Framed holonomic knots (Q1597796)

A {holonomic plane curve} \(c(t)\) and a {holonomic space curve} \(C(t)\) associated to a smooth function \(f:S^1\to \mathbb R\) are \(c(t)=(f(t), f'(t))\subset \mathbb R^2\) and \(C(t)=(f(t), f'(t), f''(t))\subset \mathbb R ^3=(x_0, x_1, x_2)\). \textit{J.~Birman} and \textit{N.~Wrinkle} [J. Knot Theory Ramifications 9, No. 3, 293--309 (2000; Zbl 1001.57015)] showed that two holonomic knots that are topologically isotopic are holonomically isotopic. The authors study the category of framed holonomic knots \(C(t),\) i.e., holonomic knots such that \(c(t)\) is an immersion. In this case \(C(t)\) is equipped with the natural framing \(\partial x_2\). They construct examples of non-isotopic framed holonomic knots that are regular isotopic. This shows that the theory of framed holonomic knots is nontrivial, unlike the theory of ordinary holonomic knots. The authors show that the Whitney index \(W\) of a framed holonomic knot \(K\) is bounded above by the negative of the braid index of \(K\) and that for the self-linking number \(S\) of \(K\) the sum \(W+| S| \) is bounded from above by the negative of the Euler characteristic of a Seifert surface of \(K\). Surprisingly enough, the authors show that \(S\) restricted to framed holonomic knots with \(W=m\) splits into \(n=\lfloor \frac {| m| }{2}\rfloor\) invariants \(S_j, j=1, \dots, n,\) with \(\sum_j S_j=S\). They construct two framed holonomic knots \(K_1, K_2\) with \(W(K_1)=W(K_2), S(K_1)=S(K_2)\) that are regular isotopic but are distinguished by the invariants \(S_j\).