Apollonius Circle

Imagine tracing a path where your distance from one landmark is always kept in a constant ratio to your distance from a second. What shape would you create? This intriguing geometric puzzle was solved over two millennia ago by the Greek geometer Apollonius of Perga, and the answer is a perfect circle. This "Apollonius circle" is more than a mathematical curiosity; it is a fundamental pattern that emerges unexpectedly across various scientific and mathematical landscapes. This article delves into the elegant world of the Apollonius circle, addressing the core principles behind this surprising geometric fact and its far-reaching implications.

First, in the "Principles and Mechanisms" chapter, we will formally define the Apollonius circle and use the power of complex numbers and Möbius transformations to understand why this constant ratio of distances generates a circle. We will explore the properties of the entire family of these circles, uncovering their hidden symmetries and structure. Following this, the "Applications and Interdisciplinary Connections" chapter will take us on a journey to find the Apollonius circle in the wild, revealing its role in describing physical phenomena like electric fields and fluid flows, and its deep connections to foundational concepts in complex analysis, partial differential equations, and even topology.

Imagine you are on a vast, flat plain at night. In the distance, you see two lighthouses, A and B. You have a special device that measures the ratio of the brightness of the light you receive from A to the brightness you receive from B. We know from physics that the intensity of light falls off with the square of the distance. This means that keeping the ratio of the light intensities constant is equivalent to keeping the ratio of your distances to the lighthouses constant. Let's say you decide to walk along a path where you are always exactly twice as far from lighthouse B as you are from lighthouse A. What shape would you trace in the dark? A straight line? An ellipse? Something more exotic?

The ancient Greek geometer Apollonius of Perga posed and solved this very question more than two millennia ago. The surprising and elegant answer is that your path would be a perfect circle. This is the ​​Circle of Apollonius​​.

Let's state this remarkable idea more formally. A Circle of Apollonius is the set of all points $P$ for which the ratio of the distance from $P$ to a fixed point $F_1$ to the distance from $P$ to another fixed point $F_2$ is a constant positive number, $k$. The two fixed points, $F_1$ and $F_2$, are called the ​​foci​​ of the circle.

$\frac{|PF_1|}{|PF_2|} = k$

There is one special case. What if $k=1$? If the distances are equal, $|PF_1| = |PF_2|$, you must be standing on the perpendicular bisector of the line segment connecting the two foci. This is a straight line. But for any other positive value of $k$, the locus of points is a circle. This simple definition hides a deep and beautiful geometric structure that we are about to uncover. And it's not just an abstract curiosity; this very principle can govern the navigation path of an autonomous submarine homing in on acoustic beacons.

But why is it a circle? It’s not at all obvious from just looking at the definition. Let's see if a little algebra can illuminate the hidden geometry. A powerful way to handle problems of distance and geometry is to use the complex plane, where points are represented by complex numbers. Let's place our foci at the points represented by complex numbers $z_1$ and $z_2$, and let our moving point be $z$. The definition becomes:

$|z - z_1| = k |z - z_2|$

This equation looks innocent enough. To get rid of the absolute value signs, which represent distance, we can square both sides. For any complex number $w$, we know that $|w|^2 = w \bar{w}$. Applying this gives:

$|z - z_1|^2 = k^2 |z - z_2|^2$ $(z - z_1)(\bar{z} - \bar{z}_1) = k^2 (z - z_2)(\bar{z} - \bar{z}_2)$

If we expand this expression, we get an equation relating $z$ and its conjugate $\bar{z}$: $z\bar{z} - z\bar{z}_1 - \bar{z}z_1 + z_1\bar{z}_1 = k^2(z\bar{z} - z\bar{z}_2 - \bar{z}z_2 + z_2\bar{z}_2)$

Now, let's gather all the terms on one side and group them: $(1-k^2)z\bar{z} - (z\bar{z}_1 - k^2z\bar{z}_2) - (\bar{z}z_1 - k^2\bar{z}z_2) + (|z_1|^2 - k^2|z_2|^2) = 0$

This equation may still look daunting, but it has the unmistakable form of a circle's equation in the complex plane. After some algebraic manipulation—the kind of satisfying work demonstrated in problems like and—it can be rearranged into the standard form $|z-c|^2 = R^2$. The algebra doesn't lie: the geometric locus defined by a constant ratio of distances is, indeed, a circle.

Now that we are convinced it's a circle, we can ask about its defining features: its center and radius. The same algebraic workout that proves it's a circle also yields a wonderfully compact formula for the complex number representing its center, $c$:

$c = \frac{z_1 - k^2 z_2}{1 - k^2}$

This formula is quite revealing. The center of the circle is not simply halfway between the foci; its position is a sort of "weighted average" of the foci, but with the strange-looking weights $1$ and $-k^2$. The presence of $k^2$ implies that the ratio $k$ has a powerful, non-linear influence on the circle's position.

Let's explore this. What about the circle corresponding to the ratio $1/k$? Its center, $c_{1/k}$, would be: $c_{1/k} = \frac{z_1 - (1/k)^2 z_2}{1 - (1/k)^2} = \frac{k^2 z_1 - z_2}{k^2 - 1}$

A remarkable symmetry is hidden here. If you take the midpoint of the centers of the two circles, $c_k$ and $c_{1/k}$, a delightful cancellation occurs: $\frac{c_k + c_{1/k}}{2} = \frac{1}{2} \left( \frac{z_1 - k^2 z_2}{1 - k^2} + \frac{k^2 z_1 - z_2}{k^2 - 1} \right) = \frac{1}{2(k^2 - 1)} \left( -(z_1 - k^2 z_2) + (k^2 z_1 - z_2) \right) = \frac{1}{2(k^2-1)} \left( (k^2-1)z_1 + (k^2-1)z_2 \right) = \frac{z_1 + z_2}{2}$

The midpoint of the centers is simply the midpoint of the foci themselves. This tells us that the circles for ratios $k$ and $1/k$ form a symmetric pair. For example, if the foci are at $-a$ and $a$ on the real axis, the center of the circle for $k=2$ and the center for $k=1/2$ will be at equal distances from the origin, but on opposite sides.

What happens if we consider all possible values of $k$? We don't get just one circle, but a whole ​​family​​ of them. For a fixed pair of foci $z_1$ and $z_2$, as we vary $k$ from just above 0 to very large numbers, we trace out a set of nested, non-intersecting circles. This is known as a ​​coaxal system​​ of circles.

Let's think about the extreme cases, which often reveal the most about a system.

• What happens as $k$ gets very close to 0? The condition $|z-z_1| \approx 0 \cdot |z-z_2|$ means that the distance $|z-z_1|$ must be practically zero. In the limit as $k \to 0$, our circle shrinks down to a single point: the focus $z_1$.
• What happens as $k$ becomes infinitely large? We can rewrite the definition as $|z-z_2| = (1/k)|z-z_1|$. As $k \to \infty$, the term $1/k \to 0$, so the condition becomes $|z-z_2| \approx 0$. In this limit, the circle shrinks down to the other focus, $z_2$.

This is a profound insight: the original foci are themselves members of the family they generate! They are the ​​limit points​​ of the system—in essence, "point-circles" of radius zero. They form the two bookends of our entire family of circles. And what about the straight line we found for $k=1$? It acts as the "spine" of this family, a circle of infinite radius that separates the circles with $k1$ (which enclose $z_1$) from the circles with $k>1$ (which enclose $z_2$). The way these families interact holds further secrets; for instance, two Apollonian circles from different families can be orthogonal, which leads to a beautifully simple relationship between their ratios.

So far, we have a beautiful but somewhat complex picture: a family of nested circles with two point-like limit points. Is there a way to see this more simply? Is there a change of perspective that makes this whole structure obvious? In mathematics, and especially in complex analysis, such "magic lenses" exist. They are called ​​Möbius transformations​​.

A Möbius transformation is a function of a complex variable $z$ of the form $f(z) = \frac{az+b}{cz+d}$. They are famous for their geometric superpower: they map circles and lines to other circles and lines.

Let's invent a transformation specifically tailored to our foci, $z_1$ and $z_2$: $w = f(z) = \frac{z - z_1}{z - z_2}$

Now, let's see what this transformation does to a point $z$ that lies on an Apollonian circle. The defining equation of that circle is $|z-z_1| = k|z-z_2|$, which we can rewrite as $\frac{|z-z_1|}{|z-z_2|} = k$. But the expression on the left is just the modulus of our transformation!

$|w| = \left| \frac{z - z_1}{z - z_2} \right| = \frac{|z - z_1|}{|z - z_2|} = k$

Look at what happened! The entire, seemingly complicated family of nested Apollonian circles in the $z$-plane has been transformed into a simple family of concentric circles, $|w|=k$, centered at the origin in the $w$-plane. This is a stunning simplification. It's like finding the perfect pair of glasses that makes a confusing picture perfectly clear. The two foci, $z_1$ and $z_2$, which were the limit points of the original family, are mapped to $w=0$ and $w=\infty$—the center and the point at infinity for the new concentric family. The perpendicular bisector (where $k=1$) is mapped to the unit circle $|w|=1$.

This transformation reveals the true, simple nature of the Apollonius circle family: it's just a warped version of the simplest family of circles imaginable. The complexity was an illusion of our coordinate system. This deep connection also explains why the ​​cross-ratio​​, a fundamental quantity in complex analysis that is invariant under Möbius transformations, can be used to define Apollonian circles. A condition like $|(z, 1; i, -i)| = 2$ is just a compact way of stating that a specific transformation maps the point $z$ to a circle of radius 2.

This "Apollonian" property is fundamental and robust. If you take any Apollonian circle and apply a different Möbius transformation, such as the inversion $M(z) = 1/z$, the result is yet another Apollonian circle, with transformed foci and a new ratio constant. The structure persists. Ultimately, this entire web of ideas can be connected to the classical geometric concept of ​​inversion​​, where the limit points are found to be a pair of points that are inverses of each other with respect to every single circle in the family, tying everything back together into a single, unified, and beautiful geometric picture.

Now that we have acquainted ourselves with the Apollonius circle through its formal definition, you might be tempted to file it away as a neat, but perhaps niche, piece of classical geometry. To do so would be to miss the real magic. The fun is not just in knowing what a thing is, but in discovering where it is. And the Apollonius circle, it turns out, is a surprisingly frequent guest in the house of nature. It appears, often in disguise, across different branches of science and mathematics, a recurring pattern that hints at a deep, underlying unity. Let us go on a hunt for it.

Our first stop is the world of physics, specifically the study of fields. Imagine two infinitely long, parallel wires. Let’s give one a positive electric charge and the other an equal negative charge, turning them into what physicists call line charges. Now, what does the landscape of electric potential look like in the space around them? A test charge placed very near the positive wire will feel a high potential, and near the negative wire, a low potential. What about the surfaces of constant potential—the equipotentials?

You might guess they are simple circles centered on each wire, but the influence of the other wire distorts them. And the shape of this distortion? It is perfect. Each and every equipotential surface is an Apollonius circle, with the two line charges serving as the foci! This is not an approximation; it is an exact mathematical consequence of the laws of electrostatics.

This provides a wonderfully clever trick known as the method of images. Suppose you have two non-concentric conducting cylinders, one held at a potential $V_1$ and the other at $V_2$. Calculating the electric field in the space between them seems horribly complicated. But we now know the secret: this physical setup is electrically indistinguishable from the field of two fictitious, cleverly placed line charges. The surfaces of the real cylinders are simply two of the Apollonian equipotentials generated by the imaginary line charges. By finding the right line charges, a fiendishly hard problem becomes manageable.

Now, let's change the scenery completely. Let's trade electricity for water and charges for vortices. Consider an ideal fluid and imagine a vortex spinning clockwise (a sink) and an equal and opposite vortex spinning counter-clockwise (a source) placed some distance apart. What path does a tiny speck of dust follow as it's carried by the flow? These paths are called streamlines. You can almost feel the punchline coming: the family of streamlines for this vortex pair forms a set of Apollonius circles.

The mathematics is so strikingly similar that one can be forgiven for mixing them up. The electrostatic potential in one problem plays the role of the stream function in the other. It is a beautiful example of how nature uses the same mathematical script for entirely different plays. The geometry of a constant ratio of distances governs both the invisible landscape of electric potential and the visible flow of a fluid.

The repeated appearance of this structure suggests we might be looking at it from the wrong angle. Physics gives us the "what," but mathematics can often give us the "why." The natural language for two-dimensional geometry and fields is the language of complex numbers.

In the complex plane, the definition of an Apollonius circle is breathtakingly simple: $|z - z_1| = k |z - z_2|$ for two fixed points (foci) $z_1$ and $z_2$ and a constant ratio $k$. This compact expression contains everything. When you work through the algebra, it blossoms into the familiar Cartesian equation of a circle, with a center and radius determined by $z_1$, $z_2$, and $k$.

But the real insight comes when we consider not just one family of curves, but two. The first is our family of Apollonian circles, defined by a constant ratio of distances to $z_1$ and $z_2$. The second is the family of all circles (and one straight line) that pass through both $z_1$ and $z_2$. If you draw these two families of curves together, an astonishing pattern emerges: they form a perfect grid. Every curve from the first family intersects every curve from the second family at a precise right angle ($\pi/2$ radians).

This is no coincidence. This orthogonal grid is the visual representation of the level curves of a single analytic function, $f(z) = \ln\frac{z-z_1}{z-z_2}$. The Apollonian circles are the curves where the real part of $f(z)$ (which is $\ln|z-z_1| - \ln|z-z_2|$) is constant, and the circles passing through the foci are where the imaginary part of $f(z)$ (the difference in angles) is constant. The fundamental rules of complex analysis—the Cauchy-Riemann equations—guarantee that these two sets of level curves must be orthogonal.

We can take this beautiful structure and view it on a grander scale. Using a map called the stereographic projection, we can wrap the entire infinite complex plane onto the surface of a sphere, the Riemann sphere. This map is conformal, meaning it preserves angles. Our orthogonal grid of Apollonian and through-point circles on the plane transforms into a beautiful, globally consistent grid of orthogonal circles on the sphere. The flat, infinite, and local picture becomes a finite, curved, and global one, all without losing its essential Apollonian character.

Having seen the Apollonius circle in the physical world and in the elegant machinery of complex analysis, we can ask if it appears at an even more fundamental level. It does.

In the study of partial differential equations, a powerful tool called the Green's function is used to understand how a "disturbance" at one point affects the potential or temperature everywhere else within a domain. For the unit disk in the complex plane—arguably the simplest and most important domain—if you fix a source point $w_0$ inside it, the level curves of the Green's function $G(z, w_0)$ are, once again, Apollonius circles. The foci in this case are the source point $w_0$ and its "image point" $1/\bar{w}_0$, a reflection across the boundary of the disk. This reveals that the Apollonian structure is not just about two arbitrary points; it is intrinsically tied to the way potentials behave inside boundaries.

As a final leap into abstraction, let's visit the field of topology, which studies the most fundamental properties of shape and space. A famous result, Urysohn's Lemma, states that if you have two disjoint closed sets in a "nice" space, you can always construct a continuous function that is, say, $0$ on one set and $1$ on the other. For the simple case of two points $p_A$ and $p_B$ in the Euclidean plane, one explicit way to build this function $f$ is through distance: $f(\mathbf{x}) = d(\mathbf{x}, p_A) / (d(\mathbf{x}, p_A) + d(\mathbf{x}, p_B))$.

What are the level sets of this function? Where is $f(\mathbf{x})$ equal to a constant $c$? A little algebra shows that this is equivalent to the condition $d(\mathbf{x}, p_A) / d(\mathbf{x}, p_B) = c / (1-c)$, which is precisely the definition of an Apollonius circle. So, this geometric object is hardwired into one of the most basic ways mathematicians have for distinguishing between two points in a space.

From electric fields to fluid flows, from the complex plane to the sphere, and from the solution of physical equations to the foundational concepts of topology, the Apollonius circle weaves its way through. It is far more than a classroom exercise; it is a fundamental pattern, a simple rule that nature, in its efficiency and elegance, seems to have discovered long before Apollonius of Perga ever did.