-e_{3})}},t\geq e_{1}, z\in(0,\omega _{1}],
23.6.33z=\frac{i}{2}\int _{{-\infty}}^{t}\frac{du}{\sqrt{(e_{1}-u)(e_{2}-u)(e_{3}-u)}},t\leq e_{3}, z\in(0,\omega _{3}].
23.6.342\omega _{1}=\int _{{e_{1}}}^{\infty}\frac{du}{\sqrt{(u-e_{1})(u-e_{2})(u-e_{3})}}=\int _{{e_{3}}}^{{e_{2}}}\frac{du}{\sqrt{(e_{1}-u)(e_{2}-u)(u-e_{3})}},
23.6.352\omega _{3}=i\int _{{e_{2}}}^{{e_{1}}}\frac{du}{\sqrt{(e_{1}-u)(u-e_{2})(u-e_{3})}}=i\int _{{-\infty}}^{{e_{3}}}\frac{du}{\sqrt{(e_{1}-u)(e_{2}-u)(e_{3}-u)}}.

For (23.6.30)–(23.6.35) and further identities see Lawden (1989, §6.12).

See also §§19.2(i), 19.14, and Erdélyi et al. (1953b, §13.14).

For relations to symmetric elliptic integrals see §19.25(vi).

General Lattice

Let z be a point of \Complex different from e_{1},e_{2},e_{3}, and define w by

23.6.36w=\int _{z}^{\infty}\frac{du}{\sqrt{4u^{3}-g_{2}u-g_{3}}}=\frac{1}{2}\int _{z}^{\infty}\frac{du}{\sqrt{(u-e_{1})(u-e_{2})(u-e_{3})}},

where the integral is taken along any path from z to \infty that does not pass through any of e_{1},e_{2},e_{3}. Then z=\mathop{\wp\/}\nolimits\!\left(w\right), where the value of w depends on the choice of path and determination of the square root; see McKean and Moll (1999, pp. 87–88 and §2.5).