On Holditch’s theorem

In this paper we present a certain modification of the Holditch construction. This construction allows to consider a geometric family of pairs of ring domains. It is proved that the ratio of areas of ring domains of each pair belonging to this family is constant. Problems on extremal chords of constant length sliding around a given oval with both endpoints on it are also considered.


Introduction
Hamnet Holditch, president of Caius College in Cambridge, published in [13] a remarkable theorem. Let C be a convex curve, and a chord h of length a + b be divided into parts of lengths a and b by a point A. Let C a,b denote a curve traced out by the point A when the chord h slides around with both endpoints on C. Holditch proved that the area of a ring domain bounded by C and C a,b is equal to πab, see Fig 1. Arne Broman proved in [4] and [3] a much more general theorem and gave some kinematic applications. Further applications to mechanics were given in [11,12], and [14]. Some additional remarks on Holditch's theorem can be found also in [1,8,9], and [18], and recent related investigations are given in [15,16], and [10].
In this paper we modify the Holditch construction in which one ring domain is considered. In our modification we deal with a family of pairs of ring domains and obtain a natural geometric generalization. As an application we derive some Crofton-type formula for a ring domain.
We denote by C * the family of all closed strictly convex curves of class C 1 . Let C ∈ C * and let p denote a fixed support function of C. The parametric representation of the curve C has the form where the dot denotes differentiation with respect to t, see [2] and [17]. We denote by C a subfamily of C * defined as follows: a curve C ∈ C * belongs to C if and only if the function R = p +p satisfies the inequality Note that the function R is the curvature radius of C if the curve is of class We fix α ∈ (0, π), and we denote by z α (t) the intersection point of the tangent lines at z (t) and We will use the notations introduced in [6] and [7], namely where then we have see [7]. We denote by ξ α (t) the intersection point of the normal lines at z (t) and z (t + α). We have Simple calculations lead us to the formulas (1.13) Comparing (1.6) and (1.13), we get (1.14) Moreover, it is easy to verify that

Extremal chords
Let us fix a curve C ∈ C. We consider lengths of chords joining a fixed point z (t) and z (t + α) for α ∈ (0, 2π). Let We denote by −, − the Euclidean scalar product. Differentiating the function H t given by the formula (2.1) and making use of (1.8) and (1.15), we obtain The above formula implies immediately the following statement.
If the chord joining a fixed point z (t) and a point z (t + α) for some α ∈ (0, 2π) has maximal length, then the normal line at z (t + α) intersects C at z (t). Let Now we consider the particular but important case of ellipses.
Then the maximal length of chords is given by the formula Proof. For a given point P (r, s) of E we consider the normal line to E at P . This normal line intersects E at the second pointP (r,s), where Hence the distance between the points P andP is d(P,P ) = 2(b 4 r 2 + a 4 s 2 ) Since a 2 s 2 = a 2 b 2 − b 2 r 2 , it suffices to find the minimum of the function Example. Let us fix an ellipse E, x 2 a 2 + y 2 b 2 = 1. Each chord of the length m which slides around the curve E with endpoints on E determines some curve E m . We find the equation of the curve E m . For this aim we solve the system of equations where the second equation represents the line dual to an exterior point (u, v) with respect to E. Then its intersection points with E are where we substituted x, y instead of u, v, respectively.

Sliding a chord around a curve
We denote by N a family of functions ν : [0, +∞) → R of the class C 1 (0, +∞) satisfying the following conditions: Let C be a curve t → z(t) given by (1.1), and ν ∈ N be a function. We associate with C a vector field Q along the curve C, defined as follows: In view of (1.8) we have Differentiating (3.5) and using the formulas given in [7], we obtaiṅ We note that Hence we have equivalence of the following conditions: With respect to (3.9) we consider the implicit equation Differentiating the above equation and using the formulas (1.2) and (1.4), we get (R(t + Γ) − λ sin Γ)Γ = λ sin Γ − R(t + Γ) + R cos Γ, and thereforeΓ If the maximal width is attained at t = t 0 , then Γ(t 0 ) = π. We note that Γ(t) > π 2 , since for the orthoptic curve we have f = −μ = 0 and the considered function Γ is differentiable.
We associate with the curve C and ν ∈ N a curve C ν defined by

The main theorem
Now we assume that a chord of constant length m slides around with both endpoints on C which is given by formula (1.1). The endpoints of the sliding chord determine an increasing function ν ∈ N . We assume that |z(0)−z(t 0 )| = m for some t 0 ∈ (0, 2π). Thus the function ν satisfies the differential equation (3.9) with the initial condition ν(0) = t 0 . For a fixed ξ ∈ [0, 1] we consider a curve C(m, ξ) given by the formula Obviously, we have |Q| ≡ m. We note that Hence we get immediately We note that the graphs of the curves C(m, 0) and C(m, 1) coincide with the graph of C. Thus for ξ = 1 from (4.2) it follows immediately that Letting m = a + b and ξ = a a+b , we get the well-known Holditch formula Now we associate with C and C ν a certain curve D ν,γ defined as follows: where γ is a nonnegative constant.
As corollaries of Theorem 4.1 we have the following Holditch-type formulas.

Crofton-type integral formulas
In this section we provide an interesting application of the developed theory and derive a new, geometrically justified Crofton-type formula.
The above calculations imply that Now, we consider the ring domain CC r , and we introduce the notations as in Fig. 6, maintaining at the same time the notations of Santaló from [17].
Let R 1 (x, y), R 2 (x, y) denote the radii of curvature of C at the tangent points A 1 , A 2 , respectively.
Crofton proved the integral formula where ext C denotes the exterior of C, see [13]. We will prove some Croftontype theorem, namely Theorem 5.1. If C ∈ C and r < ww (C), then the following integral formula holds: Proof. We consider a mapping T : (0, 2π) × (0, r) → interior of CC r \{some segment} defined by the formula We note that T is a bijection and where h is some function. Thus the Jacobian JT of T at (t, m) is given by the formula On the other hand, since α (t, m) = ν (t, m) − t, so ∂ν ∂m = ∂α ∂m and sin 2 α ∂λ ∂m = sin 2 α ∂ ∂m Thus the Jacobian of T at (t, m) has the form

Now we have
CCr Open Access. This article is licensed under a Creative Commons Attribution 4.0 International License, which permits use, sharing, adaptation, distribution and reproduction in any medium or format, as long as you give appropriate credit to the original author(s) and the source, provide a link to the Creative Commons licence, and indicate if changes were made. The images or other third party material in this article are included in the article's Creative Commons licence, unless indicated otherwise in a credit line to the material. If material is not included in the article's Creative Commons licence and your intended use is not permitted by statutory regulation or exceeds the permitted use, you will need to obtain permission directly from the copyright holder. To view a copy of this licence, visit http://creativecommons. org/licenses/by/4.0/.
Publisher's Note Springer Nature remains neutral with regard to jurisdictional claims in published maps and institutional affiliations.