Let's assume that $f$ is a quasiconformal homeomorphism of $\mathbb{C}$ with Beltrami coefficient $\mu = \frac{\bar{\partial} f}{\partial f}$. Notice that by definition $\Vert \mu \Vert _{L^{\infty}} < 1$. It is well known that if $\mu_n \rightarrow \mu$ in $L^{\infty}$ then $f_n \rightarrow f$ (the f's are all normalised to fix three points) in the compact open topology (hence pointwise at least). I believe (but am not sure) that the rate of convergence is also known.

My question is - In the above setting if $\mu_n \rightarrow \mu$ in $L^p$ for some $\infty >p>1$ then can we say (preferably with effective convergence estimates) that $f_n$ still goes to $f$ in the compact-open topology (or atleast pointwise)? It is also assumed that $\Vert \mu _n \Vert _{L^{\infty}} \rightarrow \Vert \mu \Vert _{L^{\infty}}$.


1 Answer 1


EDIT: The old answer to the originally asked question (without the additional assumption of convergence of the $L^\infty$ norms) is below.

I am not sure about effective convergence estimates, but the answer to the question is yes, even under slightly weaker assumptions. It directly follows from the Bers-Bojarski convergence theorem, which can be found in the book of Lehto and Virtanen. The statement of the theorem is that if $(f_n)$ is a sequence of $K$-qc mappings converging locally uniformly to a $K$-qc mapping $f$, and if their complex dilatations $\mu_n \to \mu$ a.e., then $\mu$ is a.e. the complex dilatation of $f$.

In the situation of your question, if $(f_n)$ is a sequence of normalized $K$-qc maps with dilatations $\mu_n\to\mu$ in $L^p$, any subsequence has a subsequence which converges locally uniformly to a $K$-qc limit (by compactness of the family of normalized $K$-qc maps.) Now you can pass to yet another subsequence where $\mu_n \to \mu$ a.e. Applying the Bers-Bojarski theorem gives that any such subsequential $K$-qc limit $f$ has complex dilatation $\mu$, which implies that the original sequence converges locally uniformly to $f$, where $f$ is the normalized qc map with complex dilatation $\mu$.

Note that you really only need $\mu_n \to \mu$ locally in measure, and uniform quasiconformality of the sequence $(f_n)$.

Old Answer

(Without the assumption that $\| \mu_n \|_\infty \to \| \mu \|_\infty$.)

I assume you want to have $f$ to be normalized in some way, otherwise convergence of the maps will not hold even in the case of $L^\infty$ convergence of the complex dilatations. However, even for normalized qc maps, without additional assumptions $L^p$ convergence of the complex dilatations does not imply convergence of the maps.

As an example, consider the family $f_k$ of quasiconformal mappings defined in polar coordinates as $(r,\theta) \mapsto (g_k(r),\theta)$, where $g_k$ is the identity for $r > 1$, $g_k(r) = r^{k^2}$ for $1-\frac1k \le r \le 1$, and $g_k$ is linear on $[0,1-1/k]$. Then $f_k$ is quasiconformal in the plane, $\mu_k$ is supported on the annulus $1-\frac1k \le r \le 1$, so $\mu_k \to 0$ in $L^p$ for any $p<\infty$, and $f_k$ converges to the identity outside the unit disk, and to the constant $0$ inside the disk, so the limit is not even continuous.

Maybe you want to require that the $f_k$ are uniformly quasiconformal, i.e., that there exists $\kappa<1$ such that $\| \mu_k \|_\infty \le \kappa$ for all $k$?

  • $\begingroup$ Fair enough. Indeed I want them to be uniformly quasiconformal. In fact, stronger than that. $\endgroup$
    – Vamsi
    Commented Nov 21, 2014 at 3:38
  • $\begingroup$ I added the answer to the updated question. $\endgroup$ Commented Nov 21, 2014 at 20:47

Your Answer

By clicking “Post Your Answer”, you agree to our terms of service and acknowledge you have read our privacy policy.

Not the answer you're looking for? Browse other questions tagged or ask your own question.