---
title: Semi-Discrete Conjugate Surfaces
url: https://www.emergentmind.com/topics/semi-discrete-conjugate-surfaces
type: topic
---

# Semi-Discrete Conjugate Surfaces

Semi-discrete conjugate surfaces are surfaces with one discrete and one smooth parameter for which the mixed smooth/discrete directions satisfy a conjugacy condition. The literature does not present a single universal definition. Instead, closely related formulations appear in semi-discrete conjugate nets \(x(k,t)\) on \(\mathbb Z\times\mathbb R\), in sequences of smooth curves whose neighboring strips are developable ruled surfaces, and in the analytic theory of hyperbolic differential-difference equations underlying classical conjugate-net geometry [1709.07373], [1208.4427], [2508.13343], [2506.18603].

## 1. Foundational definitions

A standard semi-discrete domain is a subdomain \(\mathbb D\subset \mathbb Z\times\mathbb R\), with discrete variable \(k\in\mathbb Z\), smooth variable \(t\in\mathbb R\), and notation
\[
x=x(k,t), \qquad x_1=x(k+1,t), \qquad \partial x=\frac{dx}{dt}, \qquad \Delta x=x_1-x.
\]
In one formulation, a semi-discrete surface is assumed to be a conjugate net when \(\partial x\), \(\partial x_1\), and \(\Delta x\) lie in a \(2\)-plane in the ambient \(V^4\); that \(2\)-plane is the tangent plane of the surface at the edge \([x,x_1]\). In another formulation, a semi-discrete surface is called conjugate when \(\partial x\), \(\Delta x\), and \(\partial\Delta x\) are linearly dependent [1709.07373], [1208.4427].

A strip-wise formulation replaces a pointwise net by a sequence of smooth curves. For
\[
U=\{0,\dots,n\}\times[0,T], \qquad F:U\to\mathbb R^d, \qquad F_i(t)=F(i,t),
\]
the strip between neighboring curves is represented by
\[
(u,v)\mapsto (1+i-u)F_i(v)+(u-i)F_{i+1}(v), \qquad u\in[i,i+1),\ v\in[0,T].
\]
A semi-discrete surface is then called conjugate when each such ruled strip is developable. Remark 2.3 gives the equivalent condition that \(\Delta F_i(t)\), \(\dot F_i(t)\), and \(\dot F_{i+1}(t)\) are linearly dependent for all \(i,t\); in \(\mathbb R^3\), this is
\[
\det\big(\Delta F_i,\dot F_i,\dot F_{i+1}\big)=0.
\]
This formulation makes explicit that the smooth direction is tangent to each curve, while the discrete direction is carried by rulings between neighboring curves [2508.13343].

Regularity is expressed by independence of the smooth and discrete directions. In the strip model, the pairs \((\dot F_i,\Delta F_i)\) and \((\dot F_i,\Delta F_{i-1})\) are required to be linearly independent for all \(i,t\). In the net-based literature, the same nondegeneracy appears through tangent-plane and curvature-line assumptions [2508.13343], [1709.07373].

## 2. Isothermic, Legendre, and minimal structures

A major branch of the subject places semi-discrete conjugate surfaces inside curvature-line and isothermic geometry. A semi-discrete Legendre immersion \((x,n)\) requires three compatibility conditions: \(\partial n\), \(\partial n_1\), and \(\Delta n\) lie in the tangent plane at \([x,x_1]\); \(\Delta x\), \(n_1\), and \(n\) lie in one \(2\)-dimensional plane; and \(n\) is perpendicular to \(\partial x\). Curvature-line parametrization is then imposed by
\[
\Delta n \parallel \Delta x, \qquad \partial x \parallel \partial n,
\]
together with the tangent cross ratio
\[
cr(x,x_1):= \partial x \cdot (\Delta x)^{-1} \cdot \partial x_1 \cdot (\Delta x)^{-1} < 0.
\]
The reality of this cross ratio implies the circularity condition: there is a circle through \(x\) and \(x_1\) tangent to \(\partial x\) at \(x\) and \(\partial x_1\) at \(x_1\) [1709.07373].

Semi-discrete isothermicity is characterized by factorization of the tangent cross ratio,
\[
cr(x(k,t),x(k+1,t))=\frac{\tau(t)}{\sigma(k)}<0,
\]
with \(\tau\) depending only on the smooth variable and \(\sigma\) only on the discrete variable. In the older minimal-surface formulation, a circular semi-discrete surface is isothermic when there exist positive functions \(\nu,\sigma,\tau\) such that
\[
\|\Delta x\|^2=\sigma\nu\nu_1,\qquad \|\partial x\|^2=\tau\nu^2,\qquad \partial \sigma=0,\quad \Delta\tau=0.
\]
These two descriptions belong to the same integrable framework of semi-discrete isothermic geometry [1709.07373], [1208.4427].

The mixed-area formalism extends curvature theory to pairs of compatible semi-discrete conjugate surfaces. For two semi-discrete conjugate surfaces \(x,y\) satisfying
\[
\partial x \parallel \partial y, \qquad \Delta x \parallel \Delta y,
\]
the mixed area is
\[
A(x,y) := \frac{1}{4} \big((\partial x+\partial x_1) \wedge \Delta y + (\partial y+\partial y_1) \wedge \Delta x\big).
\]
For a curvature-line parametrized semi-discrete surface \((x,n)\), Gaussian and mean curvature on edges are defined by
\[
A(n,n)=K\cdot A(x,x), \qquad A(x,n)=-H\cdot A(x,x).
\]
In the same setting, the paper gives explicit principal-curvature formulas and a duality relation between surface and normal,
\[
K^x=\frac{1}{K^n}, \qquad H^x=\frac{H^n}{K^n},
\]
together with parallel families \(x_\theta=x+\theta n\) in \(\mathbb R^3\) and \(\mathbb R^{2,1}\), and
\[
x_{\theta}:=\cosh \theta \cdot x +\sinh \theta \cdot n, \qquad
n_\theta :=\sinh \theta \cdot x+\cosh \theta \cdot n
\]
in \(\mathbb H^3\) and \(\mathbb S^{2,1}\) [1709.07373].

Minimality is expressed through duality. If \(x\) and \(x^*\) are conjugate semi-discrete surfaces, they are dual when
\[
\partial x^*=-\frac{1}{\nu^2}\partial x,\qquad \Delta x^*=\frac{1}{\nu\nu_1}\Delta x.
\]
A semi-discrete isothermic surface is minimal when \(x^*\) is inscribed in a sphere. The Weierstrass theorem states that every semi-discrete minimal surface is reconstructed from a semi-discrete holomorphic function \(g\), and conversely every semi-discrete minimal surface arises in that way. In that representation, the dual spherical surface is the inverse stereographic image of \(g\) [1208.4427].

## 3. Hyperbolic differential-difference equations and Darboux-Laplace theory

The analytic counterpart of semi-discrete conjugate geometry is a hyperbolic differential-difference equation. The relevant operator is
\[
{\cal L}_j=\partial_x T_n+a_{j,n}\,\partial_x+b_{j,n}T_n+c_{j,n},
\]
where \(n\in\mathbb Z\), \(x\in\mathbb R\), and \(T_n\psi_n(x)=\psi_{n+1}(x)\). This is the semi-discrete analog of the smooth hyperbolic operator \(\partial_x\partial_y+\cdots\) and the fully discrete hyperbolic operator \(T_nT_m+\cdots\) [2506.18603].

The operator admits two first-order factorizations,
\[
{\cal L}_j=(\partial_x+b_{j,n})(T_n+a_{j,n})+a_{j,n}k_{j,n},
\]
and
\[
{\cal L}_j=(T_n+a_{j,n})(\partial_x+b_{j,n-1})+a_{j,n}h_{j,n},
\]
where \(k_{j,n}\) and \(h_{j,n}\) are the semi-discrete Laplace invariants. Vanishing of \(k_{j,n}\) or \(h_{j,n}\) is equivalent to factorization. Under Laplace transformation, neighboring operators satisfy \(k_{j+1,n}=h_{j,n}\), and the invariants obey one form of the semi-discrete two-dimensional Toda lattice [2506.18603].

This framework is directly tied to conjugate-net theory. The paper starts from the classical fact that a parametrization of a surface in \(\mathbb R^3\) whose second fundamental form is diagonal is a conjugate net, and such nets satisfy a linear hyperbolic second-order PDE. In the semi-discrete case, the same role is played by a scalar hyperbolic differential-difference equation; the paper states that a semi-discrete surface coordinate function \(r_n(x)\) in an affine or projective ambient space is typically governed componentwise by exactly such an equation [2506.18603].

For finite Laplace series, the theory becomes explicit. If forward transforms terminate at \(h_r=0\) and backward transforms at \(k_{-s}=0\), then the general solution of \({\cal L}_0\psi=0\) has the form
\[
\psi_{0,n}=A_{0,n} N_n+A_{1,n} N_{n+1}+\dots+A_{r,n} N_{n+r}
+ B_{0,n} X+B_{1,n} X'+\dots+B_{s,n} X^{(s)},
\]
where \(N_n\) is an arbitrary function of the discrete variable and \(X(x)\) is an arbitrary function of the continuous variable. The paper also proves a semi-discrete Darboux determinant formula, giving the general solution as a single determinant built from derivatives of \(X\), shifts of \(N_n\), basis functions \(\xi_i(x)\), \(\nu_{i,n}\), and a gauge factor \(\omega_n\) [2506.18603].

A plausible implication is that finite-Laplace-series semi-discrete conjugate surfaces form the directly solvable subclass of the theory: the coordinate functions are reconstructed from arbitrary smooth data in the continuous direction and arbitrary shifted data in the discrete direction.

## 4. Developable strips, frameworks, and liftings

A geometric application of the strip model is the semi-discrete Maxwell-Cremona correspondence. A planar semi-discrete framework is a map
\[
f:\{-1,0,\dots,n+1\}\times[0,T]\to\mathbb R^2,
\]
that is, a discrete sequence of smooth planar curves \(f_i(t)\). A stress consists of functions \(\lambda_i(t)\) along the curves and \(\mu_i(t)\) along the strip direction between \(f_i\) and \(f_{i+1}\). Writing the tangential force as
\[
V_i(t)=\lambda_i(t)\dot f_i(t),
\]
the local equilibrium equation is
\[
\dot V_i+\mu_i\,\Delta f_i-\mu_{i-1}\,\Delta f_{i-1}=0,
\]
or equivalently
\[
\dot\lambda_i\,\dot f_i+\lambda_i\,\ddot f_i+\mu_i\,\Delta f_i-\mu_{i-1}\,\Delta f_{i-1}=0.
\]
A pair \((\lambda,\mu)\) satisfying this equation is a self-stress [2508.13343].

The paper defines a semi-discrete height function \(H_\gamma\) by summing discrete jump terms and continuous strip integrals along an increasing semi-discrete path. Theorem 3.4 states that \((\lambda,\mu)\) is a self-stress if and only if \(H_\gamma(f(p))\) is independent of the chosen path. This produces a lifting
\[
L(p)=(f(p),H_\gamma(f(p)))\in \mathbb R^2\times\mathbb R.
\]
For neighboring curves, the lifted vectors
\[
\Delta L_{k-1}(t),\quad \dot L_{k-1}(t),\quad \dot L_k(t)
\]
are linearly dependent, so each strip is developable and \(L\) is a semi-discrete conjugate surface [2508.13343].

The principal theorem states both directions of the correspondence: if \((\lambda,\mu)\) is a self-stress for a framework \(f\), then its semi-discrete lifting is a semi-discrete conjugate surface in \(\mathbb R^3\); conversely, if \(F\) is a semi-discrete conjugate surface in \(\mathbb R^3\) and its orthogonal projection \(f\) to \(\mathbb R^2\) is regular, then \(f\) is a stressable framework. In the abstract, this is summarized as the statement that stressable semi-discrete frameworks in the plane are precisely the orthogonal projections of semi-discrete conjugate surfaces in \(3\)-space [2508.13343].

The same paper records further geometric consequences. Frameworks with vanishing boundary forces \(\mu_{-1}=\mu_n=0\) have liftings with planar boundary curves. For a one-strip framework with no boundary forces, liftability is characterized by an explicit equation, labeled (4.1), involving derivatives of the two boundary curves [2508.13343].

## 5. Globally developable nets and rigid-ruling deformations

A more specialized theory studies globally developable semi-discrete conjugate nets arising from crease-rule patterns. The basic object is a sequence of smooth curves
\[
\gamma_0(t),\gamma_1(t),\dots,\gamma_{n+1}(t),
\]
with ruled strips between adjacent curves. A single ruled patch is written as
\[
X(t,u)=\gamma(t)+uR(t),
\]
where \(\gamma\) is the directrix and \(R(t)\in S^2\) is the unit ruling direction. Developability is imposed by
\[
\det(\gamma'(t),R(t),R'(t))=0.
\]
The paper interprets these objects simultaneously as curved-crease origami and as developable semi-discrete conjugate nets [2603.06420].

The central deformation problem is rigid-ruling folding: a continuous family of non-trivial folded states preserving the rulings, equivalently a conjugate-net-preserving isometry. For a single crease, the folded state is governed by
\[
\varphi'(t)=\frac12 s'(t)k(t)\bigl(\cot\theta_R(t)+\cot\theta_L(t)\bigr)\tan\varphi(t),
\]
\[
\tau(t)=\frac12 s'(t)k(t)\bigl(\cot\theta_R(t)-\cot\theta_L(t)\bigr)\tan\varphi(t),
\]
\[
K(t)=\frac{k(t)}{\cos\varphi(t)}.
\]
For multiple creases, compatibility across a common strip is encoded by equality of the ruling curvature
\[
V(t)=s'(t)k(t)\tan\varphi(t)\frac{1}{\sin\theta(t)}.
\]
The paper states that the ruling curvature determines the bend configuration of a developed patch up to Euclidean motion [2603.06420].

For a pair of creases, the main theorem gives necessary and sufficient conditions for a rigid-ruling folding motion. With \(I_i(t)\) and \(F_i(t)\) defined from the crease data, the conditions are
\[
\frac{F_1'(t)}{F_1(t)}-\frac{F_2'(t)}{F_2(t)}+I_1'(t)-I_2'(t)=0
\tag{A}
\]
and
\[
F_2(t)^2 I_1'(t)=F_1(t)^2 I_2'(t).
\tag{B}
\]
The paper also gives integrated versions \((A')\), \((B')\), and \((C)\). It proves a local-to-global assembly lemma: if each pair of adjacent creases with their incident surfaces can undergo a rigid-ruling folding motion, then the entire regular crease-rule pattern can too [2603.06420].

Several structural restrictions follow. If a candidate crease-rule pattern with rigid-ruling motion contains one constant fold-angle crease, then all creases are constant fold-angle creases. A Combescure transformation preserves existence of folded states, existence of rigid-ruling folding motions, and whether a crease is planar or constant fold-angle. The paper also derives explicit nonlinear third-order ODEs for appending new compatible creases in the cylindrical and conical cases, and notes that tangent-developable strips reduce by a semi-discrete Combescure transformation to the conical case. In each setting, the appended crease generally depends on three free initial values [2603.06420].

## 6. Related traditions, neighboring theories, and scope

The phrase “conjugate surface” is not used uniformly across the literature. In semi-discrete differential geometry, it usually refers to a parametrization or strip geometry, not necessarily to a transformed partner surface. One paper explicitly notes that it does not develop a separate theory explicitly called “semi-discrete conjugate surfaces” in the classical sense of pairs of conjugate surfaces, even though its basic objects are semi-discrete conjugate nets and its mixed-area formalism is defined for two semi-discrete conjugate surfaces satisfying
\[
\partial x \parallel \partial y, \qquad \Delta x \parallel \Delta y
\]
[1709.07373].

In the smooth CMC tradition, by contrast, “conjugate” often refers to sister-surface correspondences rather than conjugate-net parametrization. The Daniel sister correspondence establishes an isometric duality between minimal immersions
\[
\widetilde\phi:\Sigma\to \mathbb{E}(4H^2+\epsilon,H)
\]
and \(H\)-immersions
\[
\phi:\Sigma\to \mathbb{M}^2(\epsilon)\times\mathbb{R},
\]
with the phase-rotation relations
\[
\mathrm d\phi^{-1}(T)=J\,\mathrm d\widetilde\phi^{-1}(\widetilde T), \qquad S=J\widetilde S.
\]
This is structurally relevant to semi-discrete conjugate-surface theory, but it belongs to a different, fully smooth usage of “conjugate” [1802.04070].

A separate integrable bridge comes from the discretization principle via permutability. Starting from a smooth surface and two commuting transforms, one obtains a Bianchi quadrilateral and then a discrete net by evaluating transformed surfaces at a fixed point. The paper emphasizes that if one lets one variable remain the smooth parameter and only iterates one transform direction discretely, one gets a family \(f_m(x)\) depending smoothly on \(x\) and discretely on \(m\). It does not develop this semi-discrete geometry, but presents the mechanism as an obvious precursor [2603.21870].

Two neighboring theories supply additional context. “Principal binets” are fully discrete, not semi-discrete, but they separate conjugacy from orthogonality by using a pair of primal and dual conjugate nets, Möbius/Laguerre/Lie lifts, and a consistency principle on higher-dimensional lattices. Their direct contribution to semi-discrete theory is therefore structural rather than formal [2409.11322]. Likewise, the paper on explicit semi-discrete surfaces built from Jacobi elliptic functions does not define semi-discrete conjugate surfaces in the standard net-theoretic sense, but studies an adjacent integrable class of semi-discrete surfaces and discrete \(K\)-surfaces, with the edge relation
\[
\Gamma_{m+1}-\Gamma_m=\varepsilon\, B_{m+1}\times B_m
\]
as a central compatibility formula [2405.19619].

Taken together, these strands show that semi-discrete conjugate surfaces occupy an interface between projective conjugacy, curvature-line and isothermic geometry, developable-strip kinematics, and hyperbolic differential-difference equations. The field is technically unified by mixed smooth/discrete compatibility, but not by a single canonical definition.

Source: https://www.emergentmind.com/topics/semi-discrete-conjugate-surfaces