ON CERTAIN SUBCLASSES OF UNIVALENT

FUNCTIONS AND RADIUS PROPERTIES

M. OBRADOVIĆ and S. PONNUSAMY

Let S denote the class of normalized univalent functions f in the unit disk ∆.
One of the problems addressed in this paper is that of the F -radius in G when
F , G ⊂ S, namely the maximum value of r0 such that r−1 f (rz) ∈ G for all f ∈ F
and 0 < r ≤ r0 . The investigations are concerned primarily with the classes U
and P(2) consisting of univalent functions satisfying
 2  00
z z
f 0 (z) − 1 ≤ 1 and ≤ 2,
f (z) f (z)
respectively, for all |z| < 1. Similar radius properties are also obtained for a
geometrically motivated subclass Sp ⊂ S. Several new sufficient conditions for f
to be in the class U are also presented.
AMS 2000 Subject Classification: 30C45.

Key words: coefficient inequality, analytic, univalent and starlike functions.

1. INTRODUCTION AND PRELIMINARIES

Denote by A the class of all functions f , normalized by f (0) = f 0 (0) −

1 = 0, that are analytic in the unit disk ∆ = {z ∈ C : |z| < 1}, and
by S the subclass of univalent functions in ∆. Denote by S ∗ the subclass
consisting of functions f in S that are starlike (with respect to origin), i.e.,
tw ∈ f (∆) whenever t ∈ [0, 1] and w ∈ f (∆). Analytically, f ∈ S ∗ if and only
if Re (zf 0 (z)/f (z)) > 0 in ∆. A function f ∈ A is said to belong to the class
U if
 2
0 z
f (z) − 1 ≤ 1, z ∈ ∆.
f (z)
In [6], the authors introduced a subclass P(2) of U, consisting of functions f
for which
 00
z
≤ 2, z ∈ ∆.
f (z)

REV. ROUMAINE MATH. PURES APPL., 54 (2009), 4, 317–329

318 M. Obradović and S. Ponnusamy 2

We have the strict inclusion P(2) ( U ( S (see [1, 6, 10] for a proof). An
interesting fact is that each function in
 
z z z z
SZ = z, , , ,
(1 ± z)2 1 ± z 1 ± z 2 1 ± z + z 2
belongs to U. Also, it is well-known that these are the only functions in S
having integral coefficients in the power series expansions of f ∈ S (see [2]).
From the analytic characterization of starlike functions, it is a simple exercise
to see that SZ ⊂ S ∗ .
Further work on the classes U and P(2), including some interesting gen-
eralizations of these classes, may be found in [7, 9, 11]. A function f ∈ S ∗ is
said to be in T ∗ if it can be expressed as
∞
X
f (z) = z − ak z k ,
k=2
where ak ≥ 0 for k = 2, 3, . . .. Functions of this form are discussed in de-
tail by Silverman [13, 14]. The work of Silverman led to a large number of
investigations for univalent functions of the above form.
In this paper we shall be mainly concerned with functions f ∈ A of
the form
z
(1) f (z) = ∞ , z ∈ ∆.
n
P
1+ bn z
n=1
The class of functions f of this form for which bn ≥ 0 is especially interesting
and deserves a separate discussion. We remark that if f ∈ S then z/f (z) is
nonvanishing in the unit disk ∆. Hence it can be represented as Taylor series
of the form
∞
z X
=1+ bn z n , z ∈ ∆.
f (z)
n=1
The above representation is convenient for our investigation.
Now, we introduce a subclass Sp of starlike functions, namely,
zf 0 (z) zf 0 (z)
 
∗
Sp = f ∈ S : − 1 ≤ Re , z∈∆ .
f (z) f (z)
Geometrically, f ∈ Sp if and only if the domain values of zf 0 (z)/f (z), z ∈ ∆,
is the parabolic region (Im w)2 ≤ 2 Re w − 1. It is well-known [12, Theorem 2]
that f (z) = z + an z n is in Sp if and only if (2n − 1)|an | ≤ 1.
Let F and G be two subclasses of A. If for every f ∈ F, r−1 f (rz) ∈ G
for r ≤ r0 , and r0 is the maximum value for which this holds, then we say that
3 Univalent functions and radius properties 319

r0 is the G-radius in F. There are many results of this type in the theory of
univalent functions. For example, the Sp -radius in S ∗ was found by Rønning
in [12] to be 1/3. Moreover, the class Sp and its associated class of uniformly
convex functions, introduced by Goodman [4, 5], have been investigated in
[12]. We recall here the following result.
1
Theorem A [12, Theorem 4]. If f ∈ S then r f (rz) ∈ Sp for 0 < r ≤
0.33217 . . ..
The paper is organized as follows. We investigate the P(2)-radius in F,
where F is the subclasses of U consisting of functions f ∈ U of the form (1)
∞
P
that satisfies either the condition (n − 1)|bn | ≤ 1 (see Theorem 1) or bn ≥ 0
n=2
(see Corollary 1). In Theorem 2 we obtain a necessary coefficient condition
for a function f of the form (1) with bn ≥ 0 to be in Sp , while in Theorem 3
we obtain a sufficient coefficient condition for a nonvanishing analytic function
z/f (z) of the form (1) (where bn ∈ C) to be in Sp . In Theorem 4 we derive the
value of the S-radius in Sp . In Theorems 5 and 6 we establish new necessary
and sufficient conditions for a function to belong to the class U. Finally, in
Corollary 2 we show that T ∗ ⊂ U, which is somewhat surprising.

2. LEMMAS

For the proof of our results we need the following result (see [3, Theo-
rem 11 on p. 193 of Vol. 2]) which reveals the importance of the area theorem
in the theory of univalent functions.
∞
Lemma 1. Let µ > 0 and f ∈ S be of the form (z/f (z))µ = 1 + bn z n .
P
n=1
∞
(n − µ)|bn |2 ≤ µ.
P
Then we have
n=1

We also have
∞
bn z n be a non-vanishing analytic
P
Lemma 2 ([9]). Let φ(z) = 1 +
n=1
function in ∆ and f (z) = z/φ(z). Then
∞
P
(a) f ∈ U if (n − 1)|bn | ≤ 1;
n=2
∞
P
(b) f ∈ P(2) if n(n − 1)|bn | ≤ 2.
n=2
320 M. Obradović and S. Ponnusamy 4

3. MAIN RESULTS

It is easy to see that the rational function f (z) = z/(1 + Az 3 ) belongs to

U if and only if |A| ≤ 1/2. Further, for |A| = 1/2 we have
|(z/f (z))00 | = |6Az| ≤ 3|z| ≤ 2
provided |z| ≤ 2/3. It seems reasonable to expect that the P(2)-radius in U
is at least 2/3, and we formulate our first result.
∞
P
Theorem 1. If f ∈ U is of the form (1) such that (n − 1)|bn | ≤ 1,
n=2
then 1r f (rz) ∈ P(2) for 0 < r ≤ 2/3; and 2/3 is the largest number with this
property, especially in the class for which b1 = b2 = 0.
Proof. Let f ∈ U be of the form (1). We need to show that 1r f (rz) ∈ P(2)
for 0 < r ≤ 2/3. Using (1), for 0 < r ≤ 1 we can write
∞
z X
1 =1+ (bn rn )z n .
r f (rz) n=1

According to Lemma 2(b), it suffices to show that

∞
X
n(n − 1)|bn |rn ≤ 2
n=2

for 0 < r ≤ 2/3. It is easy to see by induction that nrn ≤ 3r for all 0 <
r ≤ 2/3 and for n ≥ 2. In view of this observation, and the assumption that
∞
P
(n − 1)|bn | ≤ 1, we obtain
n=2
∞
X
n(n − 1)|bn |rn ≤ 3r ≤ 2 for r ≤ 2/3.
n=2

Hence, by Lemma 2(b), 1r f (rz) ∈ P(2) for 0 < r ≤ 2/3.

To prove the sharpness, we consider fθ (z) = z/(1 + eiθ z 3 /2). Then we
observe that fθ ∈ U, but it does not belongs to P(2). We see that 1r fθ (rz) ∈
P(2) for 0 < r ≤ 2/3 and r = 2/3 is the largest value with the desired
property. 
An interesting consequence of Theorem 1 is stated later in Corollary 1.
Theorem 2. If a function f of the form (1) with bn ≥ 0 is in Sp , then
∞
X
(2) (2n − 1)bn ≤ 1.
n=1
5 Univalent functions and radius properties 321

Proof. Let f ∈ Sp . Then

 0  2
z z z
(3) z = − f 0 (z).
f (z) f (z) f (z)

Therefore, as f ∈ Sp is of the form (1), we have

0
  0
z z z
−z −z
zf 0 (z)
 0 
zf (z) f (z) f (z) f (z)
− 1 ≤ Re ⇔ z ≤ Re z
f (z) f (z)
f (z) f (z)
∞  ∞ 
nbn z n nbn z n
P P
−
n=1  n=1 
⇔ ∞ ≤ Re 
 1 − ∞
.

n n
P P
1+ bn z 1+ bn z
n=1 n=1

If z ∈ ∆ is real and tends to 1− through reals, then from the last inequality
we have
∞
P  P∞ 
nbn nbn
n=1 1 − n=1
 
∞ ≤ Re  ∞
,
P P 
1+ bn 1+ bn
n=1 n=1
∞
P
from which we obtain the desired inequality (2n − 1)bn ≤ 1. 
n=1

Remark 1. Condition (2) for functions of the form (1) with nonnegative
coefficients bn is not sufficient for the corresponding f to be in the class Sp .
As an example, consider the function f (z) = z/(1 + z). It is easy to see that
the condition for the class Sp , namely,

zf 0 (z) zf 0 (z)
− 1 ≤ Re ,
f (z) f (z)
√
does not hold for all z ∈ ∆, for example at the boundary point z = (−1+i)/ 2,
hence at some points in ∆.

Remark 2. Let 0 < λ < 1 and f (z) = z − λz m , where m ≥ 2. Then

∞
z 1 X
= = 1 + λk z k(m−1) ,
f (z) 1 − λz m−1
k=1
322 M. Obradović and S. Ponnusamy 6

which is nonvanishing in the unit disk. It follows from the previous theorem
that if f ∈ Sp , the coefficient must satisfy the condition
∞
X
[2k(m − 1) − 1]λk ≤ 1,
k=1
which simplifies to λ(2m − 1) ≤ 1. Thus, a necessary condition for f to belong
to Sp is 0 ≤ λ ≤ 1/(2m − 1). It is a simple exercise to see that this condition
also is a sufficient condition for f ∈ Sp (see also [12, Theorem 2]). Thus,
the upper bound for λ cannot be improved. This observation shows that the
constant 1 on the right hand side of inequality (2) cannot be replaced by a
larger constant. In this sense, condition (2) is sharp.
Theorem 3. Let f (z) be a nonvanishing analytic function in 0 < |z| < 1
of the form (1). Then the condition
X∞
(4) (2n + 1)|bn | ≤ 1
n=1
is sufficient for f to belong to the class Sp .
Proof. As in the proof of Theorem 2, we notice that
∞  ∞ 
n n
P P
nb n z nb n z
zf 0 (z) zf 0 (z)
 
⇔ − n=1 ∞ n=1
 
− 1 ≤ Re ≤ Re  1 − ∞
.
f (z) f (z) P n
 P n

1+ bn z 1+ bn z
n=1 n=1
Thus, to show that f is in Sp , it suffices to show that the quotient
∞
nbn z n
P
n=1
− ∞
bn z n
P
1+
n=1
lies in the parabolic region (Im w)2 ≤ 1 + 2 Re w. Geometric considerations
show that this condition holds if
∞
nbn z n
P
n=1 1
(5) ∞ ≤ , z ∈ ∆.
2
bn z n
P
1+
n=1
∞
(2n + 1)|bn | |z|n ≤ 1 and so
P
From condition (4) we obtain that
n=1
∞ ∞
!
X 1 X
n|bn | |z|n ≤ 1− |bn | |z|n .
2
n=1 n=1
7 Univalent functions and radius properties 323

Finally, we find that

∞ ∞
nbn z n |bn | |z|n
P P
1−
n=1 1 n=1 1
∞ ≤ ∞ = .
2 2
bn z n |bn | |z|n
P P
1+ 1−
n=1 n=1

This means that inequality (5) holds and, therefore, f ∈ Sp . 

Theorem 4. If f ∈ S is given by (1), then 1r f (rz) ∈ Sp for 0 < r ≤

r0 , where r0 , which depends on the second coefficient of f , is the root of the
equation
4 4
(6) + − (8 + 12r2 ) − 9r2 ln(1 − r2 ) = (1 − (3/2)|f 00 (0)|r)2 .
(1 − r2 )2 1 − r2
Proof. Let f ∈ S be given by (1). Then z/f (z) is nonvanishing and for
0 < r ≤ 1 we have
z
1 = 1 + (b1 r)z + (b2 r2 )z 2 + · · · (b1 = −f 00 (0)/2),
r f (rz)
and if
∞
X
(7) (2n + 1)|bn |rn ≤ 1
n=1
1
for some r, then r f (rz) ∈ Sp by Theorem 3. By Lemma 1 with µ = 1, we have
∞
X
(8) (n − 1)|bn |2 ≤ 1.
n=2

Now, by the Cauchy-Schwarz inequality and (8),

∞ ∞
X
n
X √ 2n + 1 n
(2n + 1)|bn |r = n − 1|bn | √ r
n=2 n=2
n−1
∞
!1 ∞
!1
2
X X (2n + 1)2 2n 2
≤ (n − 1)|bn |2 r
n−1
n=2 n=2
∞
!1  21
2
(2n + 1)2 16r4 − 12r6
X 
2n 2 2
≤ r = − 9r ln(1 − r ) .
n−1 (1 − r2 )2
n=2

In particular, for 0 < r ≤ r0 , the last expression is less than or equal to

1 − 3|b1 |r. Therefore, (7) holds, concluding the proof. 
324 M. Obradović and S. Ponnusamy 8

Remark 3. One can easily show that equation (6) has a unique solution
for 0 < r ≤ 1 and |b1 | ≤ 1/3. Indeed, let
4 4
G(r) = + − (8 + 12r2 ) − 9r2 ln(1 − r2 ) − (1 − 3|b1 |r)2
(1 − r2 )2 1 − r2
and 1 − r2 = x. Now, for 0 ≤ x < 1 we consider the new function
4 4 √
H(x) = 2 + + 12x − 20 − 9(1 − x) ln x − (1 − 3|b1 | 1 − x)2 .
x x
For this function, we see that H(x) → +∞ when x → 0+, H(1) = −1, and
√
 
0 8 4 1−x
6|b1 |
H (x) = − 3 − 2 + 12 − 9 − ln x + − 1 − 3|b1 | 1 − x √
x x x 1−x
3 2 √
     
1−x 1−x 1−x 6|b1 |
= −8 3
−4 2
+ 9 ln x − 9 − (1 − 3|b1 | 1 − x) √ ,
x x x 1−x
which is negative for 0 < x < 1 while |b1 | ≤ 1/3, showing that equation (6)
has a unique solution in the interval (0, 1).
Also, in Theorem 4, we have actually obtained F-radius in S, where F
is the subclass of Sp consisting of functions f given by (1) with coefficients
∞
P
satisfying the condition (2n + 1)|bn | ≤ 1.
n=1

Remark 4. For f 00 (0) = 0 in Theorem 4, we have r0 = 0.30066 . . ., and

the result is the best possible, the extremal function being of the form
∞
z X 2n + 1
=1+ rn z n .
f (z) n−1 0
n=2

To see this, for |ζ| < 1 we have

∞
r0 ζ X
=1+ bn ζ n ,
f (r0 ζ)
n=2

where
2n + 1 2n
bn = r
n−1 0
and
∞ ∞
X X (2n + 1)2
(2n + 1)|bn | = r02n = 1,
n−1
n=1 n=2

by the definition of r0 from (6). This means that 1r f (rz) belongs to Sp for
∞
P
0 < r ≤ r0 . Moreover, for |z| = r > r0 we have (2n + 1)|bn | > 1. Therefore,
n=1
9 Univalent functions and radius properties 325

f is extremal for the class F of functions f given by (1) with coefficients

∞
P
satisfying the condition (2n + 1)|bn | ≤ 1. In this sense, the result is sharp.
n=1
On the other hand, the function f is univalent because it can be easily
seen that f ∈ U. Indeed, we have
∞ ∞
X X 2r0 r02
(n − 1)|bn | − 1 = (2n + 1)r0n − 1 = − 2r 0 + −1
(1 − r0 )2 1 − r0
n=2 n=2
(1 − 3r0 )[r0 (1 − r0 ) + 1]
=− < 0.
(1 − r0 )2

According to Lemma 2(a), f ∈ U, hence f is univalent. Finally, we only need

to prove that the function
∞
z X 2n + 1 (r0 z)2
=1+ r0n z n = 1 + 2 − 3r0 z log(1 − r0 z)
f (z) n−1 1 − r0 z
n=2

has no zeros in the unit disk. This is easy because

  ∞ ∞
z X 2n + 1 n X
Re =1− r ≥1− (2n + 1)r0n > 0.
f (z) n−1 0
n=2 n=2

Thus, we have established that r0 in Theorem 4 is the best possible radius

when f 00 (0) = 0. In other words, if F is the subclass of functions f ∈ S of the
form (1) such that f 00 (0) = 0, then 1r f (rz) belongs to Sp for 0 < r ≤ r0 , where
r0 is the largest value with the desired property.

√ U ( S. In [8], the authors have shown that the U-radius

It is known that
in the class S is 1/ 2. Our next result is simple but is surprising as it identifies
an important subclass of S which lies in U. We remark that a function f ∈ U
does not necessarily imply that Re(f 0 (z)) > 0 throughout |z| < 1, see [7].

Theorem 5. If f is given by (1) with bn ≥ 0 and such that Re(f 0 (z)) > 0
for z ∈ ∆, then f ∈ U.

Proof. Remark that if f ∈ A satisfies Re(f 0 (z)) > 0 for z ∈ ∆, then f

must be univalent in ∆ (see [3]). Also, notice that
 0 ∞
z z
(n − 1)bn z n
P
−z 1−
0 f (z) f (z) n=2
Re(f (z)) > 0 ⇔ Re 2 > 0 ⇔ Re  2 > 0.
∞

z
bn z n
P
1+
f (z) n=1
326 M. Obradović and S. Ponnusamy 10

For z → 1− along the positive real axis, the last inequality above becomes
P∞
1− (n − 1)bn
n=2
Re 
∞
2 ≥ 0,
P
1+ bn
n=1

∞
P
which gives (n − 1)bn ≤ 1 and so f ∈ U, by Lemma 2(a). 
n=2

Theorem 6. A function f of the form (1) with bn ≥ 0 and z/f (z) 6= 0

in ∆, is in U if and only if
∞
X
(9) (n − 1)bn ≤ 1.
n=2

Proof. On account of Lemma 2(a), it suffices to prove the necessary part.

To do this, we let f ∈ U of the form (1). This means that
 2  0 ∞
z 0 z z X
f (z) − 1 = −z −1 = (n − 1)bn z n < 1.
f (z) f (z) f (z)
n=2

Choosing values of z on the real axis and then letting z → 1− through real
values, we obtain the coefficient condition (9). 
For example, by (9), the functions
z z z z
, , and
(1 + z)2 1 + z 1 + z 2 1 + z + z2
are in U.
As an immediate consequence of Theorems 1 and 6, we have the following
result.

Corollary 1. If f ∈ U is of the form (1) such that bn ≥ 0, then

1
r f (rz)∈ P(2) for 0 < r ≤ 2/3; and 2/3 is the largest number with this
property, at least when b1 = 0 = b2 .

We next show that a certain class of functions in S ∗ is in U, which is

again a surprising simple result. Using this result, we can generate functions
in S ∗ that are also in U.

Theorem 7. If f ∈ S ∗ is of the form (1) with bn ≥ 0, then the coefficient

inequality (9) holds.
11 Univalent functions and radius properties 327

Proof. Suppose that f ∈ S ∗ is of the form (1) with bn ≥ 0. We have

 0 
zf (z) zf 0 (z) zf 0 (z)
Re >0⇔ −1 < +1
f (z) f (z) f (z)
 0 ∞
z
nbn z n
P
−z −
f (z) n=1
⇔  0 < 1 ⇔ ∞ < 1.
z z 2 + b z −
P
(n − 2)b z n
2 −z 1 n
f (z) f (z) n=3

For z → 1− through real values, from the last inequality we obtain that
P∞
nbn
n=1
∞ ≤ 1,
P
2 + b1 − (n − 2)bn
n=3
which is equivalent to (9). Therefore, f ∈ U. 

Remark 5. Although condition (9) will be a useful necessary condition

for a rational function f of the form (1) (with bn ≥ 0) to be starlike, it is not
a sufficient condition for the starlikeness for functions f ∈ U. To prove this,
we consider the function
z
f1 (z) = .
1 + 12 z + 12 z 3
By Theorem 6, f1 ∈ U. On the other hand, it is easy to see that
zf10 (z) 1 − z3
=
f1 (z) 1 + 12 z + 12 z 3
√
and at the boundary point z0 = (−1 + i)/ 2, we have
√ √
z0 f10 (z0 ) 2−2 2 1−2 2
= + i,
f1 (z0 ) 3 3
which implies that Re {z0 f10 (z0 )/f1 (z0 )} < 0. Consequently, there are points
in the unit disk |z| < 1 for which Re {zf10 (z)/f1 (z)} < 0, which shows that the
function f1 is not starlike in ∆.
∞
an z n is in S ∗ , where an ≥ 0 for n ≥ 2,
P
Corollary 2. If f (z) = z −
n=2
then f ∈ U.
Proof. Let f ∈ S ∗ . Then z/f (z) is nonvanishing in the unit disk. So,
z/f (z) can be expressed as
z 1
= = 1 + b1 z + b2 z 2 + · · · ,
f (z) 1 − a2 z − a3 z 2 − · · ·
328 M. Obradović and S. Ponnusamy 12

where bn ≥ 0 for all n ∈ N. Then, by Theorem 7, the inequality

∞
X
(n − 1)bn ≤ 1
n=2

holds. Hence, by Theorem 6, f ∈ U. 

Corollary 2 is especially helpful in obtaining functions that are both
starlike as well as in U, as there are numerous results concerning starlike
functions with negative coefficients. For example, fm (z) = z − z m /m is in S ∗ ,
∞ ∞
|an |z n is in S ∗ if and only if
P P
hence in U. Since f (z) = z − n|an | ≤ 1
n=2 n=2
(see [13, Theorem 2]), this result can be used to generate functions f ∈ U that
are not starlike.

Acknowledgements. The authors thank the referee for their suggestions for the
improvement of the presentation. The work was initiated during the visit of the
second author to the University of Turku, Finland. The visit was supported by the
Commission on Development and Exchanges (CDE) of the International Mathematical
Union and this author thanks CDE for its support.

REFERENCES

[1] L.A. Aksentév, Sufficient conditions for univalence of regular functions. Izv. Vysš.
Učebn. Zaved. Matematika 1958, 3 (4), (1958), 3–7. (Russian)
[2] B. Friedman, Two theorems on schlicht functions. Duke Math. J. 13 (1946), 171–177.
[3] A.W. Goodman, Univalent functions, Vols. 1–2. Mariner, Tampa, Florida, 1983.
[4] A.W. Goodman, On uniformly convex functions. Ann. Polon. Math. 56 (1991), 87–92.
[5] A.W. Goodman, On uniformly starlike functions. J. Math. Anal. Appl. 155 (1991),
364–370.
[6] M. Nunokawa, M. Obradović and S. Owa, One criterion for univalency. Proc. Amer.
Math. Soc. 106 (1989), 1035–1037.
[7] M. Obradović and S. Ponnusamy, New criteria and distortion theorems for univalent
functions. Complex Variables Theory Appl. 44 (2001), 173–191.
[8] M. Obradović and S. Ponnusamy, Radius properties for subclasses of univalent functions.
Analysis (Munich) 25 (2006), 183–188.
[9] M. Obradović, S. Ponnusamy, V. Singh and P. Vasundhra, Univalency, starlikeness and
convexity applied to certain classes of rational functions. Analysis (Munich) 22 (2002),
225–242.
[10] S. Ozaki and M. Nunokawa, The Schwarzian derivative and univalent functions. Proc.
Amer. Math. Soc. 33 (1972), 392–394.
[11] S. Ponnusamy and P. Vasundhra, Criteria for univalence, starlikeness and convexity.
Ann. Polon. Math. 85 (2005), 121–133.
[12] F. Rønning, Uuniformly convex functions and a corresponding class of starlike functions.
Proc. Amer. Math. Soc. 118 (1993), 189–196.
13 Univalent functions and radius properties 329

[13] H. Silverman, Univalent functions with negative coefficicnts. Proc. Amer. Math. Soc. 51
(1975), 109–116.
[14] H. Silverman, E.M. Silvia and D.N. Telage, Locally univalent functions and coefficient
distortions. Pacific J. Math. 77 (1978), 533–539.

Received 5 December 2008 Faculty of Civil Engineering

Department of Mathematics
Bulevar Kralja Aleksandra 73
11000 Belgrade, Serbia
[email protected]
and
Indian Institute of Technology
Department of Mathematics
Madras, Chennai–600 036, India
[email protected]