“Resterait à voir si cet ensemble, sans être continu, peut avoir la puissance du continu”.^{
$\text{1}$
}
According to Alan Sokal (private communication), the same question arises in holomorphic dynamics, when one tries to extend to holomorphic families of transcendental entire functions the results of Lyubich [8] on holomorphic families of rational functions.

In this paper, we show that in general, the result of Lelong and Tsuji is best possible: every closed set of zero capacity can occur as an exceptional set (Theorem 1). Then we study a related problem of dependence of Picard exceptional values of the function
$z\mapsto f(z,w)$
on the parameter
$w$
(Theorem 2).

It is known that the exceptional set
$P$
is discrete in the important case that
$z\mapsto f(z,w)$
are functions of finite order. This was discovered by Lelong in [6] and later the result was generalized to the case of multi-dimensional parameter
$w$
in [7,Thm.3.44] .

We also mention that the set
$P$
has to be analytic in certain holomorphic families of entire functions with finitely many singular values, considered in [1, 2] . These families may consist of functions of infinite order.

We begin with a simple proof of a version of Lelong's theorem on functions of finite order.

Proposition 1. Let
$D$
be a complex manifold, and
$f:\mathbf{C}\times D\to \mathbf{C}$
an analytic function, such that the entire functions
$z\mapsto f(z,w)$
are not identically equal to zero and are of finite order for all
$w\in D$
. Then the set

is analytic.

$$\begin{array}{c}P=\{w\in D:(\forall z\in \mathbf{C}\left)f\right(z,w)\ne 0\}\end{array}$$ | (1) |

Corollary. Let
$D$
be a region in
$\mathbf{C}$
, and
$f:\mathbf{C}\times D\to \mathbf{C}$
an analytic function, such that entire functions
$z\mapsto f(z,w)$
are of finite order for all
$w\in D$
. Then the set
$P$
as in 1
$\left(\right)$
is discrete or
$P=D$
.

Indeed, the set
$A=\{w\in D:f(.,w)\equiv 0\}$
is discrete (unless
$f=0$
when there is nothing to prove). So there exists a function
$g$
holomorphic in
$D$
whose zero set is
$A$
, and such that
$f/g$
satisfies all conditions of Proposition 1.

$\square $
Remarks. 1. In general, when
$D$
is of dimension greater than one, and the set
$A$
is not empty, one can only prove that
$A\cup P$
is contained in a proper analytic subset of
$D$
, unless
$A\cup P=D$
, [7,Thm3.44] .

2. If the order of
$f(.,w)$
is finite for all
$w\in D$
then this order is bounded on compact subsets of
$D$
[7,Thm.1.41] .

Proof of Proposition 1. We assume without loss of generality that

(shift the origin in
$\mathbf{C}$
and shrink
$D$
, if necessary), and that the order of the function
$f(.,w)$
does not exceed
$\lambda $
for all
$w\in D$
(see Remark 2 above).

$$\begin{array}{c}f(0,w)\ne 0\text{for}w\in D.\end{array}$$ | (2) |

Let
$p$
be an integer,
$p>\lambda $
. Then, for each
$w$
,
$f$
has the Weierstrass representation
$$f(z,w)={e}^{{c}_{0}+\dots +{c}_{p}{z}^{p}}{\prod}_{a:f(a,w)=0}\left(1-\frac{z}{a}\right){e}^{z/a+\dots +{z}^{p}/p{a}^{p}},$$
where
$a$
are the zeros of
$f(.,w)$
repeated according to their multiplicities, and
${c}_{j}$
and
$a$
depend on
$w$
. Taking the logarithmic derivative, differentiating it
$p$
times, and substituting
$z=0$
, we obtain for each
$w\in D$
:

$${F}_{p}\left(w\right)={\frac{{d}^{p}}{d{z}^{p}}\left(\frac{df}{fdz}\right)|}_{z=0}=p!{\sum}_{a:f(a,w)=0}{a}^{-p-1}.$$
The series in the right hand side is absolutely convergent because of our choice of
$p$
. The functions
${F}_{p}$
are holomorphic in
$D$
, in view of ( 2 ). Clearly
$w\in P$
implies
${F}_{p}\left(w\right)=0$
for all
$p>\lambda $
. In the opposite direction,
${F}_{p}\left(w\right)=0$
for all
$p>\lambda $
means that all but finitely many derivatives with respect to
$z$
at
$z=0$
of the unction
$df/fdz$
meromorphic in
$\mathbf{C}$
are equal to zero, so this meromorphic function is a polynomial, and thus
$f(z,w)=exp({c}_{0}+\dots +{c}_{p}{z}^{p}),$
that is
$w\in P$
.
$\square $
The following result is due to Lelong and Tsuji but they state it only for the case
$dimD=1$
, and we need a multi-dimensional version in the proof of Theorem 2 below.

Proposition 2. Let
$D$
be a connected complex manifold, and
$f:\mathbf{C}\times D\to \mathbf{C}$
an analytic function, such that the entire functions
$z\mapsto f(z,w)$
are not identically equal to zero. Then the set
$$P=\{w\in D:(\forall z\in \mathbf{C}\left)f\right(z,w)\ne 0\}$$
is closed and pluripolar or coincides with
$D$
.

Proof. Suppose that
$P\ne D$
. It is enough to show that every point
${w}_{0}\in D$
has a neighborhood
$U$
such that
$P\cap U$
is pluripolar. By shifting the origin in
${\mathbf{C}}_{z}$
and shrinking
$U$
we may assume that
$f(0,w)\ne 0$
for all
$w\in U$
. Let
$r\left(w\right)$
be the smallest of the moduli of zeros of the entire function
$z\mapsto f(z,w)$
. If this function has no zeros, we set
$r\left(w\right)=+\infty $
. We are going to prove that
$logr$
is a continuous plurisuperharmonic function.

First we prove continuity of
$r:U\to (0,+\infty ]$
. Indeed, suppose that
$r\left({w}_{0}\right)<\infty $
, and let
$k$
be the number of zeros of
$f(.,{w}_{0})$
on
$\left|z\right|=r\left({w}_{0}\right)$
, counting multiplicity. Let
$\epsilon >0$
be so small that the number of zeros of
$f(.,{w}_{0})$
in
$\left|z\right|\le r\left({w}_{0}\right)+\epsilon $
equals
$k$
. Then
$${\int}_{\left|z\right|=r\left({w}_{0}\right)\pm \epsilon}\frac{df}{fdz}dz=\{\begin{array}{c}2\pi ik,\\ 0.\end{array}$$
As the integrals depend on
$w$
continuously, we conclude that
$r$
is continuous at
${w}_{0}$
. Consideration of the case
$r\left({w}_{0}\right)=+\infty $
is similar.

Now we verify that the restriction of
$logr$
to any complex line is superharmonic.

Let
$\zeta \to w\left(\zeta \right)$
be the equation of such line, where
$w\left(0\right)={w}_{0}$
. Let
${z}_{0}$
be a zero of
$f(.,{w}_{0})$
of the smallest modulus. We verify the inequality for the averages of
$logr$
over the circles
$\left|\zeta \right|=\delta $
, where
$\delta $
is small enough. According to the Weierstrass Preparation theorem, the set
$Q=\left\{\right(z,\zeta ):f(z,w\left(\zeta \right))=0\}$
is given in a neighborhood of
$({z}_{0},0)$
by an equation of the form
$$(z-{z}_{0}{)}^{p}+{b}_{p-1}(\zeta \left)\right(z-{z}_{0}{)}^{p-1}+\dots +{b}_{0}\left(\zeta \right)=0,$$
where
${b}_{j}$
are analytic functions in a neighborhood of
$0$
,
${b}_{j}\left(0\right)=0$
. We rewrite this as

where
${c}_{j}$
are analytic functions, and

Let
$V$
be a punctured disc around
$0$
; we choose its radius so small that
${c}_{0}\left(\zeta \right)\ne 0$
in
$V$
. For
$\zeta $
in
$V$
, let
${z}_{i}\left(\zeta \right),i=1,\dots ,p$
be the branches of the multi-valued function
$z\left(\zeta \right)$
defined by equation ( 3 ). Then

where we used ( 4 ) and harmonicity of
$log\left|{c}_{0}\right|$
in
$V\cup \left\{0\right\}$
. This completes the proof of plurisuperharmonicity. As
$P=\{w:logr(w)=+\infty \}$
we conclude that
$P$
is pluripolar.
$\square $
Our first theorem answers the question of Julia; it shows that the restriction of finiteness of order cannot be removed in Proposition 1, and that Proposition 2 is best possible, at least when
$dimD=1$
.

$$\begin{array}{c}{z}^{p}+{c}_{p-1}\left(\zeta \right){z}^{p-1}+\dots +{c}_{0}\left(\zeta \right)=0,\end{array}$$ | (3) |

$$\begin{array}{c}{c}_{0}\left(0\right)=(-{z}_{0}{)}^{p}.\end{array}$$ | (4) |

$$\begin{array}{c}\frac{1}{2\pi}{\int}_{-\pi}^{\pi}logr\left(w\right(\delta {e}^{i\theta}\left)\right)d\theta \le \frac{1}{2\pi}{\int}_{-\pi}^{\pi}{min}_{i}log\left|{z}_{i}\right(\delta {e}^{i\theta}\left)\right|d\theta \end{array}$$ |

$$\begin{array}{c}\end{array}$$ |

$$\begin{array}{c}\le \frac{1}{2\pi p}{\int}_{-\pi}^{\pi}log{\prod}_{i}\left|{z}_{i}\right(\delta {e}^{i\theta}\left)\right|d\theta =\frac{1}{2\pi p}{\int}_{-\pi}^{\pi}log\left|{c}_{0}\right(\delta {e}^{i\theta}\left)\right|d\theta \end{array}$$ |

$$\begin{array}{c}\end{array}$$ |

$$\begin{array}{c}=log\left|{z}_{0}\right|=logr\left({w}_{0}\right),\end{array}$$ |

Theorem 1. Let
$D\subset \mathbf{C}$
be the unit disc, and
$P$
an arbitrary compact subset of
$D$
of zero capacity. Then there exists a holomorphic function
$f:\mathbf{C}\times D\to \mathbf{C}$
, such that for every
$w\in P$
the equation
$f(z,w)=0$
has no solutions, and for each
$w\in D\backslash P$
it has infinitely many solutions.

It is not clear whether a similar result holds with multidimensional parameter space
$D$
and arbitrary closed pluripolar set
$P\subset D$
.

Proof. Let
$\phi :D\to D\backslash P$
be a universal covering. Let
$S$
be the set of singular points of
$\phi $
on the unit circle. Then
$S$
is a closed set of zero Lebesgue measure.

(We recall a simple proof of this fact. As a bounded analytic function,
$\phi $
has radial limits almost everywhere. It is easy to see that a point where the radial limit has absolute value
$1$
is not a singular point of
$\phi $
. Thus the radial limits exist and belong to
$P$
almost everywhere on
$S$
. Let
$u$
be the “Evans potential” of
$P$
, that is a harmonic function in
$D\backslash P$
, continuous in
$\overline{D}$
and such that
$u\left(\zeta \right)=0$
for
$\zeta \in \partial D$
and
$u\left(\zeta \right)=-\infty $
for
$\zeta \in P$
. Such function exists for every compact set
$P$
of zero capacity. Now
$v=u\circ \phi $
is a negative harmonic function in the unit disc, whose radial limits on
$S$
are equal to
$-\infty $
, thus
$\left|S\right|=0$
by the classical uniqueness theorem.) According to a theorem of Fatou, (see, for example, [3,Ch.VI] ), for every set
$S$
of zero Lebesgue measure on
$\partial D$
, there exists a holomorphic function
$g$
in
$D$
, continuous in
$\overline{D}$
and such that
$$\{\zeta \in \overline{D}:g(\zeta )=0\}=S.$$
In particular,
$g$
has no zeros in
$D$
.

Now we define the following set
$Q\subset \mathbf{C}\times D$
:

$$Q=\left\{\right(1/g\left(\zeta \right),\phi \left(\zeta \right)):\zeta \in D\}.$$
It is evident that the projection of
$Q$
on the second coordinate equals
$D\backslash P$
.

It remains to prove that the set
$Q$
is analytic. For this, it is enough to establish that the map
$$\Phi :D\to \mathbf{C}\times D,\Phi \left(\zeta \right)=(1/g(\zeta ),\phi (\zeta \left)\right)$$
is proper. Let
$K\subset \mathbf{C}\times D$
be a compact subset. Then the closure of
${\Phi}^{-1}\left(K\right)$
in
$\overline{D}$
is disjoint from
$S$
because
$g$
is continuous and
$1/g\left(\zeta \right)\to \infty $
as
$\zeta \to S$
.

On the other hand, for every point
$\zeta \in \partial D\backslash S$
, the limit
$${lim}_{\zeta \to {\zeta}_{0}}\phi \left(\zeta \right)$$
exists and has absolute value
$1$
. So
${\Phi}^{-1}\left(K\right)$
is compact in
$D$
. Now the existence of the required function
$f$
follows from the solvability of the Second Cousin problem [4,sect.5.6] .
$\square $
Notice that the map
$\Phi $
constructed in the proof is an immersion.

We recall that a point
$a\in \mathbf{C}$
is called an exceptional value of an entire function
$f$
if the equation
$f\left(z\right)=a$
has no solutions. Picard's Little theorem says that a non-constant entire function can have at most one exceptional value.

Let
$f$
be an entire function of
$z$
depending of the parameter
$w$
holomorphically, as in Propositions 1 and 2, and
$\text{assume in the rest of the paper that for all}w\in D\text{,}f(.,w)\ne const\text{.}$
Let
$n\left(w\right)\in \{0,1\}$
be the number of exceptional values of
$f(.,w)$
.

Question 1. What can be said about
$n\left(w\right)$
as a function of
$w$
?

Example 1.
$f(z,w)={e}^{z}+wz$
. We have
$n\left(0\right)=1$
and
$n\left(w\right)=0$
for
$w\ne 0$
.

Example 2.
$f(z,w)=({e}^{wz}-1)/w$
for
$w\ne 0$
, and
$f(z,0)=z$
. We have
$n\left(0\right)=0$
, while
$n\left(w\right)=1$
for
$w\ne 0$
. The exceptional value
$a\left(w\right)=-1/w$
tends to infinity as
$w\to 0$
.

Thus
$n$
is neither upper nor lower semicontinuous.

Example 3. Let
$$f(z,w)={\int}_{-\infty}^{z}(\zeta +w){e}^{-{\zeta}^{2}/2}d\zeta ,$$
where the contour of integration consists of the negative ray, passed left to right, followed by a curve from
$0$
to
$z$
. We have
$f(z,0)=-{e}^{-{z}^{2}/2}$
, which has exceptional value
$0$
, so
$n\left(0\right)=1$
. It is easy to see that there are no exceptional values for
$w\ne 0$
, so
$n\left(w\right)=0$
for
$w\ne 0$
. Thus
$n\left(w\right)$
is the same as in Example 1, but this time we have an additional feature that the set of singular values of
$f(.,w)$
is finite for all
$w\in \mathbf{C}$
, namely, there is one critical value
$f(-w,w)$
and two asymptotic values,
$0$
and
$\sqrt{2\pi}w$
. Question 2. Suppose that
$n\left(w\right)\equiv 1$
, and let
$a\left(w\right)$
be the exceptional value of
$f(.,w)$
. What can be said about
$a\left(w\right)$
as a function of
$w$
?

For functions of finite order, Question 2 was addressed by Nishino [9] who proved the following:

Let
$f$
be an entire function of two variables such that
$z\mapsto f(z,w)$
is a non-constant function of finite order for all
$w$
. If
$n\left(w\right)=1$
for all
$w$
in some set having a finite accumulation point, then there exists a meromorphic function
$\stackrel{~}{a}\left(w\right)$
such that
$a\left(w\right)=\stackrel{~}{a}\left(w\right)$
when
$a\left(w\right)\ne \infty $
, and
$f(.,w)$
is a polynomial when
$\stackrel{~}{a}\left(w\right)=\infty .$
Example 4. (Nishino) Let
$f(z,w)=({e}^{w{e}^{z}}-1)/w,w\ne 0$
and
$f(z,0)={e}^{z}.$
Then
$f$
is an entire function of two variables and
$n\left(w\right)\equiv 1$
. However
$a\left(w\right)=-1/w,w\ne 0$
and
$a\left(0\right)=0$
, so
$a$
is a discontinuous function of
$w$
.

Nishino also proved that for arbitrary entire function
$f$
of two variables, with
$n\left(w\right)=1$
in some region
$D$
, the set of discontinuity of the function
$a\left(w\right)$
is closed and nowhere dense in
$D$
.

Our Theorem 2 below gives a complete answer to Question 2. We first prove the following semi-continuity property of the set of exceptional values which holds for all meromorphic functions holomorphically depending on parameter. We always assume that
$z\mapsto f(z,w)$
is non-constant for all
$w$
. Denote
$$A\left(w\right)=\{a\in \overline{\mathbf{C}}:(\forall z\in \mathbf{C}\left)f\right(z,w)\ne a\}.$$
Proposition 3. For every
${w}_{0}\in D$
and every
$\epsilon >0$
there exists
$\delta >0$
such that
$|w-{w}_{0}|<\delta $
implies that
$A\left(w\right)$
is contained in the
$\epsilon $
-neighborhood of
$A\left({w}_{0}\right)$
with respect to the spherical metric.

Proof. Let
$U$
be the open
$\epsilon $
-neighborhood of
$A\left({w}_{0}\right)$
. Then
$K=\overline{\mathbf{C}}\backslash U$
is compact, so there exists
$r>0$
such that the image of the disc
$\left|z\right|<r$
under
$f(.,{w}_{0})$
contains
$K$
. Then by Hurwitz's theorem, for every
$w$
close enough to
${w}_{0}$
the image of the disc
$\left|z\right|<r$
under
$f(.,w)$
will also contain
$K$
.
$\square $
Corollary 1. The set of meromorphic functions having no exceptional values on the Riemann sphere is open in the topology of uniform convergence on compact subsets of
$\mathbf{C}$
with respect to the spherical metric in the image. The set of entire functions whose only exceptional value is
$\infty $
is not open as Example 2 above shows.

Corollary 2. Suppose that
$f(.,w)$
is entire and has an exceptional value
$a\left(w\right)\in \mathbf{C}$
for all
$w$
on some subset
$E\subset D$
. If
$a\left(w\right)$
is bounded on
$E$
then its restriction on
$E$
is continuous.

Example 4 shows that
$a$
can be discontinuous.

Theorem 2. Let
$f:\mathbf{C}\times D\to \mathbf{C}$
, be a holomorphic function, where
$D$
is a region in
$\mathbf{C}$
, and
$z\mapsto f(z,w)$
is not constant for all
$w\in D$
. Assume that for some function
$a:D\to \mathbf{C}$
we have
$f(z,w)\ne a\left(w\right)$
for all
$z\in \mathbf{C}$
.

Then there exists a discrete set
$E\subset D$
such that
$a$
is holomorphic in
$D\backslash E$
, and
$a\left(w\right)\to \infty $
as
$w\to {w}_{0}$
, for every
${w}_{0}\in E$
.

So the singularities of
$a$
can be only of the type described in Example 4.

The main ingredient of the proof of Theorem 2 is the following recent result of N. Shcherbina [11] :

Theorem A. Let
$h$
be a continuous function in a region
$G\in {\mathbf{C}}^{n}$
. If the graph of
$h$
is a pluripolar subset of
${\mathbf{C}}^{n+1}$
then
$h$
is analytic.

Proof of Theorem 2. Consider the analytic set
$$Q=\left\{\right(z,w,a)\in {\mathbf{C}}_{z}\times D\times {\mathbf{C}}_{a}:f(z,w)-a=0\}.$$
Let
$R$
be the complement of the projection of
$Q$
onto
$D\times {\mathbf{C}}_{a}$
. Then it follows from the assumptions of Theorem 2 and Picard's theorem that
$R$
is the graph of the function
$w\mapsto a\left(w\right)$
. So
$R$
is a non-empty proper subset of
$D\times {\mathbf{C}}_{a}$
. Proposition 1 implies that
$R$
is closed in
$D$
.

According to Proposition 2,
$R$
is pluripolar.

It follows from Proposition 3 that
$w\mapsto \left|a\right(w\left)\right|$
is lower semi-continuous, so the sets
$${E}_{n}=\{w\in D:|a\left(w\right)|\le n\},\text{where}n=1,2,3\dots $$
are closed. We have
${E}_{1}\subset {E}_{2}\subset \dots $
, and the assumptions of Theorem 2 imply that
$D=\cup {E}_{j}$
. Let
${E}_{j}^{0}$
be the interiors of
${E}_{j}$
, then
${E}_{1}^{0}\subset {E}_{2}^{0}\subset \dots $
, and the set
$G=\cup {E}_{j}^{0}$
is open. We claim that
$G$
is dense in
$D$
. Indeed, otherwise there would exist a disc
$U\subset D$
which is disjoint from
$G$
. But then the closed sets
${E}_{j}\backslash {E}_{j}^{0}$
with empty interiors would cover
$D$
, which is impossible by the Baire category theorem.

This proves the claim.

As
$a$
is locally bounded on
$G$
, Corollary 2 of Proposition 3 implies that
$a$
is continuous in
$G$
, so by Shcherbina's theorem,
$a$
is analytic in
$G$
. It follows from the definition of
$G$
that
$a$
does not have an analytic continuation from any component of
$G$
to any boundary point of
$G$
. Our goal is to prove that
$D\backslash G$
consists of isolated points.

If
$G$
has an isolated boundary point
${w}_{0}$
then we have

Indeed, by Proposition 3, the limit set of
$a\left(w\right)$
as
$w\to {w}_{0},w\in G$
consists of at most two points,
$a\left({w}_{0}\right)$
and
$\infty $
. On the other hand, this limit set is connected. So the limit exists. If the limit is finite, then it is equal to
$a\left({w}_{0}\right)$
and the Removable Singularity Theorem gives an analytic continuation of
$a$
to
$G\cup {w}_{0}$
, contradicting the statement above that there is no such continuation.

$$\begin{array}{c}{lim}_{w\to {w}_{0}}a\left(w\right)=\infty .\end{array}$$ | (5) |

So the limit is infinite and ( 5 ) holds.

We add to
$G$
all its isolated boundary points, thus obtaining new open set
${G}^{\prime}$
containing
$G$
. Our function
$a$
has a meromorphic continuation to
${G}^{\prime}$
, which we call
$\stackrel{~}{a}$
. This meromorphic continuation coincides with
$a$
in
$G$
but does not coincide at the added points
${G}^{\prime}\backslash G$
.

We claim that
${G}^{\prime}$
has no isolated boundary points. Indeed, suppose that
${w}_{0}$
is an isolated boundary point of
${G}^{\prime}$
. By the same argument as above, the limit ( 5 ) exists and is infinite. Then
$\stackrel{~}{a}$
can be extended to
${w}_{0}$
such that the extended function has a pole at
${w}_{0}$
, but then
${w}_{0}$
would be an isolated boundary point of
$G$
(poles cannot accumulate to a pole) so
${w}_{0}\in {G}^{\prime}$
by definition of
${G}^{\prime}$
. This contradiction proves the claim.

Let
${F}^{\prime}$
be the complement of
${G}^{\prime}$
with respect to
$D$
. Then
${F}^{\prime}$
is closed, nowhere dense subset of
$D$
. Furthermore,
${F}^{\prime}$
has no isolated points, because such points would be isolated boundary points of
${G}^{\prime}$
. So
${F}^{\prime}$
is perfect or empty. Our goal is to prove that
${F}^{\prime}$
is empty.

Assume the contrary, that is that
${F}^{\prime}$
is perfect. The closed sets
${E}_{n}$
cover the locally compact space
${F}^{\prime}$
, so by the Baire category theorem one of these
${E}_{n}$
contains a relatively open part of
${F}^{\prime}$
. This means that there exists a positive integer
$n$
and an open disc
$U\subset D$
intersecting
${F}^{\prime}$
and such that

By Corollary 2 of Proposition 3, this implies that the restriction of
$a$
on
$U\cap {F}^{\prime}$
is continuous.

$$\begin{array}{c}\left|a\right(w\left)\right|\le n\text{for}w\in U\cap {F}^{\prime}.\end{array}$$ | (6) |

We are going to prove that
$\stackrel{~}{a}$
has a continuous extension from
${G}^{\prime}$
to
$U\cap {F}^{\prime}$
, and this extension agrees with the restriction of
$a$
on
$U\cap {F}^{\prime}$
. Let
$W$
be a point of
$U\cap {F}^{\prime}$
, and
$\left({w}_{k}\right)$
a sequence in
${G}^{\prime}$
tending to
$W$
. Choosing a subsequence we may assume that there exists a limit

finite or infinite. By a small perturbation of the sequence that does not change the limit of
$\stackrel{~}{a}\left({w}_{k}\right)$
we may assume that
${w}_{k}$
are not poles of
$\stackrel{~}{a}$
so
$a\left({w}_{k}\right)=\stackrel{~}{a}\left({w}_{k}\right)$
. By Proposition 3, the limit ( 7 ) can only be
$a\left(W\right)$
or
$\infty $
.

$$\begin{array}{c}{lim}_{k\to \infty}\stackrel{~}{a}\left({w}_{k}\right),\end{array}$$ | (7) |

To prove continuity we have to exclude the latter case. So suppose that

Let
${C}_{k}$
be the component of the set
$${G}^{\prime}\cap \{w\in U:|w-W|<2|{w}_{k}-W\left|\right\},$$
that contains
${w}_{k}$
, and let

We claim that
${m}_{k}\to \infty $
. Indeed, suppose this is not so, then choosing a subsequence we may assume that
${m}_{k}\le m$
for some
$m>n$
. Then there exists a curve in
${C}_{k}$
connecting
${w}_{k}$
to some point
${w}_{k}^{\prime}\in {C}_{k}$
such that
$\left|\stackrel{~}{a}\right({w}_{k}^{\prime}\left)\right|\le m+1$
.

$$\begin{array}{c}{lim}_{k\to \infty}\stackrel{~}{a}\left({w}_{k}\right)=\infty .\end{array}$$ | (8) |

$$\begin{array}{c}{m}_{k}=inf\left\{\right|\stackrel{~}{a}\left(w\right)|:w\in {C}_{k}\}.\end{array}$$ | (9) |

As
$\left|a\right|$
is continuous in
${C}_{k}$
, ( 8 ) implies that this curve contains a point
${y}_{k}$
such that
$\left|\stackrel{~}{a}\right({y}_{k}\left)\right|=m+1.$
By selecting another subsequence we achieve that
$${lim}_{k\to \infty}\stackrel{~}{a}\left({y}_{k}\right)=y,\text{where}\left|y\right|=m+1>n.$$
As
$\left|a\right(W\left)\right|\le n$
, we obtain a contradiction with Proposition 3. This proves our claim that
${m}_{k}\to \infty $
in ( 9 ).

So we can assume that

Let us show that this leads to a contradiction. Fix
$k$
, and consider the limit set of
$\stackrel{~}{a}\left(w\right)$
as
$w\to \partial {C}_{k}\cap U$
from
${C}_{k}\cap U$
. In view of ( 6 ), ( 10 ) and Proposition 3, this limit set consists of the single point, namely
$\infty $
. To see that this is impossible, we use the following Lemma. Let
$V$
and
$C$
be two intersecting regions in
$\mathbf{C}$
, and
$g$
a meromorphic function in
$C$
such that
$${lim}_{w\to W,w\in C}g\left(w\right)=\infty \text{for all}W\in V\cap \partial C.$$
Then
$V\cap \partial C$
consists of isolated points in
$V$
and
$g$
has meromorphic extension from
$C$
to
$V\cup C$
.

$$\begin{array}{c}{m}_{k}\ge n+1\text{for all}k.\end{array}$$ | (10) |

Proof. By shrinking
$V$
, we may assume that
$\left|g\right(w\left)\right|\ge 1$
for
$w\in C\cap V$
Then
$h=1/g$
has a continuous extension from
$C$
to
$C\cup V$
by setting
$h\left(w\right)=0$
for
$w\in V\backslash C$
. The extended function is holomorphic on the set
$$\{w\in C\cap V:h(w)\ne 0\},$$
so by Rado's theorem [10,Thm.3.6.5] ,
$h$
is analytic in
$V\cup C$
,
$1/g$
gives the required meromorphic extension of
$g$
.
$\square $
Applying this Lemma with
$$C={C}_{k},V=\{w\in U:|w-W|<2|{w}_{k}-W\left|\right\}$$
and
$g=\stackrel{~}{a}$
, and taking into account that
$\partial {C}_{k}\cap V\subset {F}^{\prime}$
, and
${F}^{\prime}$
has no isolated points, we arrive at a contradiction which completes the proof that
$\stackrel{~}{a}$
has a continuous extension to
${F}^{\prime}\cap U$
, an extension which agrees with
$a$
on
${F}^{\prime}\cap U$
. From Theorem A we obtain now that
$a$
is analytic on
${F}^{\prime}\cap U$
, which contradicts the fact stated in the beginning of the proof that
$a$
has no analytic continuation from
$G$
.

This contradiction shows that
${F}^{\prime}=\varnothing $
which proves the theorem.
$\square $
The author thanks Alan Sokal for asking the question which is the subject of this paper, and Adam Epstein, Laszlo Lempert, Pietro Poggi-Corradini, Alexander Rashkovskii and Nikolay Scherbina for stimulating discussions.

References

- A. E. Eremenko and M. Yu. Lyubich, Structural stability in some families of entire functions, Funct. Anal. Appl. 19 (1985) 323–324.
- A. E. Eremenko and M. Yu. Lyubich, Dynamical properties of some classes of entire functions, Ann. Inst. Fourier, Grenoble 42 (1992) 989–1020.
- Kenneth Hoffman, Banach spaces of analytic functions, Prentice-Hall, Englewood Cliffs, N. J. 1962.
- L. Hörmander, An introduction to complex analysis in several variables, D. van Nostrand, Princeton, NJ, 1966.
- Gaston Julia, Sur le domaine d'existence d'une fonction implicite définie par une relation entière $G(x,y)=0$ , Bull. soc. math. France 54 (1926) 26-37; CRAS 182 (1926) 556.
- P. Lelong, Sur les valeurs lacunaires d'une relation a deux variables, Bull. Sci. math. ${2}^{e}$ série, t. 66 (1942) 103–108, 112–125.
- P. Lelong and L. Gruman, Entire functions of several complex variables, Springer, Berlin, 1986.
- M. Yu. Lyubich, Investigation of the stability of the dynamics of rational functions, Teor. Funktsiĭ Funktsional. Anal. i Prilozhen. No. 42 (1984) 72–91. (Russian.) Translated in Selecta Math. Soviet. 9 (1990) no. 1 69–90.
- Toshio Nishino, Sur les valeurs exceptionnelles au sens de Picard d'une fonction entière de deux variables, J. Math. Kyoto Univ., 2-3 (1963) 365–372.
- Thomas Ransford, Potential theorem in the complex plane, Cambridge UP, Cambridge 1995.
- N. V. Shcherbina, Pluripolar graphs are holomorphic, http://xxx.lanl.gov/abs/math.CV/0301181, accepted in Acta math.
- Masatsugu Tsuji, On the domain of existence of an implicit function defined by an integral relation $G(x,y)=0$ , Proc. Imperial Academy 19 (1943) 235–240.
- M. Tsuji, Potential theory in modern function theory, Maruzen, Tokyo 1959.

^{
$\text{1}$
}
It remains to see whether this set, without being a continuum, can have the power of continuum.