The Wonderful World of Organic Geometry

Spatial Visualization of Polar euclidean circles

This is an AI-generated translation of the text of Polareuklidische Kreise (PDF). It is intended to be read in conjunction with the figures in the German original.

Spatial Visualization of Polar-Euclidean Circles #

Thomas Neukirchner

October 2026


1 Polar-Euclidean Circles #

The starting point of this consideration is the question of how one can gain an understanding of the fact that in polar-Euclidean geometry, the dual of an ordinary circle is a conic section, one of whose foci coincides with the near point. The question immediately arises as to what significance the second focus then has when one dualizes the conic section back again.

In the case of conic sections, as the name itself suggests, it is often obvious to make use of three-dimensional space. A conic section with one of its foci can be represented spatially in the simplest way as follows: A circular cone intersects a plane in a conic section. A sphere is inscribed in the cone and the plane. Then, according to Dandelin, the point of contact $N$ of the sphere with the plane is one of the foci of the conic section.

Figure 1: Left: A dual-Euclidean circle as an oblique section of a right circular cone. Polar to it on the right: A Euclidean circle as a stereographic projection of an intersection circle with the sphere.

We regard the plane as a dual-Euclidean plane $\mathbb{D}$ with near point $N$. All dual-Euclidean circles in this plane result as sections with cones that touch this distinguished sphere. Strictly speaking, we must imagine the cone as being formed of planes whose section with the distinguished plane produces a line-based conic section.

The dualization of such a cone configuration should now lead to a Euclidean circle. For this purpose, we use the polarity at the distinguished sphere: The cone consists of all tangent planes of the sphere through the point $P$. The polar entity thus consists of all points of the sphere that are contained in the polar planes, that is, of the intersection points of the sphere with the plane $p$ polar to $P$, an intersection circle. Finally, the cone was intersected with the point and line field $\mathbb{D}$. Polarly, therefore, the intersection circle on the sphere is connected with the bundle $\mathbb{D}^*$ polar to $\mathbb{D}$, which has its carrier point in $N$. We are not used to doing Euclidean geometry in a bundle, so we intersect the bundle with a suitable plane. The plane polar to the near point $N$ is $\mathbb{D}$. Thus, this corresponds to the line at infinity in the bundle $\mathbb{D}^*$. Therefore, as the cutting plane, we must choose a plane parallel to $\mathbb{D}$, e.g., the plane at the south pole of the sphere. The resulting section is now indeed a Euclidean circle, since the projection of an intersection circle of the sphere from the north pole onto the tangent plane at the south pole is again a circle, as has been known since Ptolemy as the circle-preserving property of stereographic projection. A nice elementary geometric proof of this can be found in [Gla22] pp. 175f].


2 Center and Center Line of Circles #

Next, we want to investigate the question of how to find the center of the Euclidean circle and the center line of the dual-Euclidean circle in the above image. The center line is the polar to the near point, i.e., the directrix of the conic section with respect to this focus. This also follows directly from the Dandelin figure: It is the intersection of the planes $p$ and $\mathbb{D}$. Polarly to this, one connects the cone apex $P$ with $\mathbb{D}^*$ and projects into the plane $\mathbb{E}$.

Figure 2: Center line $m$ as the line of intersection $p \cap \mathbb{D}$ and center point $M$ corresponding to the ray $P \cup \mathbb{D}^*$.

This results in a simple picture of dual-Euclidean circles with a common center line corresponding to concentric circles in Euclidean geometry. For the latter, one considers a family of cones whose vertices lie on a fixed line $m^*$ through $N$. The polar planes are contained in the pencil of lines $m$ polar to $m^*$ and intersect the sphere in a family of concentric circles. The sections of the cones thus generate d-circles with the common center line $m$. One could call them con-peripheral.

Figure 3: A con-peripheral d-circle family with common center line $m$ and, polar to it, a concentric circle family.


3 Euclidean, Dual-Euclidean, Polar-Euclidean #

Let us take a slightly closer look at how polar-Euclidean geometry manifests itself in this spatial picture. The dual-Euclidean plane $\mathbb{D}$ with its distinguished near point $N$ is realized as the tangent plane at the north pole of a unit sphere. The polarity at this sphere first transforms the field $\mathbb{D}$ into a bundle $\mathbb{D}^*$, which becomes a Euclidean bundle due to the distinguished plane polar to $N$. By intersection with a plane $\mathbb{E}$ parallel to $\mathbb{D}$, we obtain the Euclidean plane with the ordinary position of the line at infinity. We choose the tangent plane at the south pole $S$ of the sphere for this purpose. However, this point is not a distinguished point in $\mathbb{E}$. Finally, polarizing again leads to a dual-Euclidean bundle $\mathbb{E}^*$.

The special feature of polar-Euclidean geometry is precisely that one has to think of the two planes $\mathbb{D}$ and $\mathbb{E}$ as united, as if looking from the point at infinity of the north-south axis and identifying corresponding points in $\mathbb{E}$ and $\mathbb{D}$ with each other. In other words: By reflecting across the equatorial plane, one interchanges the roles of $\mathbb{D}$ and $\mathbb{E}$. In this way, one can look at a dual-Euclidean configuration with Euclidean concepts; one can Euclideaniize it. We will make use of this in the next section. More on this in [Die21] and [Bau26].

Figure 4: Left: The spatial model of polar-Euclidean geometry presented here. Right: The second focus $F$ of a d-circle.


4 Confocal d-Circles and Conchordal e-Circles #

Next, we want to address the question of the second focus of a d-circle. It must be emphasized again at this point that within pure dual-Euclidean geometry, this question cannot be answered. In it, d-circles simply appear as arbitrary conic sections, without a distinguished position of the near point. Only the agreement that near rays to orthogonal points should also appear orthogonal in the Euclidean sense leads to the fact that the near point must lie in a focus of the conic section, because there conjugate rays are orthogonal. This property naturally also holds for the second focus, whereby with the consideration of orthogonal lines one has finally arrived at the Euclidean perspective.

According to Dandelin, the second focus $F$ also results as a point of contact with a sphere inscribed in the cone and the plane $\mathbb{D}$, see Fig. 4. From this figure, it becomes immediately clear that the second focus $F$ lies on the connecting line $PS$. The plane $f$ polar to it therefore passes through the intersection line of the plane $p$ with the south pole plane $\mathbb{E}$. How does this line behave relative to the polar circle?

Figure 5: Confocal conic sections as d-circles: one focus is $N$, the other an arbitrary fixed point. Polar to this, an Apollonian family of circles, i.e., a family of circles having a common radical axis and additionally containing the point $S$ as a circle with radius zero.

To get a feel for this, we put the figure in motion: Instead of a single d-circle, we consider the entire family of confocal conic sections with second focus $F$. This family results as the intersection of all cones tangent to the sphere whose vertex $P$ lies on the connecting line $PS$. The family of circles polar to it in $\mathbb{E}$ yields a familiar picture: They are obviously the circles of Apollonius, where $S$ is a point of vanishing radius (limiting point) of the circle family.

However, the family of circles of Apollonius can also be characterized as all circles with a common radical axis. This property of the circles is easy to see with the help of the spatial viewpoint: The radical axis of two intersecting circles is the line connecting the two points of intersection. In stereographic projection, the radical axis thus results from the projection of the line of intersection of the two planes whose intersection with the sphere corresponds to the two circles, see Figure 6 left. However, this property now also transfers to circles with imaginary intersection points.

Figure 6: The radical axis of two circles results as the projection of the line of intersection of their cutting planes.


5 Overview #

Finally, we want to give an overview of the Euclidean elements introduced here and their polar-Euclidean counterparts in the spatial model of a sphere.

  • (i) Dual-Euclidean plane $\mathbb{D}$ with near point $N$: Tangent plane at the north pole $N$ of the sphere. $\leftrightarrow$ Euclidean plane $\mathbb{E}$: Bundle $\mathbb{D}^*$ with carrier point $N$, intersected with the parallel plane at the south pole $S$.

  • (ii) d-circle: Point $P$ in space, connected with the sphere, intersected with $\mathbb{D}$.

  • $P$ on sphere $\Rightarrow$ Degeneration to a line.

  • $P$ in $\mathbb{D} \Rightarrow$ Degeneration to a point.

$\leftrightarrow$ e-circle: Plane $p$ in space, intersected with the sphere, projected from $\mathbb{D}^*$.

  • $p$ tangent to sphere $\Rightarrow$ Degeneration to a point.

  • $p$ in $\mathbb{D}^* \Rightarrow$ Degeneration to a line.

  • (iii) Center line of a d-circle: Intersection of the plane $p$ polar to $P$ with $\mathbb{D}$. $\leftrightarrow$ Center of an e-circle: Projection of the point $P$ polar to $p$ from $\mathbb{D}^*$.

  • (iv) Con-peripheral d-circles: Range of points $m^*$ containing the point $N$. The common center line is the line $m$ polar to it. $\leftrightarrow$ Concentric e-circles: Pencil of planes $m$ containing the plane $\mathbb{D}$. The common center belongs to the polar line $m^*$.

  • (v) Confocal d-circles: Range of points $m^*$ containing the south pole $S$. The intersection point with $\mathbb{D}$ is the common focus of the conic sections. $\leftrightarrow$ Conchordal e-circles containing $S$: Pencil of planes $m$ containing the plane $\mathbb{E}$. $m$ is the common radical axis of all circles.

The descriptions given here refer to the spatial description. As a projection into a polar-Euclidean plane, $N$ and $S$ coincide. The conchordal family of circles polar to the confocal d-circles thus contains the near point $N$ as a circle with radius zero, just as the confocal d-circle family contains the line at infinity as a degeneration.

Next, it would be worthwhile to look at the orthogonal family of circles to the Apollonian circles: This family of circles is also a conchordal family of circles. The spatial lines corresponding to these two families of circles are polar to each other. In particular, the relationship shown in Figure 7 between polar orthogonality and mutually polar spatial lines holds.

Figure 7: The projection from $\mathbb{D}^*$ to two mutually polar lines $g$ and $g^*$ yields an orthogonal pair of lines in $\mathbb{E}$. Polar to this, the intersection of the two lines with $\mathbb{D}$ yields an orthogonal pair of points.

Furthermore, one could consider general conchordal families of circles in which the near point $N$ is no longer part of the family of circles but has a general position. What properties do the d-circle families polar to them have?


6 The Sphere as an Imaginary Regulus #

Stereographic projection can also be transferred to other quadrics, such as a regulus (ruled surface): From a fixed point $O$ of the surface, all points of the surface are projected onto an arbitrary plane. The distinguished directrix and generator through $O$ thereby become two points $O'$ and $O''$; all other generators are projected onto the pencil at $O'$, and the directrices onto the pencil at $O''$. An arbitrary plane section of the ruled surface becomes, upon projection, a conic section through $O'$ and $O''$.

When transferring this construction to the sphere, the generators and directrices become conjugate complex lines. The two distinguished lines through the north pole $N$ meet the absolute quadric of Euclidean geometry in two points $I$ and $J$, which also lie in the opposite tangent plane $\mathbb{E}$ at the south pole $S$ of the sphere. The projections of plane sections of the sphere are thus projected onto conic sections that pass through the two points $I$ and $J$, i.e., onto circles! This makes the circle-preserving property of stereographic projection understandable from a higher standpoint.

Figure 8: Stereographic projection from the point $O$ of a regulus from [Kle26][p. 185].


References #

  • [Bau26] Peter Baum. Skizze der polareuklidischen Geometrie (Sketch of Polar-Euclidean Geometry). Unpublished manuscript, 2026.

  • [Die21] Immo Diener. Polareuklidische Geometrie (Polar-Euclidean Geometry). Springer Berlin Heidelberg, 2021.

  • [Gla22] Georg Glaeser. Geometrie und ihre Anwendungen in Kunst, Natur und Technik (Geometry and its Applications in Art, Nature, and Technology). Springer-Verlag Berlin Heidelberg, 2022.

  • [Kle26] Felix Klein. Vorlesungen über höhere Geometrie (Lectures on Higher Geometry). Springer-Verlag Berlin, 1926.