A1A5802C8577
(The expanding mixed-sign quadric family) For $n\geq 3$, $n/2<k<n$, and $S=2k-n$, set
On $(Q_-\times \mathbb R)/((x,s)\sim (-x,s+\pi ))$ this is a complete proper embedded Lagrangian in $\mathbb C^ n$, of dimension and codimension $n$. It satisfies $\vec H=X^\perp /2$, and $L_ t=\sqrt t\, X(Q_-\times \mathbb R)$ for $t>0$. This is an exact parametrization subject to one quadric constraint, not an ODE approximation. See [LeeWang2010], Propositions 2.1–2.2, with weights $+1$ and $-1$.
The fiber is $\mathbb R^ k\times S^{n-k-1}$. The mapping-torus quotient is connected and has one end. If $n-k=1$, its two fiber components are exchanged by the gluing, and the total manifold is diffeomorphic to $\mathbb R^ k\times S^1$, hence orientable. For $n-k\geq 2$ the fiber is connected, the antipodal monodromy must be retained, and the total manifold is orientable exactly when $n$ is even. In particular, the $n=3,k=2$ case is orientable: one must pass around the base twice to return to the same component. These topology statements are obtained from the quotient directly; the product and parity shorthand in the source's Proposition 2.2 requires this disconnected-fiber qualification.
The family is noncompact and Hamiltonian stationary, with $SO(k)\times SO(n-k)$ and circle symmetries, and is asymptotic to the same cone as Example 2.50. Its planar-coordinate quadric description gives polynomial volume growth and finite centered Gaussian area; this entry records no entropy value or stability result. No surface genus is assigned to these $n\geq 3$ examples. The source's weak continuation at $t=0$ is a separate varifold statement; the smooth time interval here is $t>0$.
For the illustrated case $n=3$, $k=2$, the full expander has the global parametrization
This parametrization is obtained directly from the unit-weight construction in [LeeWang2010], Theorem 1.1 and Section 2.2: one takes the positive component of the real quadric and a full circle of phases. The real slice, obtained at $s=0$ and $s=\pi $, consists of the two graphs
Their tips are at $x_3=\pm \sqrt{2t}$. The figure compares $t=1/4,1,4$ with the singular cone limit $t=0$, using the same spatial scale and the fixed display cutoff $r\leq 3.2$. The complete slices are unbounded. These are exact two-dimensional slices of a three-dimensional flow in $\mathbb R^6$; they are not themselves mean curvature flows in $\mathbb R^3$.
The circle direction explains why two real sheets do not mean two components of the expander. For a point $x=(x_1,x_2,x_3)$ on the positive real sheet, the curve
lies in $L_ t$ and joins $x$ to $-x$ on the other real sheet. It is a semicircle in a real two-plane in $\mathbb R^6$, since the two spanning vectors are orthogonal and have equal length. The circle-phase control follows this path at fixed time; only the time control changes the flow.
Comments
Comments are not enabled yet.