Slings and Spirals
Spirals whose curvatures are exponential functions of the arclength are denser near their centers than logarithmic spirals (see Figure 1). I stumbled upon these exponential spirals thanks to an inspiring conversation with mathematician Tadashi Tokieda.

In a previous SIAM News article, I discussed the following parody of the sling [1]. If we want to spin a stone (held by a rope) in a circle with a hand that is spinning in a smaller concentric circle—where the rope forms a fixed angle \(\alpha > \pi /2\) with the hand’s velocity—then we must accelerate the hand in direct proportion to its speed: \(\dot v = c v ^2\) (with some \(c= \textrm{const.}\) that depends on the circles’ radii and on \(\alpha\)). Such \(v=v(t)\) blows up in finite time, meaning that maintaining constant \(\alpha\) is possible only for that long.
Here is a reasonable modification of the problem: What path must my hand take if I no longer insist on moving it in a circle, but instead move it with constant speed (say, \(v=1\)) while still keeping \(\alpha = \textrm{const.}\)?1 An exponential spiral turns out to be the answer, and the following is a precise statement:
Tracking at a Constant Angle
Consider a pendulum (i.e., a point mass on a massless rod) in a plane, with no gravity (see Figure 2). If the pivot \(P\) moves with constant speed \(v=1\) in such a way as to keep the angle \(\alpha\) in Figure 2 constant, then the path of \(P\) is an exponential spiral with curvature
\[k = \omega_0 e^{-as}, \quad \textrm{where} \quad a = \frac{\cos \alpha}{r}. \tag1\]
Here, \(s=t\) is the arclength \(=\) time (recall that \(v=1\)), \(r\) is the rod length, and \(\omega_0\) is the initial angular velocity of the pendulum.
I omit the proof, which available upon request.
Some Observations
- If \(\alpha\) is obtuse (as in Figure 2) so that the pendulum lags, then \(k(s) \rightarrow \infty\) as \(s \rightarrow \infty\) according to \((1)\). The pivot moves in a tightening spiral to keep \(\alpha = \textrm{const.}\)
- The curvature \(k\) plays a double role in that it is also the angular velocity of the rod. Indeed, the angle between the rod and the velocity \(\mathbf{v}_P\) is constant. For obtuse \(\alpha\), \(k\) grows exponentially — as does the angular velocity of the rod. The bob then spirals towards a circle of radius \(r\), while \(P\) spirals towards a point. And since the bob’s speed approaches infinity, so does the force with which the hand at \(P\) pulls on the rod.
-
Figure 2. If \(P\) moves with constant speed to keep \(\alpha = \textrm{const.}\), then the curvature of \(P\)'s path is an exponential function of time (and thus of arclength).
Some Questions
- At what rate do the coils of an exponential spiral accumulate? (Providing a precise meaning of this rate is part of the question.)
- Does an exponential spiral have an asymptote? More generally, consider a curve that is defined by its curvature \(k=k(s)\). How fast must \(k\) approach \(0\) as \(s \rightarrow \infty\) (or \(- \infty\)) for the curve to have an asymptote?
Do exponential spirals arise in any other settings? It would seem likely that such natural objects exist somewhere in nature, but I couldn’t find any other examples.
1 As before, \(\alpha\) is the angle between the hand’s velocity vector and the rope.
2 Except, perhaps, for the initial part of the swing when one applies torque to the handle to begin the swing.
The figures in this article were provided by the author.
References
[1] Levi, M. (2018, January 29). Slings, bullets, blow-up, and linearity. SIAM News, 51(1), p. 6.
About the Author
Mark Levi
Professor, Pennsylvania State University
Mark Levi (levi@math.psu.edu) is a professor of mathematics at the Pennsylvania State University.
Related Reading



Stay Up-to-Date with Email Alerts
Sign up for our monthly newsletter and emails about other topics of your choosing.