Erlangen Program at Large-1: Geometry of Invariants

For any type of the Clifford algebra {\ensuremath{ \mathcal{C}\kern-0.12em\ell(\sigma)}}:

1. The subgroup N defines shifts ue_0+ve_1 \mapsto (u+\nu)e_0+ve_1 along the “real” axis U by \nu.
   The vector field of the derived representation is dN_a{(u,v)}=(1,0).

2. The subgroup A defines dilations ue_0+ve_1 \mapsto \alpha^{-2}(ue_0+ve_1) by the factor \alpha^{-2} which fixes origin (0,0).
   The vector field of the derived representation is dA_a{(u,v)}=(2u,2v).

A K-orbit in \ensuremath{\mathbb{R}^{\sigma}_{}{}} passing the point (0,t) has the following equation:

The curvature of a K-orbit at point (0,t) is equal to

(c@enumiaeph)(c@enumiaeph)

1. For {\ensuremath{ \mathcal{C}\kern-0.12em\ell(e)}} the orbits of K are circles, they are coaxal [CoxeterGreitzer]*§ 2.3 with the real line being the radical axis. A circle with centre at (0, (v+v^{-1})/2) passing through two points (0,v) and (0,v^{-1}).
   The vector field of the derived representation is dK_e{(u,v)}=(u^2-v^2+1, 2uv).

2. For {\ensuremath{ \mathcal{C}\kern-0.12em\ell(p)}} the orbits of K are parabolas with the vertical axis V. A parabola passing through (0,v/2) has horizontal directrix passing through (0, v-v^{-1}/2) and focus at (0,(v+v^{-1})/2).
   The vector field of the derived representation is dK_p{(u,v)}=(u^2+1, 2uv).

3. For {\ensuremath{ \mathcal{C}\kern-0.12em\ell(h)}} the orbits of K are hyperbolas with asymptotes parallel to lines u=\pm v. A hyperbola passing through the point (0,v) has the focal distance 2p, where p=\frac{v^2+1}{\sqrt{2}v} and the upper focus is located at (0,f) with:

   f=\left\{\begin{array}{ll} p-\sqrt{\frac{p^2}{2}-1}, &\textrm{ for } 0<v<1; \textrm{ and }\\ p+\sqrt{\frac{p^2}{2}-1}, & \textrm{ for } v\geq 1. \end{array}\right.

   The vector field of the derived representation is dK_h{(u,v)}=(u^2+v^2+1, 2uv).

K-orbits may be obtained cases as follows:

(c@enumiaeph)(c@enumiaeph)

1. elliptic K-orbits are sections of cones (?) by the plane z=0 (EE' on Fig. ?);

2. parabolic K-orbits are sections of (?) by the plane y=\pm z (PP' on Fig. ?);

3. hyperbolic K-orbits are sections of (?) by the plane y=0 (HH' on Fig. ?);

Moreover, each straight line generating a cone from the family (?) is crossing corresponding elliptic, parabolic and hyperbolic K-orbits at points with the same value of parameter \phi (?) of the subgroup K.

The fix group of the point e_1 is

(c@enumiaeph)(c@enumiaeph)

1. the subgroup K'_e=K in the elliptic case. Thus the elliptic upper half-plane is a model for the homogeneous space \ensuremath{\ensuremath{\ifxSLl \ell_{2}^{}{}\else SL_{2}^{}{}\fi}(\ensuremath{\mathbb{R}^{}_{}{}})}/K;

2. the subgroup N'_p of matrices

   \begin{pmatrix} 1&0\\ \nu e_0 & 1 \end{pmatrix} = \begin{pmatrix} 0&e_0\\ e_0 & 0 \end{pmatrix} \begin{pmatrix} 1&\nu e_0\\ 0& 1 \end{pmatrix} \begin{pmatrix} 0&-e_0\\ -e_0 & 0 \end{pmatrix}

   in the parabolic case. It also fixes any point ve1. It is conjugate to subgroup N, thus the parabolic upper half-plane is a model for the homogeneous space \ensuremath{\ensuremath{\ifxSLl \ell_{2}^{}{}\else SL_{2}^{}{}\fi}(\ensuremath{\mathbb{R}^{}_{}{}})}/N;

3. the subgroup A'_h of matrices

   \begin{pmatrix} \cosh(\tau) & \sinh(\tau)e_0\\ -\sinh(\tau)e_0 & \cosh(\tau) \end{pmatrix} = \frac{1}{2} \begin{pmatrix} 1 & -e_0\\ -e_0 & 1 \end{pmatrix} \begin{pmatrix} e^\tau & 0\\ 0 & e^{-\tau} \end{pmatrix} \begin{pmatrix} 1 & e_0\\ e_0 & 1 \end{pmatrix},

   in the hyperbolic case. It is conjugate to subgroup A, thus two copies of the upper halfplane (see Section ?) is a model for \ensuremath{\ensuremath{\ifxSLl \ell_{2}^{}{}\else SL_{2}^{}{}\fi}(\ensuremath{\mathbb{R}^{}_{}{}})}/A.

Moreover, vectors fields of these actions are (u^2+\sigma(v^2-1), 2uv) for the corresponding values of \sigma. Orbits of the fix groups satisfy to the equation:

Möbius action of \ensuremath{\ensuremath{\ifxSLl \ell_{2}^{}{}\else SL_{2}^{}{}\fi}(\ensuremath{\mathbb{R}^{}_{}{}})} in each EPH case is generated by action the corresponding fix-subgroup (A''_h in the hyperbolic case) and actions of the ax+b group, e.g. subgroups A and N.

Möbius transformations preserve the cycles in the upper half-plane, i.e.:

(c@enumiaeph)(c@enumiaeph)

1. For {\ensuremath{ \mathcal{C}\kern-0.12em\ell(e)}} Möbius transformations map circles to circles.

2. For {\ensuremath{ \mathcal{C}\kern-0.12em\ell(p)}} Möbius transformations map parabolas to parabolas.

3. For {\ensuremath{ \mathcal{C}\kern-0.12em\ell(h)}} Möbius transformations map hyperbolas to hyperbolas.

A cycle -k(e_0u+e_1v)^2-2\left\langle (l,n),(u,v) \right\rangle\ifx\relax\relax\else_{\relax}\fi+m=0 is transformed by g\in\ensuremath{\ensuremath{\ifxSLl \ell_{2}^{}{}\else SL_{2}^{}{}\fi}(\ensuremath{\mathbb{R}^{}_{}{}})} into the cycle -\tilde{k}(e_0u+e_1v)^2-2\left\langle (\tilde{l}, \tilde{n}),(u,v) \right\rangle\ifx\relax\relax\else_{\relax}\fi+\tilde{m}=0 such that

for any Clifford algebras {\ensuremath{ \mathcal{C}\kern-0.12em\ell(\sigma)}} and {\ensuremath{ \mathcal{C}\kern-0.12em\ell(\breve{\sigma})}}. Explicitly this means:

Any implementation of cycles shown on (?) by the dotted arrow for any combination of \sigma, \breve{\sigma} and s intertwines two actions of \ensuremath{\ensuremath{\ifxSLl \ell_{2}^{}{}\else SL_{2}^{}{}\fi}(\ensuremath{\mathbb{R}^{}_{}{}})}: by matrix conjugation (3.1) and Möbius transformations (?).

The conjugate g^{-1} Z^{s}_{\breve{\sigma}}(y)g of a \breve{\sigma}-zero-radius cycle Z^{s}_{\breve{\sigma}}(y) with g\in\ensuremath{\ensuremath{\ifxSLl \ell_{2}^{}{}\else SL_{2}^{}{}\fi}(\ensuremath{\mathbb{R}^{}_{}{}})} is a \breve{\sigma}-zero-radius cycle Z^{s}_{\breve{\sigma}}(g\cdot y) with centre at g\cdot y—the Möbius transform of the centre of Z^{s}_{\breve{\sigma}}(y).

Self-adjoint cycles form a family, which is invariant under the Möbius transformations.

The \breve{\sigma}-orthogonality condition (4.1) is invariant under Möbius transformations.

The \breve{\sigma}-orthogonality (4.1) of cycles {\tilde C^{s}_{\breve{\sigma}}} and { C^{s}_{\breve{\sigma}}} is given through their defining equation (3.1) coefficients by

or specifically by

in the elliptic, parabolic and hyperbolic cases of {\ensuremath{ \mathcal{C}\kern-0.12em\ell(\sigma)}} correspondingly.

1. A \breve{\sigma}-self-orthogonal cycle is \breve{\sigma}-zero-radius one (3.5).

2. For \breve{\sigma}=\pm 1 there is no non-trivial cycle orthogonal to all other non-trivial cycles.
   For \breve{\sigma}=0 only the real axis v=0 is orthogonal to all other non-trivial cycles.

3. For \breve{\sigma}=\pm 1 any cycle is uniquely defined by the family of cycles orthogonal to it, i.e. ({ C^{s}_{\breve{\sigma}}}^\perp)^\perp=\{{ C^{s}_{\breve{\sigma}}}\} .
   For \breve{\sigma}= 0 the set ({ C^{s}_{\breve{\sigma}}}^\perp)^\perp consists of all cycles which have the same roots as { C^{s}_{\breve{\sigma}}}, see middle column of pictures in Fig. ?.

A cycle { C^{s}_{\breve{\sigma}}} is \breve{\sigma}-orthogonal to \sigma-zero-radius cycle Z^{s}_{\sigma }(u,v) if

i.e. \sigma-implementation of { C^{s}_{\breve{\sigma}}} is passing through the point (u, \breve{\sigma} v), which \breve{\sigma}-centre of Z^{s}_{\sigma}(u,v).

Let T: \ensuremath{\mathbb{P}^{3}_{}{}} \rightarrow \ensuremath{\mathbb{P}^{3}_{}{}} is an orthogonality preserving map of the cycles space, i.e. \left\langle { C^{s}_{\breve{\sigma}}},{\tilde C^{s}_{\breve{\sigma}}} \right\rangle\ifx\relax\relax\else_{\relax}\fi=0 \ \Leftrightarrow \left\langle T{ C^{s}_{\breve{\sigma}}},T{\tilde C^{s}_{\breve{\sigma}}} \right\rangle\ifx\relax\relax\else_{\relax}\fi=0. Then for \sigma\neq 0 there is a map T_\sigma : \ensuremath{\mathbb{R}^{\sigma}_{}{}} \rightarrow\ensuremath{\mathbb{R}^{\sigma}_{}{}}, such that Q intertwines T and T_\sigma:

Let T_i: \ensuremath{\mathbb{P}^{3}_{}{}} \rightarrow \ensuremath{\mathbb{P}^{3}_{}{}}, i=1,2 are two orthogonality preserving maps of the cycles space. If they coincide on the subspace of \breve{\sigma}-zero-radius cycles, \breve{\sigma}\neq 0, then they are identical in the whole \ensuremath{\mathbb{P}^{3}_{}{}}.

Let cycles { C^{}_{\breve{\sigma}}} and {\tilde C^{}_{\breve{\sigma}}} be \breve{\sigma}-orthogonal. For their \sigma-implementations we define the ghost cycle {\hat C^{}_{\breve{\sigma}}} by the following two conditions:

1. \chi(\sigma)-centre of {\hat C^{}_{\breve{\sigma}}} coincides with \breve{\sigma}-centre of { C^{}_{\breve{\sigma}}}.

2. Determinant of {\hat C^{1}_{\sigma}} is equal to determinant of { C^{\chi(\breve{\sigma})}_{\sigma}}.

Then:

1. {\hat C^{}_{\sigma}} coincides with { C^{}_{\sigma}} if \sigma \breve{\sigma}=1;

2. {\hat C^{}_{\sigma}} has common roots (real or imaginary) with { C^{}_{\sigma}};

3. In the \sigma-implementation the tangent line to {\tilde C^{}_{\breve{\sigma}}} at points of its intersections with the ghost cycle {\hat C^{}_{\sigma}} are passing the \sigma-centre of {\hat C^{}_{\sigma}}.

Let \det { C^{s}_{\breve{\sigma}}} \neq 0, then:

1. The conjugation (?) preserves the orthogonality relation (4.1).

2. The image { C^{s_2}_{\sigma}}\tilde Z^{s_1}_{\sigma}(u, v){ C^{s_2}_{\sigma}} of a \sigma-zero-radius cycle \tilde Z^{s_1}_{\sigma} under the conjugation (?) is a \sigma-zero-radius cycle \tilde Z^{s_1}_{\sigma}(u', v'), where (u',v') is calculated by the linear-fractional transformation (?) (u',v')={ C^{s_1s_2}_{\sigma}}(u,v) associated to the cycle { C^{s_1s_2}_{\sigma}}.

3. Both formulae (?) and (?) define the same transformation of the point space.

A cycle {\tilde C^{s}_{\breve{\sigma}}} is orthogonal to a cycle { C^{s}_{\breve{\sigma}}} if for any point u_1e_0+v_1e_1\in {\tilde C^{s}_{\breve{\sigma}}} the cycle {\tilde C^{s}_{\breve{\sigma}}} is also passing through its image

under the Möbius transform defined by the matrix { C^{s}_{\breve{\sigma}}}. Thus the point u_2e_0+v_2e_1={ C^{s}_{\breve{\sigma}}}(u_1e_0+v_1e_1) is the inversion of u_1e_0+v_1e_1 in { C^{s}_{\breve{\sigma}}}.

The reflection 4.2.? of a zero-radius cycle Z^{s}_{\breve{\sigma}} in a cycle { C^{s}_{\breve{\sigma}}} is given by the conjugation: { C^{s}_{\breve{\sigma}}} Z^{s}_{\breve{\sigma}}{ C^{s}_{\breve{\sigma}}}.

Let { C^{s}_{\breve{\sigma}}}=(k, l, n,m) be a cycle and for \breve{\sigma}\neq 0 the {\tilde C^{s}_{\breve{\sigma}}} be given by (k, l, n\pm\sqrt{\det{ C^{\sigma}_{\breve{\sigma}}}},m). Then

1. {\tilde C^{s}_{\breve{\sigma}}} { C^{s}_{\breve{\sigma}}} {\tilde C^{s}_{\breve{\sigma}}} = \ensuremath{\mathbb{R}^{}_{}{}} and {\tilde C^{s}_{\breve{\sigma}}} \ensuremath{\mathbb{R}^{}_{}{}} {\tilde C^{s}_{\breve{\sigma}}} = { C^{s}_{\breve{\sigma}}}

2. {\tilde C^{s}_{\breve{\sigma}}} and { C^{s}_{\breve{\sigma}}} have common roots.

3. In the \breve{\sigma}-implementation the cycle { C^{s}_{\breve{\sigma}}} passes the centre of {\tilde C^{s}_{\breve{\sigma}}}.

The inversion of second kind (?) is a composition of three inversions: in parabolas u^2- 2lu -4mv-m/k=0, u^2- 2lu-m/k=0, and the real line.

The image { C^{s_1}_{\breve{\sigma}}} R^{s}_{\breve{\sigma}} { C^{s_1}_{\breve{\sigma}}} of the real line under inversion in { C^{s_1}_{\breve{\sigma}}}=(k,l,n,m) is the cycle:

It is the real line if s\cdot\det({ C^{s_1}_{\breve{\sigma}}})\neq0 and either

1. s_1 n=0, in this case it is a composition of \ensuremath{\ensuremath{\ifxSLl \ell_{2}^{}{}\else SL_{2}^{}{}\fi}(\ensuremath{\mathbb{R}^{}_{}{}})}-action by \begin{pmatrix} l&-me_0\\ke_0 &-l \end{pmatrix} and the reflection in the real line; or

2. \breve{\sigma}=0, i.e. the parabolic case of the cycle space.

If this condition is satisfied than f-orthogonality preserved by the inversion {\tilde C^{s}_{\breve{\sigma}}}\rightarrow { C^{s_1}_{\breve{\sigma}}} {\tilde C^{s}_{\breve{\sigma}}} { C^{s_1}_{\breve{\sigma}}} in {\tilde C^{s}_{\breve{\sigma}}}.

f-orthogonality of {{ C^{s}_{p}}} to {{\tilde C^{s}_{p}}} is given by either of the following equivalent identities

Let { C^{s}_{\breve{\sigma}}} be a cycle, then its f-ghost cycle {\tilde C^{\breve{\sigma}}_{\breve{\sigma}}} = { C^{\chi(\sigma)}_{\breve{\sigma}}} \ensuremath{\mathbb{R}^{\breve{\sigma}}_{\breve{\sigma}}{}} { C^{\chi(\sigma)}_{\breve{\sigma}}} is the reflection of the real line in { C^{\chi(\sigma)}_{\breve{\sigma}}}, where \chi(\sigma) is the Heaviside function ?. Then

1. Cycles { C^{s}_{\breve{\sigma}}} and {\tilde C^{\breve{\sigma}}_{\breve{\sigma}}} have the same roots.

2. \chi(\sigma)-Centre of {\tilde C^{\breve{\sigma}}_{\breve{\sigma}}} coincides with the \breve{\sigma}-focus of { C^{s}_{\breve{\sigma}}}, consequently all lines f-orthogonal to { C^{s}_{\breve{\sigma}}} are passing one of its foci.

3. s-Reflection in { C^{s}_{\breve{\sigma}}} defined from f-orthogonality (see Definition 4.2.?) coincides with usual inversion in {\tilde C^{\breve{\sigma}}_{\breve{\sigma}}}.

The \breve{\sigma}-radius of a cycle { C^{s}_{\breve{\sigma}}} is equal to 1/k, where k is (2,1)-entry of \det-normalised matrix (see Defn. 3.3) of the cycle.

1. The cycle of the form (3.5) has zero radius.

2. The distance between two points y=e_0u+e_1v and y'=e_0u'+e_1v' in the elliptic or hyperbolic spaces is

   d^2(y, y') = \frac{ \breve{\sigma} ((u-u')^2-\sigma(v- v')^2) +4(1-\sigma\breve{\sigma}) v v'} {(u- u')^2 \breve{\sigma}-(v-v')^2} ((u-u')^2 -\sigma(v- v')^2),

   and in parabolic case it is (see Fig. ?(b) and [Yaglom79]*p. 38, (5))

   d^2(y, y') = (u-u')^2.

If cycles { C^{s}_{\breve{\sigma}}} and {\tilde C^{s}_{\breve{\sigma}}} are normalised by conditions k=1 and \tilde{k}=1 then

where \left| c-c \right|\ifx\relax\relax\else_{\relax}\fi_{\breve{\sigma}}^2=(l-\tilde{l})^2-\breve{\sigma}(n-\tilde{n})^2 is the square of \breve{\sigma}-distance between cycles’ centres, r_{\breve{\sigma}} and \tilde{r}_{\breve{\sigma}} are bs-radii of the respective cycles.

1. The \breve{\sigma}-length from the \mathring{\sigma}-centre between two points y=e_0u+e_1v and y'=e_0u'+e_1v' is

   l_{c_{\breve{\sigma}}}^2(y, y') = (u-u')^2-\sigma v'^2+2\mathring{\sigma} v v' -\breve{\sigma} v^2.

2. The \breve{\sigma}-length from the \mathring{\sigma}-focus between two points y=e_0u+e_1v and y'=e_0u'+e_1v' is

   l_{f_{\breve{\sigma}}}^2(y, y') = (\mathring{\sigma}-\breve{\sigma}) p^2-2vp,

   where:

   \begin{aligned} p & = & \mathring{\sigma}\left(-(v'-v)\pm\sqrt{\mathring{\sigma}(u'-u)^2+(v'-v)^2-\sigma\mathring{\sigma} v'^2}\right), \quad\text{if } \mathring{\sigma}\neq0;\\ p & = & \frac{(u'-u)^2-\sigma v'^2}{2(v'-v)}, \quad\text{if } \mathring{\sigma}=0. \end{aligned}

1. The distance (?) is conformal if and only if the type of point and cycle spaces are the same, i.e. \sigma\breve{\sigma}=1. The parabolic distance (?) is conformal only in the parabolic point space.

2. The lengths from centres (?) are conformal for any combination of values of \sigma, \breve{\sigma} and \mathring{\sigma}.

3. The lengths from foci (?) are conformal for \mathring{\sigma}\neq 0 and any combination of values of \sigma and \breve{\sigma}.

Let the focal length is given by the identity (?) with \sigma=\mathring{\sigma}=0, e.g.:

Then it is conformal in the sense that for any constant y=ue_0+ve_1 and y'=u'e_0+v'e_1 with a fixed u' we have:

Let the line element be \left| dy \right|\ifx\relax\relax\else_{\relax}\fi^2=du^2-\sigma dv^2 and the “length of a curve” is given by the corresponding line integral, cf. [Beardon05a]*§ 15.2:

Then the length of the curve is preserved under the Möbius transformations.

Let two points w_1 and w_2 in the upper half-plane are linked by an arc of a parabola with zero \breve{\sigma}-radius. Then the length (5.4) along the arc is invariant under Möbius transformations.

Let \overrightarrow{AB} be l_{c_{\breve{\sigma}}}-perpendicular (l_{f_{\breve{\sigma}}}-perpendicular) to a vector \overrightarrow{CD}. Then the flat cycle (straight line) AB, is (s-)orthogonal to the cycle { C^{s}_{\sigma}} with centre (focus) at A passing through B. The vector \overrightarrow{CD} is tangent to { C^{s}_{\sigma}} at B.

Let A=(u,v) and B=(u',v'). Then

1. d-perpendicular (in the sense of (?)) to \overrightarrow{AB} in the elliptic or hyperbolic cases is a multiple of the vector

   (\sigma (v-v')^3-(u-u')^2 (v+v'(1-2 \sigma \breve{\sigma})), \breve{\sigma}(u-u')^3-(u-u')(v- v')(-2 v' +(v+v') \breve{\sigma} \sigma)),

   which for \sigma\breve{\sigma}=1 reduces to the expected value (v-v', \sigma(u-u')).

2. d-perpendicular (in the sense of (?)) to \overrightarrow{AB} in the parabolic case is (0, t), t\in\ensuremath{\mathbb{R}^{}_{}{}} which coincides with the Galilean orthogonality defined in [Yaglom79]*§ 3.

3. l_{c_{\breve{\sigma}}}-perpendicular (in the sense of (?)) to \overrightarrow{AB} is a multiple of

   (\sigma v'-\mathring{\sigma} v, u-u').

4. l_{f_{\breve{\sigma}}}-perpendicular (in the sense of (?)) to \overrightarrow{AB} is a multiple of

   (\sigma v'+p, u-u'), where p is defined either by (?) or by (?) for corresponding values of \mathring{\sigma}.

Let \overrightarrow{AB}=u e_0+v e_1 and \overrightarrow{CD}=u'e_0+v'e_1, then:

(c@enumiaeph)(c@enumiaeph)

1. In the elliptic case the d-perpendicularity for \breve{\sigma}=-1 means that \overrightarrow{AB} and \overrightarrow{CD} form a right angle, or analytically u u'+v v'=0.

2. In the parabolic case the l_{f_{\breve{\sigma}}}-perpendicularity for \breve{\sigma}=1 means that \overrightarrow{AB} bisect the angle between \overrightarrow{CD} and the vertical direction or analytically:

   u'u-v'p=u'u-v'(\sqrt{u^2+v^2}-v)=0,

   where p is the focal length (?)

3. In the hyperbolic case the d-perpendicularity for \breve{\sigma}=-1 means that the angles between \overrightarrow{AB} and \overrightarrow{CD} are bisected by lines parallel to u=\pm v, or analytically u' u-v' v=0.

Let \breve{\sigma} and \mathring{\sigma} be two metric signs and let a point (u_0,v_0)\in\ensuremath{\mathbb{R}^{p}_{}{}} with v_0>0. Consider a cycle { C^{s}_{\breve{\sigma}}} defined by

where

Then

1. The point (u_0, v_0) is \mathring{\sigma}-focus of the cycle.

2. The square of \breve{\sigma}-radius is exactly -\varepsilon^2, i.e. (6.1) defines an infinitesimal radius cycle.

3. The focal length of the cycle is an infinitesimal number of order \varepsilon^2.

[Kisil05b]*§ ? Infinitesimal cycle (6.1) consists of points, which are infinitesimally close (in the sense of length from focus (?)) to its focus F=(u_0, v_0):

The image of \ensuremath{\ensuremath{\ifxSLl \ell_{2}^{}{}\else SL_{2}^{}{}\fi}(\ensuremath{\mathbb{R}^{}_{}{}})}-action on an infinitesimal radius cycle (6.1) by conjugation (3.1) is an infinitesimal radius cycle of the same order.

Image of an infinitesimal cycle under cycle conjugation is an infinitesimal cycle of the same or lesser order.

Let { C^{s}_{\breve{\sigma}}} be the infinitesimal cycle (6.1) and {\breve C^{s}_{\breve{\sigma}}}=(k,l,n,m) be a generic cycle. Then

1. The orthogonality condition (4.1) { C^{s}_{\breve{\sigma}}}\perp {\breve C^{s}_{\breve{\sigma}}} and the f-orthogonality (4.3) {\breve C^{s}_{\breve{\sigma}}} \dashv { C^{s}_{\breve{\sigma}}} both are given by:

   ku_0^2-2lu_0+m=O(\varepsilon).

   In other words the cycle {\breve C^{s}_{\breve{\sigma}}} has root u_0 in the parabolic space.

2. The f-orthogonality (4.3) { C^{s}_{\breve{\sigma}}} \dashv {\breve C^{s}_{\breve{\sigma}}} is given by:

   ku_0^2-2lu_0-2nv_0+m=O(\varepsilon).

   In other words the cycle {\breve C^{s}_{\breve{\sigma}}} passes focus (u_0,v_0) of the infinitesimal cycle in the parabolic space.

Let {\breve C^{s}_{\breve{\sigma}}} be the image under g\in\ensuremath{\ensuremath{\ifxSLl \ell_{2}^{}{}\else SL_{2}^{}{}\fi}(\ensuremath{\mathbb{R}^{}_{}{}})} of an infinitesimal cycle { C^{s}_{\breve{\sigma}}} from (6.1). Then \mathring{\sigma}-focus of {\breve C^{s}_{\breve{\sigma}}} is displaced from g(u_0,v_0) by infinitesimals of order \varepsilon^2 (while both cycles have \breve{\sigma}-radius of order \varepsilon).

Consequently \ensuremath{\ensuremath{\ifxSLl \ell_{2}^{}{}\else SL_{2}^{}{}\fi}(\ensuremath{\mathbb{R}^{}_{}{}})}-action is infinitesimally conformal in the sense of Defn. 6.2 with respect to the length from focus (Defn. 5.3) for all combinations of \sigma, \breve{\sigma} and \mathring{\sigma}.

1. Z^{}_{\infty} is the image of the zero-radius cycle Z^{}_{(0,0)}=(1,0,0,0) at the origin under reflection (inversion) into the unit cycle (1, 0,0,-1), see blue cycles in Fig. ?(b)-(d).

2. The following statements are equivalent

   1. A point (u,v)\in\ensuremath{\mathbb{R}^{\sigma}_{}{}} belongs to the zero-radius cycle Z^{}_{(0,0)} centred at the origin;

   2. The zero-radius cycle Z^{}_{(u,v)} is \sigma-orthogonal to zero-radius cycle Z^{}_{(0,0)};

   3. The inversion z\mapsto \frac{1}{z} in the unit cycle is singular in the point (u,v);

   4. The image of Z^{}_{(u,v)} under inversion in the unit cycle is orthogonal to Z^{}_{\infty}.

   If any from the above is true we also say that image of (u,v) under inversion in the unit cycle belongs to zero-radius cycle at infinity.

In the elliptic and parabolic cases the upper halfplane in \ensuremath{\mathbb{R}^{\sigma}_{}{}} is preserved by Möbius transformations from \ensuremath{\ensuremath{\ifxSLl \ell_{2}^{}{}\else SL_{2}^{}{}\fi}(\ensuremath{\mathbb{R}^{}_{}{}})}. However in the hyperbolic case any point (u,v) with v>0 can be mapped to an arbitrary point (u',v') with v'\neq 0.

1. In the elliptic case the “real axis” U is transformed to the unit circle and the upper half-plane—to the unit disk:

   \begin{aligned} \{(u,v) \,\mid\, v = 0\} &\to& \{ (u',v') \,\mid\, l^2_{c_e}(u'e_0+v'e_1)= u'^2+v'^2=1\} \\ \{(u,v) \,\mid\, v > 0\} &\to& \{(u',v') \,\mid\, l^2_{c_e}(u'e_0+v'e_1)= u'^2+v'^2<1\}, \end{aligned}

   where the length from centre l^2_{c_e} is given by (?) for \sigma=\breve{\sigma}=-1.

   On both sets \ensuremath{\ensuremath{\ifxSLl \ell_{2}^{}{}\else SL_{2}^{}{}\fi}(\ensuremath{\mathbb{R}^{}_{}{}})} acts transitively and the unit circle is generated, for example, by the point (0, 1) and the unit disk is generated by (0,0).

2. In the hyperbolic case the “real axis” U is transformed to the hyperbolic unit circle:

   \{(u,v) \,\mid\, v = 0\} \to \{ (u',v') \,\mid\, l^2_{c_h}(u',v')= u'^2-v'^2=-1\},

   where the length from centre l^2_{c_h} is given by (?) for \sigma=\breve{\sigma}=1.

   On the hyperbolic unit circle \ensuremath{\ensuremath{\ifxSLl \ell_{2}^{}{}\else SL_{2}^{}{}\fi}(\ensuremath{\mathbb{R}^{}_{}{}})} acts transitively and it is generated, for example, by point (0,1).

   \ensuremath{\ensuremath{\ifxSLl \ell_{2}^{}{}\else SL_{2}^{}{}\fi}(\ensuremath{\mathbb{R}^{}_{}{}})} acts also transitively on the whole complement

   \{(u',v') \,\mid\, l^2_{c_h}(u'e_0+v'e_1)\neq -1\}

   to the unit circle, i.e. on its “inner” and “outer” parts together.

Parabolic Cayley transform P_{\breve{\sigma}} as defined by the matrix C_{\breve{\sigma}} (?) acts on the V-axis always as a shift one unit down.

Its image can be described in term of various lengths as follows:

1. P_{\breve{\sigma}} for \breve{\sigma}\neq0 transforms the “real axis” U to the p-cycle with the p-length squared -\breve{\sigma} from its e-centre (0,-\frac{\breve{\sigma}}{2}), cf. (?):

   \{(u,v) \,\mid\, v = 0\} \to \{ (u',v') \,\mid\, l_{c_{e}}^2((0,-\frac{\breve{\sigma}}{2}), (u',v'))\cdot(-\breve{\sigma})=1\},

   where l_{c_{e}}^2((0,-\frac{\breve{\sigma}}{2}), (u',v'))= u'^2+\breve{\sigma} v', see (?).

   The image of upper halfplane is:

   \{(u,v) \,\mid\, v > 0\} \to \{(u',v') \,\mid\, l_{c_{e}}^2((0, -\frac{\breve{\sigma}}{2}), (u',v'))\cdot(-\breve{\sigma})< 1\}.

2. P_{\breve{\sigma}} with \breve{\sigma}\neq0 transforms the “real axis” U to the p-cycle with p-length squared -\breve{\sigma} (?) from its h-focus (0,-1-\frac{\breve{\sigma}}{4}), and the upper half-plane—to the “interior” part of it, cf. (?):

   \begin{aligned} \{(u,v) \,\mid\, v = 0\} &\to& \{ (u',v') \,\mid\, l_{f_{h}}^2((0,{\textstyle-1-\frac{\breve{\sigma}}{4}}), (u',v'))\cdot( -\breve{\sigma})=1\}, \\ \{(u,v) \,\mid\, v > 0\} &\to& \{(u',v') \,\mid\, l_{f_{h}}^2((0,{\textstyle-1-\frac{\breve{\sigma}}{4}}), (u',v'))\cdot( -\breve{\sigma})<1\}. \end{aligned}

3. P_{\breve{\sigma}} transforms the “real axis” U to the cycle with p-length -\breve{\sigma} from its p-focus (0,-1), and the upper half-plane—to the “interior” part of it, cf. (?):

   \begin{aligned} \{(u,v) \,\mid\, v = 0\} &\to& \{ (u',v') \,\mid\, l_{f_{p}}^2((0,-1), (u',v')) \cdot( -\breve{\sigma})=1\}, \\ \{(u,v) \,\mid\, v > 0\} &\to& \{(u',v') \,\mid\, l_{f_{p}}^2((0,-1), (u',v'))\cdot( -\breve{\sigma})<1\}, \end{aligned}

   where l_{f_{p}}^2((0,-1), (u',v'))=\frac{u'^2}{v'+1} (?).

N-orbits in the parabolic cases (Fig. ?(N_{P_{e}}, N_{P_p}, N_{P_{h}})) are concentric parabolas (or straight lines) in the sense of Defn. 2.2 with e-centres at (0,\frac{1}{2}), (0,\infty), (0,-\frac{1}{2}) correspondingly.

The Cayley transform of orbits of the parabolic fix subgroup in Fig. ?(N'_{P_e}) are parabolas consisting of points on the same l_{f_p}-length (?) from the point (0,-1), cf. 8.2.?.

Let { C^{s}_{a}} be a cycle in \ensuremath{\mathbb{R}^{\sigma}_{}{}}.

• (e, h) In the elliptic or hyperbolic cases the Cayley transform maps a cycle { C^{s}_{\breve{\sigma}}} to the composition of its inversion with the reflection {\hat C^{s}_{\breve{\sigma}}} { C^{s}_{\breve{\sigma}}}{\hat C^{s}_{\breve{\sigma}}} in the cycle { C^{s}_{\breve{\sigma}}}, where {\hat C^{s}_{\breve{\sigma}}}= \begin{pmatrix} \pm \breve{e}_1 && 1\\ 1 && \mp \breve{e}_1 \end{pmatrix} with \breve{\sigma}=\pm1 (see the first and last drawings on Fig. ?).

• (p) In the parabolic case the Cayley transform maps a cycle (k,l,n,m) to the cycle (k-2\breve{\sigma} n, l, n ,m+2\breve{\sigma} n).

1. A-orbits in transforms P_e and P_h are segments of parabolas with the focal length \frac{1}{2} passing through (0,-\frac{1}{2}). Their vertices belong to two parabolas v=\frac{1}{2}(-x^2-1) and v=\frac{1}{2}(x^2-1) correspondingly, which are boundaries of parabolic circles in P_h and P_e (note the swap!).

2. K-orbits in transform P_e are parabolas with focal length less than \frac{1}{2} and in transform P_h—with inverse of focal length bigger than -2.

An infinitesimal cycle { C^{a}_{\breve{\sigma}}} (6.1) is f-orthogonal (in the sense of Lem. 6.4.?) to a cycle {\tilde C^{a}_{\breve{\sigma}}} if and only if the Cayley transform 8.5(p) of { C^{a}_{\breve{\sigma}}} is f-orthogonal to the Cayley transform of {\tilde C^{a}_{\breve{\sigma}}}.