Prawitz’s area theorem and the mixed Aharonov sequence

Here we say g is a full mapping if the Lebesgue measure of \widehat{\mathbb{C}}\setminus g(\mathbb{D}_e) is zero.

Here we say f is a full mapping if the Lebesgue measure of \mathbb{C}\setminus f(\mathbb{D}) is zero.

From the above arguments, we see that Prawitz’s inequality (?) holds for all f\in \mathcal{U}(\mathbb{D}). The inequaity (?) has appeared in [HS-1], where (?) was used to study the integral means spectrum of univalent analytic functions in \mathbb{D}.

When \lambda=1, (?) reduces to

Here

is known as Grunsky kernel. We set

It is known that U(f; z,w) and U_f(z) also play an important role in the study of univalent analytic (meromorphic) functions and the universal Teichmüller space, see for example [Ba], [Ha], [Sh-1], [Sh-2], [TT].

When f\in \mathcal{M}(\mathbb{D}) has a first order pole z, we can define \phi_n(f; z) in view of (?) and (?) for all n \geq 1.

In the above theorems, the combinatorial number \binom{\alpha}{m} is defined as

Where \alpha\in \mathbb{R} and m is a non-negative integer number. In particular, we have

It should be pointed out that, in the recent paper [Su], Sugawa has reformulated above univalence criterion by using projective Schwarzian derivatives of higher order.

We get from (?) that \Phi_n(n\geq 0) is affine invariant in the sense that

for all affine transformation \chi(z)=az+b since \phi_0 is affine invariant and \phi_n(n\geq 1) is Möbius invariant. From this fact, we see that we can’t define \Phi_n(f; z) if f belongs to \mathcal{M}(\mathbb{D}) and z is a pole of f. We will call \{\Phi_n(f; z)\} as mixed Aharonov sequence. (?) explicitly shows how the mixed sequence is determined by the classical Aharonov invariants.

Note that (2.8) can be equivalently written as

This formulation has the advantage that both sides consist of non‑negative terms, making the role of low‑order coefficients explicit.

Due to the arbitrariness of \lambda>0, our new criteria are more flexible than the classical Aharonov criterion (which corresponds to the special case \lambda=1. The unified formulation also reveals a previously unnoticed connection between the Prawitz area theorem and Aharonov’s univalence criterion. It is hoped that the mixed Aharonov sequence and these criteria will find applications in future research.

For the well-known Koebe function

a simple computation yields that

which implies that

It follows from (Proof) and (2.9) with \zeta=0 that

for all \lambda>0. Here, the equality follows from the fact that \kappa is a full mapping. This means that our criterion reduces to an identity for all \lambda>0, while Aharonov’s original criterion only covers the single case \lambda=1. This also reveals the underlying role of the Prawitz area theorem.

We first remark that our results generalize some related ones of Aharonov in [Ah]. Let f\in \mathcal{L}(\mathbb{D}). When \lambda=1, from (?) and (?), we have

It follows that \Phi_n(f; z)=\phi_{n-1}(f; z) for n\geq 1. Then we see from [Ah, Theorem 2] that the inequality (2.3) is equivalent to

Where \mathbf{F} is defined as in Lemma 3.2.

On the other hand, when \lambda=1, we see from Lemma 3.2 that the inequality (2.8) is equivalent to

From these, we can see Theorem 2.8 as a generalization of Theorem 2.3 if we assume that f belongs to \mathcal{L}(\mathbb{D}). By the same reason, Theorem 2.9 and 2.10 also can be seen as some kind of extensions of Theorem 2.4 and 2.5, respectively.

In 1982, Harmelin proved the following property for the Aharonov sequence, see [Ha-1, Theorem 1].

We shall prove a similar result for the mixed Aharonov sequence \{\Phi_n\}.

For \lambda\in(0,1), the estimate in Proposition 4.2 reveals that the growth rate of |\Phi_n(f;z)| as |z|\to 1^- is at most (1-|z|^2)^{-n}, and the explicit bound is universal, depending only on n, |z|, and \lambda, not on the specific univalent function f.

The following result was obtained in [Ha-1, Lemma 5] by Harmelin.

By using (?) and (?), we can get the following inequality similar to (4.6).

Let k=\|\mu_f\|_{\infty}. Then it follows from (4.6) that

Here the last equality follows from the well-known integral

From (?), we see that, for any n \geq 1, it holds that

It is known that we can use the Schwarzian derivative S_f(z)=-\frac{6\Psi_1(f; z)}{(1-|z|^2)^2} and the quantity U_f(z) to describe some subclasses of the universal Teichmüller space. For example, the little Teichmüller space, see [GS], [Sh-1], [Sh-2]; Weil-Petersson Teichmüller space, see [B-1], [Sh-2], [Sh-3], [TT]; BMO(VMO)-Teichmüller space, see [AZ], [SW]. It is natural to study

If we can obtain an analogue inequality to (?), then we may use it to establish some new results about the theory of integral means spectrum.