A relay rate is not an invariant physical speed

The finite scalar envelope $\Delta /\delta$ changes when the external tick duration changes. Lemma lem:normal gives the corresponding dependence of a normal deformation component on incidence parametrization.

A physical frame-invariance statement is different from either rescaling. It requires a declared action between frames and a proof that the same physical propagation bound is preserved. Neither a successful finite relay nor the choice of units proves that statement.

As an elementary directional negative control, nearest-neighbor cardinal moves in ${\mathbb{Z}}^2$ after $n$ ticks reach exactly the lattice points with

and appropriate parity if waiting is forbidden. After division by $n$, the limiting envelope is the ${\ell }^1$ unit ball, not the Euclidean disk. The maximum displacement in direction $(1,0)$ differs from that in the Euclidean unit direction $(1,1)/\sqrt{2}$. Thus a scalar radius bound alone does not establish isotropy.

Periodic-graph velocity polytopes provide broader context Fritz,FritzCorrigendum. We do not transfer a periodic-graph obstruction to the conical incidence package or to refinement families with changing directional sets.